The Ethics of Medical Research On Humans

The ethics of medical research on humans
One of the most diYcult problems that confronts clinicians and medical
professionals is how to apply ethical principles to real decisions aVecting
patients. In this even-handed book, Claire Foster examines the three main
approaches to moral decision-making: goal-based, duty-based and right-
based. She examines the underlying philosophical arguments of each, their
relative strengths and weaknesses, and how they can actually be applied. She
also looks at the problematic boundaries where treatment ends and experi-
mentation begins. Is it ethical to experiment with new cures on people who are
probably dying anyway? And how do you assess quality of consent? This book
provides a thorough, non-partisan grounding in what the ethical principles are
and what informs them. It is an invaluable preparation both for a researcher
being interviewed by an ethics committee and for the people sitting on the
committee, and will be essential reading for all medical decision-makers.
Claire Foster has developed a unique and systematic approach to analysing the
ethics of human participant research through her experience running the
Research Ethics Committee Project at the Kings College Centre of Medical
Law and Ethics. She also has extensive experience of training members of
research ethics committees and commenting on guidelines drawn up by such
bodies as the Medical Research Council, the Royal College of Physicians and
the NuYeld Council on Bioethics.
The ethics of medical
research on humans
Claire Foster
iuniisuio n\ rui iiiss s\xoicari oi rui uxiviisir\ oi caxniioci
The Pitt Building, Trumpington Street, Cambridge, United Kingdom
caxniioci uxiviisir\ iiiss
The Edinburgh Building, Cambridge CB2 2RU, UK
40 West 20th Street, New York, NY 10011-4211, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
Ruiz de Alarcn 13, 28014 Madrid, Spain
Dock House, The Waterfront, Cape Town 8001, South Africa
http://www.cambridge.org
First published in printed format
ISBN 0-521-64196-9 hardback
ISBN 0-521-64573-5 paperback
ISBN 0-511-01800-2 eBook
Cambridge University Press 2004
2001
(netLibrary)
To members of research ethics committees past and present

MMMM
Contents
Foreword by Professor Sir David Weatherall xi
Acknowledgements xiii
1 An introduction to the ethical issues 1
Introduction 1
What is the value of research? 3
What are the limitations of research? 5
What is the right way to treat human research participants? 6
How can research participants views be respected? 8
Three areas of ethical concern in research: science, best interests
and autonomy 9
2 Goal-based morality: scientic rigour in research 13
The foundations of goal-based thinking 13
Research should maximize health and minimize harm 13
Utilitarianisms strengths and weaknesses 13
Goal-based moral thinking applied to medical research 17
The application of goal-based thinking 18
The goals of research 18
Methods of research 21
Disseminating the results of research 25
Summary and concluding remarks 30
3 Duty-based morality: acting in the research subjects
best interests 31
The foundations of duty-based thinking 31
From goals to duties 31
Natural law ethics 32
Kants categorical imperative 34
Duty-based moral thinking applied to medical research 36
vii
The application of duty-based thinking 39
Therapeutic research 39
Non-therapeutic research 43
4 Right-based morality: respecting the autonomy of
research participants 47
The foundations of right-based thinking 47
From duties to rights 47
Denitions of rights 48
The validity of rights 53
The applications of right-based thinking 57
The consent procedure 57
Concluding remarks on consent 61
Condentiality 61
5 From principles to practice 65
Introduction 65
Goal-based morality 65
Goal-based moralitys theoretical basis summarized 65
Goals of research in theory and practice 66
Research method summarized 67
The practical implications of the choice of research method 67
Summary of the problems with disseminating the results of
research 67
Introduction to practical examples of dissemination problems 68
Duty-based morality 68
Duty-based moralitys theoretical basis summarized 68
The practical implications of duty-based morality 69
Introduction to examples 69
Right-based morality 71
Right-based moralitys theoretical basis summarized 71
Introduction to examples 71
Does the three-approaches framework succeed? 72
6 Case studies of goal-based issues 73
Introduction 73
Goals of research 73
Organ transplantation 73
viii Contents
Maintaining the condition of donor organs 75
Xenotransplantation 77
Autografting using cloned embryos 79
Methods of research 80
The discovery of penicillin was due to luck and hard work,
but not planning 81
Alternative and complementary therapy research needs open
minds 85
Trials in homeopathy 86
Dissemination of the results of research 88
Results of research into futile treatment depend on what is
understood by futile 88
Results of pharmaceutical company research are always
commercially favourable 89
Results should ultimately meet public need 90
7 Case studies of duty-based issues 93
Introduction 93
Therapeutic research 94
Duty to care versus scientic goals: placebo controls in
therapeutic research 94
Trials of folic acid in pregnancy 95
Trials for treatments of peptic ulcer disease 96
Scientic arguments against the use of placebo 97
The FDAs arguments for requiring placebo 98
Meta-analysis of trials of ondansetron 100
Concluding remarks 102
Non-therapeutic research 103
Duty to care versus scientic goals: potential risks
in non-therapeutic research 103
Duty to care versus patient autonomy: non-therapeutic healthy
volunteer research indicates the need to protect subjects
from harm even if they consent 106
8 Case studies of right-based issues 113
Introduction 113
Consent 113
Right-based difculties with consent: the empirical evidence 114
Duty-based difculties with consent 121
ix Contents
Goal-based difculties with consent 123
Written consent 126
ConWdentiality 130
Should records-based or epidemiological research take place
if it compromises patient autonomy? 130
9 A framework for ethical review: researchers, research
ethics committees, and moral responsibility 133
Introduction 133
The three approaches combined 133
A framework to assist ethical review 135
Goal-based questions 135
Duty-based questions 136
Right-based questions 136
Resolving conXicts between the three approaches 136
Goal-based questions to set the context 136
Goal-based and duty-based moral imperatives in conict 136
Goal-based and right-based 137
Duty-based and right-based 138
Goal-based and right-based again 139
When the three approaches fail 139
Research ethics committees 140
The Nuremberg Code 141
Pappworths guinea pigs 141
Sir Austin Bradford-Hill and Mrs. Hodgson 142
The Medical Research Council, the World Medical Association
and the Royal College of Physicians 143
The Ministry of Health 143
The growth of research ethics committees power 143
Multi-centre research ethics committees 144
Who guards the guardians? 145
Bradford-Hills prophecy 145
Conclusion 146
References 147
Index 155
x Contents
Foreword
A few weeks ago I was sitting at my desk, when there was a knock at the door
and two characters walked in who bore an uncanny resemblance to two
parking meter attendants with whom I had had a mild altercation while
parking my car earlier in the day. We are carrying out an organ inspection
one of them said, we have to look through your oYce to see if you are
hoarding any human tissues or related material. After they had left, empty
handed I hasten to add, I mused on how close we are to Aldous Huxleys
Brave New World and pondered on the extent to which we, the doctors, have
brought all this on ourselves by centuries of insensitivity to the feelings of
society.
In the UK, the last few years have seen an endless attack on the medical
profession: clinical incompetence; abysmal communication; the storage of
childrens organs for medical research without permission from their par-
ents; and so much more. All this has brought the inevitable short-term,
knee-jerk reactions from Government, and the profession is likely to come
under increasingly close scrutiny and bureaucratic regulation. This is hap-
pening at a time when we are about to enter one of the most exciting but
complex periods of medical research. As the fall-out fromthe human genome
project is applied to clinical research and practice it will open up a whole
series of new ethical dilemmas and will undoubtedly raise many new con-
cerns about research on human beings. The more gloomy commentators on
the current scene are raising vistas of Nazi Germany, the eugenics movement
and the kind of human experimentationthat went on in concentration camps
under the guise of medical research. Already there are signs of over-reaction,
and regulations are being established which will, in the long term, have a
deleterious eVect on many areas of medical research.
Some good has come from these increasing concerns about the conduct of
medical research, however. The Government and the bodies which fund
research have become increasingly aware of the importance of the ethical
basis of clincal investigation, demanding that young clinical research workers
are exposed to instruction in good practice and that all research involving
patients is scrutinized more thoroughly by appropriate ethics committees.
xi
The bodies which control medical practice have established their own ethical
review processes and this country has taken a valuable lead in establishing an
entirely independent body of this type, free of any inXuence from Govern-
ment or the medical establishment and industry, the NuYeld Council on
Bioethics. Most medical schools are now providing students with courses on
ethics at diVerent stages of their training and many universities are establish-
ing departments of bioethics.
While all this activity is commendable, and will undoubtedly do much to
restore peoples faith in the activities of the medical profession, there is a
danger that the simple principles on which the ethical basis of medical
researchare founded are lost in a mist of political correctness. In this splendid
new book, Claire Foster has provided an excellent account of the underlying
philosophy on which the ethics of medical research is based. Building on this
foundation she has examined some of the major issues which cause so much
confused thinking about the ethics of experimentation on humans and, using
a series of well chosen examples, has provided sound and commonsense
guidelines for medical research workers.
Commendably short and free of the jargon which haunts this Weld, this
introduction to the really important issues of research on humans should
provide young clinical research workers with a solid basis on which to
develop their programmes, and, at the same time, help to remind established
clinical scientists about the central ethical issues which underline this Weld,
and which are so often lost in the complex maze of ill-thought-through
responses to new problems which, sadly, typiWes the current biomedical
research scene.
Claire Foster has done us all a great service in writing this small and
extremely clear account of the fundamentals of bioethics as they apply to
research on human beings. I wish this book all the success it deserves.
D. J. Weatherall
Oxford, June 2001
xii Foreword
Acknowledgements
I would like to thank all those who had a hand in making this book possible.
All my teachers played their part in this, but I want to thank Sophie Botros in
particular, whose careful imparting to me of her approach to moral philos-
ophy and medical ethics has so inXuenced this book. I should also like to
thank David Lloyd, Laura Wilson, James and Katie Glover and Cyril Chap-
man for their very helpful comments on earlier drafts. Jo Sumner, Madeleine
Barnes and Rebecca Fallon all provided vital support when it was needed, for
which grateful thanks are due. Most of all I want to thank David for his
unfailing, gentle presence and unselWsh support.
xiii
aaaaa
1
1
An introduction to the ethical issues
Introduction
It must be the dreamof any ill person to be cured eVectively and immediately,
with no side eVects. Every doctors dream must be to provide such a precise
service. It might happen sometimes, but the reality is rarely so satisfactory.
Even when miracle cures like penicillin are discovered, the appearance of
absolute cure with no side eVects turns out to be diVerent from the actual
experience, sometimes long after the medicine has been discovered. How-
ever, it is just as well that throughout the history of medicine, some doctors
have never accepted the idea that complete cures are a delusion and stopped
looking for them. For if research is not undertaken, medicine would not
progress in the remarkable ways that it has. There may not be many complete
cures, but there are treatments for numerous conditions that previously
would have killed or disabled for life. It has also been established that some
treatments are useless or even harmful. The ultimate goal of medical research
must be to Wnd complete cures; the more prosaic actual achievements do,
nevertheless, help a great deal.
To improve medical care as much as we can, if not to perfect it, means that
we have to accept the need for research. Some argue that the real art of
medical care is to prevent people falling ill in the Wrst place. Prevention is
better than cure, particularly if it does not involve taking drugs. Even to
establish what constitutes healthy living requires research, however. In any
case, prevention is helpful to those who have not yet succumbed to the eVects
of unhealthy living, but for those for whom it is too late, treatment is needed.
Also, there are many causes of conditions which, not being understood, or
being understood but not being controllable, cannot be avoided or changed.
Research into causes is needed, and so is research into treatment of the
conditions as they present themselves. Whatever the condition or its cause,
medical research is needed. What is more, that research is almost always
going to take the form of steps on the way to complete cures, rather than
reaching the goal in one go. Giant leaps in understanding and treatment are
not, by their very nature, planned, as the story of penicillins discovery
demonstrates. Meanwhile, the pedestrian plodding of routine research has to
2 An introduction to the ethical issues
go on. Over time it can show startlingly good results, such as the hard-won
50% improvement in the treatment of childhood leukaemia.
The recognition of the need for careful research, and participation in it,
requires sacriWces on the part both of patients and of doctors. Doctors have to
recognize that what knowledge they have had imparted to them is not
complete, and that there is always more to learn and pass on within their
discipline. To learn, doctors have to be ready to question their established
practices and beliefs, and to recognize the possibility of really diVerent ways
of treating diseases. To pass on research results, doctors have to be able to
communicate with their peers. Research means detailed and disciplined
work. Research projects have to be planned and carried out, and their results
disseminated. Patients, who would far rather not be treated as guinea pigs,
have to be encouraged to want to help. Doctors may risk losing patients to
colleagues who do not ask them to take part in research programmes.
Doctors who do undertake research need to remember that even in the midst
of a researchproject their patients still require their best interests to be served,
and that those interests come before the successful completion of a project,
should there be a conXict. Enthusiasm for reaching the goals of research
should not make doctors view their patient participants merely as good
clinical material.
Patients have to recognize that if medical care is to continue to improve
then they must play their part too, and allow their treatments to be oVered as
part of research programmes, if that is the best way to ensure continuing
improvement. If research is well designed then their treatment should not be
inferior, but they may have to accept that a computer, not a doctor, will
allocate the treatment they receive, so that the doctors bias is factored out
and the results of the research are more reliable. Patients have to understand
that, if their doctor says that she does not know which is the best treatment
for their condition, but that they can participate in a trial to help Wnd out, she
is being a better doctor than the one who wrongly claims absolute knowledge,
despite the (false) security conveyed by the second sort of doctor. If the
doctor then goes on to suggest that her patients participate in a trial to help
discover which treatment is best, the patients have to believe that their doctor
has not suddenly transformed from a genial do-gooder to a sinister re-
searcher in a white coat who from now on will not consider them as human
beings. That is a big step for many people, who shudder at the thought of
human subject research. The staV member at Kings College, London, who
used to serve our lunch when I was running courses called The ethics of
research on humans, said as much. I saw the posters for your course, she
told me. It looks horrible.
Such gut reactions are probably typical. They reXect the perception that
being a good and caring doctor and being a good researcher simultaneously is
not possible. They also indicate a lack of understanding of the need for
3 The value of research
medical research to underpin good doctoring. But the Kings College staV
member could be the next patient in hospital eligible for enrolment in a
research project. If her autonomy is to be respected, she is going to need to be
fairly persuaded that the research is worth doing, and can be done in an
ethical way.
This book is not written for the staV member, but the concerns that lie
behind her shudder are where the ethics of research on humans are located.
So although some might argue that medical research needs no justiWcation, I
would like to consider Wrst of all the reasons why medical research has to be
undertaken, and then go on to the question of how it can be done ethically.
What is the value of research?
Those who support the need for research argue that no newtreatment should
be oVered outside the context of a controlled trial, so that the treatments
eVectiveness and eYcacy can be measured ab initio, not only for the sake of
the patient receiving it but also for future patients. This view entails that
patients should by custom and practice also be experimental subjects. Few
would rather be a guinea pig than the recipient of tried and tested treatment,
but the proponents of clinical trials point out that even an established
treatment which is given outside the context of a trial is more often than not
untested and unproven. Hence patients receiving it are, de facto if not de jure,
guinea pigs in a uncontrolled trial whose outcomes are not being measured
consistently.
Baum (1986) explained that the surgeon who carries out mastectomy for
early breast cancer for 10 years and then switches to lumpectomy for the next
10 years, because custom and practice have changed, is in fact conducting a
research project involving haphazard allocation. The surgeons patients are
not receiving the best known treatment, they are receiving the treatment that
she thinks is best on the basis of unreliable data. Because the surgeon is acting
solely in what she believes to be the best interests of her patients, and her
intentions towards them are unmixed with the desire to gain knowledge
which will not be of direct beneWt to them, the ethics of her behaviour have
not, in the past, been openly questioned.
On the other hand, the surgeon might undertake a properly designed
controlled trial, in which half her patients were chosen randomly (a technical
word which means that patients treatment is determined by the equivalent of
tossing a coin rather than anyones deliberate choice) to receive mastectomy
and half to receive lumpectomy. If she then compared the treatment out-
comes for each group, she would produce objectively convincing evidence for
which of the two treatments is the better, instead of continuing in uncertainty
or, worse still, thinking she knew which was better when she did not.
Whilst few would deny the need to demonstrate greater certainty than
subjective observation allows, the attitude of the surgeon to the individual
patients in the trial might nevertheless then be open to rebuke, because
arguably she is not doing her best for each one, but treating each as a means to
her own end: that of answering the question of whether mastectomy or
lumpectomy is the better treatment for early breast cancer. However, Baum
would argue that in fact the surgeon is doing the best for each of her patients
because she is oVering a 50% chance of receiving the best treatment, which-
ever it is. Since she does not know for certain which is better, it would be
wrong for her to oVer treatment in any other way than randomly (under-
stood in its technical sense). If she switched to lumpectomy, without Wnding
out for certain whether it was the better treatment, she might be exposing her
patients to unknown risks and uncertain beneWts. Hence, for the surgeon,
putting her patients into such a trial means they are better oV than if she
oVered just one of the two treatments.
The treatment for childhood leukaemia and the side eVects of diethylstyl-
boesterol are examples of why research is so important. For some decades in
the UK, researchinto treatments for childhood leukaemia has been organized
nationally, so that most children presenting with leukaemia will (with their
parents consent) be randomly allocated either the latest proven treatment or
the latest novel and experimental treatment. As a result of this collaborative
and carefully orchestrated activity, treatment for leukaemia has moved from
being mostly unsuccessful to being 50% successful, in that mortality from the
condition has dropped from 100% to 50%. The story of treatment develop-
ment for leukaemia is an astounding success, from the point of view of its
consequences. It has not been the result of some single, radical discovery like
that of penicillin, which brought about a complete shift in the paradigm of
treatment of those diseases which antibiotics can cure. Rather, it took, and
continues to take, painstakingly small steps that have inched forward over
years. Had such careful research not been conducted, successful treatments
for leukaemia may have been developed, but, as Baum would argue, the
discoveries would have been haphazard, subject to chance, and unlikely to
have been received into the generality of practice so readily, since their
eYcacy would not be proven in the eyes of others.
Diethylstyboesterol was Wrst synthesized in 1938 and administered to
several million pregnant women to prevent spontaneous abortion and pre-
mature delivery (Dodds et al., 1938). Much later, in the 1970s, a group of
doctors noticed a connection between the mothers exposure to the drug and
the likelihood of her child, if female, contracting vaginal cancer (Noller and
Fish, 1974). Now there are indications that the cancer risk is not so great as
the 1970s observations indicated (Hatch et al., 1998). If the drug had been
introduced by means of a properly conducted randomized controlled trial,
with follow up of the patients in the trial, the questions about cancer would
5 The limitations of research
have been discovered by the 1940s or 1950s at the latest. Because the drug was
administered in an uncontrolled fashion, risks were only noticed much later
on by chance observation and linking of factors in the women concerned, and
even then they were not properly quantiWed.
What are the limitations of research?
Black warns against a total surrender to the scepticism that is theoretically
demanded by the scientist (1998). Some aspects of medicine are inappropri-
ate for the researcher to question. Black argues that although there is a
growing minority of situations in which scientiWc medicine makes all the
diVerence between a cure and a disaster . . . we should keep in mind that they
are still, however important, only a small part of the whole province of
medicine, and that we have a duty of care as well as a duty of cure . He
quotes Robert Platt:
However far the science of medicine and of surgery advances, the art of medicine will
remain: the art of Wrst identifying the patients problem (which is something more
than merely diagnosing his disease) and the art of applying the science to the need of
the individual patient. (Platt, 1972, p. 27)
Imagine a patient who presents his medical problemto his doctor, and asks
her to do what is best for him. This doctor is persuaded of the reliability and
desirability of randomized controlled trials. Armed with data, she explains
the treatment options for her patients condition, their probable success rate
and the statistical likelihood of unwanted side eVects. However, the patient,
instead of considering in a scientiWc way the relative merits of the options in
front of him, repeats his question, asking the doctor to do what she believes is
best for him. The patient trusts in and wants his doctors feelings and
intuitions about himself in particular, not numbers and risk percentages for
people who are like but not him. The doctor, unless she has trained herself to
be mechanical in her response to her patients, is unlikely not to have such
feelings and intuitions. Now the proponent of the randomized controlled
trial will argue that unless knowledge is generated from the controlled
situation of a trial, it is unreliable and should be ignored. This viewunderesti-
mates what Black would call the ability and tendency of clinicians to work on
a balance of probabilities, to which I would add their capacity for responding
in detail to precisely what is in front of them. There are all sorts of diVerences
between patients that may have important implications for the kinds of
treatments they should receive. These diVerences may or may not be com-
mensurable. The controlled trial may be capable of counting them in, or it
may need to factor them out by means, inter alia, of randomization and
sample size. Other kinds of research methods may be employed to measure
the less quantiWable aspects of the medical encounter, and these will be
discussed elsewhere. But the fact remains that nothing can replicate precisely
this encounter. Hence, the doctor in the example should bring her intelli-
gence, intuition and experience to the particular patient in front of her.
It would be reasonable to combine both the science of medicine and the art
of it as interwoven activities (Black, 1998). Doctors need to know what
research has discovered and then apply it to individual cases as they think Wt,
trusting in their intuitive responses, but also testing them against good
empirical data to ensure they have not wandered oV track in any way. Good
doctors would be expected to know the relevant data gathered from well-
designed, reliable research, and also to respond fully to the individual pa-
tients in front of them.
Of course, if doctors are to use data which are reliable, the research which
generates them needs to be scientiWcally sound. There are a number of
problems with this, and numerous possible solutions, because diVerent
research questions need appropriate methods for answering them. Finding
out whether a particular drug lowers patients blood pressure will require one
kind of method; Wnding out how the drug aVects patients quality of life will
require another. Debates persist over the extent to which diVerent research
designs are capable of producing reliable answers. Some designs rely heavily
on subjective interpretation, for example. This sort of method is regarded
with suspicion by many doctors. The issue is signiWcant because the purpose
of research is to produce data that are widely accepted. If a doctor goes to the
trouble of conducting a research project whose results are disregarded be-
cause nobody thinks the study was properly designed, her endeavour, and all
the resources it used, is wasted.
What is the right way to treat human research participants?
Dealing with such scientiWcally thorny questions as what constitutes reliable
evidence would be diYcult enough if the research subjects were, say, plants.
But they are human beings, often patients, who must be treated with due
respect. I have already moved away from the theoretically simple perspective
of Baum and others, in which, provided there is genuine uncertainty about
diVerent treatments, the most appropriate treatment for a patient would be
that which was allocated to him within a clinical trial. For, if doctors proper
care of their patients consists of more than a machine-like response to
symptoms, then more will be required of them than just to be in a state of
uncertainty about treatments before routinely oVering them as part of a
randomized controlled trial. Their duty of care includes taking into account
each individual patient, towards whom, as I have suggested, doctors will have
quite speciWc responses, inappropriate to blind allocation of treatment. It is
7 The right way to treat human research participants
the suggestion that the controlled trial removes this latter aspect of the
doctors duty of care that gives people in general a feeling of discomfort
about, or even fear of, the notion of being research participants or of doctors
being researchers, using their patients as their guinea pigs. It can be argued
that by comparison with what is on oVer outside the context of a clinical trial,
being a participant in a trial is by far the preferable option (Chalmers, 1994).
But that does not make the fact of becoming a statistic rather than being
recognized as an individual any more desirable.
The problem is made more acute by the observation that the scientiWc
validity of the trial, that is, its capacity to yield reliable results, depends upon a
mechanical method of allocating treatments. In recognition of precisely this
natural tendency of doctors to have speciWc responses to individual patients,
the process of random allocation of treatments within trials has been made
doctor-proof . It used to be the case that doctors participating in research
would be given a series of sealed envelopes, and asked to open themin turn as
each patient presented for treatment within the trial. The envelopes would
contain instructions as to which treatment should be oVered, and the envel-
opes were randomized. Doctors would, notoriously, open the next envelope
for the next patient, decide that it was not the best treatment for this patient,
and keep on opening envelopes until the treatment instructions inside
accorded with their idea of what this particular patient needed. Most ran-
domized controlled trials now use a computer to allocate treatment, and
participating doctors will be required to telephone a study centre as each
patient in the trial presents, to be told by a computer which is the treatment
to be allocated. For one of the purposes of random allocation is to erase all
possibility of the results of the research being arrived at by virtue of a doctors
bias. But which patient is going to be happy for a computer to choose his
treatment rather than a doctor? A qualitative study of parents of children in
the national extra corporeal membrane oxygenation (ECMO) study found
this attitude amongst parents interviewed:
They [the parents] said they could not understand how a decision could be based on
only a name, but also had problems when they considered the possibility that the
computer had the information about their daughters case. To them, she clearly
needed treatment other than the conventional care she was receiving and with her
details to hand it was incomprehensible that she was not given ECMO. They felt the
computer had made the wrong decision. (Snowdon et al., 1997)
It could be argued that the patients feelings about how treatments are
allocated in a trial can be set aside if the doctors are satisWed that the method
works in the best interests of their patients. If a doctor is for any reason
unhappy for her patient to be randomly allocated a treatment then the
patient should not be invited to enter the trial. Hence, the problems with
blind allocation might only be found within patient perceptions and not in
reality. The problem would therefore be solved by better communication
with patients. That doctors are clear in their own minds about when patients
are eligible for either treatment and therefore for the trial is belied by the
envelope problem. However, this too may be an issue to do with doctors
misperceptions in that they too may believe wrongly in the eYcacy of a
particular treatment, or its beneWt for a particular patient.
These observations hang upon certain assumptions about the objective
reality of treatment success and the part that attitudes and beliefs of both
doctors and patients have to play in the healing process. Randomized con-
trolled trials have been useful because they can factor out both doctors and
patients attitudes and beliefs. Because of random allocation, it does not
matter whether the patient or the doctor believes the treatment will work. It
will work objectively, or not at all. The question then remains whether
successful medicine is dependent only on objective factors, or whether beliefs
play their part as well, or, even, whether the willingness to accept objective
factors depends upon prior beliefs. For example, is it necessary that a patient
should have faith in his doctor for the doctor to be able to help him at all? At a
minimum level the answer must be yes, because why otherwise would her
patient seek and then act upon her advice? Suppose, then, that research was
able to establish objectively that some medicines were better than others, but
that in the process of establishing that fact in the process of research
patients lost their faith in the medical profession because of the constant use
of computers to allocate treatment?
Arguably, then, there is a dilemma: while the results of research are needed
because they should form part of any good doctors decision-making process,
the means by which that information is obtained presents ethical problems.
Although it is possible to show that the treatments oVered within a controlled
trial are the best available, what is lost is the individualized care, which was
described earlier as an essential complement to the scientiWcally supported
knowledge of appropriate treatments.
How can research participants views be respected?
The case is made more complicated if it is considered that the ideal, individ-
ualized encounter between patients and doctors ought to include due con-
sideration, not only of each patients clinical condition, but also of his own
thoughts, feelings, concerns and beliefs about both the condition and the
proposed treatment. Random treatment allocation ensures that patient bias
as well as doctor bias is factored out as a cause of the results of the research,
since by the same coin-tossing chance the patients views on which is the
treatment of choice are discounted.
Perhaps, however, instead of complicating the matter further, the intro-
9 Three areas of ethical concern
duction of the wishes and needs of individual patients is a way out of this
impasse between the scientiWc need to generate unbiased data, and the
doctors duty to respond sensitively to each individual. Suppose that patients
actively wish to participate in research? Provided that their wish is based
upon adequate information and is freely made, it could be argued that
doctors then have no right to override them. The matter becomes one of
patients right to self-determination; doctors need only ensure that the
patients have suYcient information and then leave them to make their own
decisions. If the problem of the ethical acceptability of research is shifted on
to the shoulders of the research subject, I have, arguably, retained the
individually tailored response, since it is precisely an individual choice
whether or not to take part.
However, by suggesting that the ethical question is answered by passing it
on to the patient, the issue has been fudged. All the patient is agreeing to is
treatment allocation by chance, which, after all, is all that the doctor was
accepting. The individual choice of a particular treatment is still not being
made. Moreover, there is always the danger that patients will agree, not
because they think it is the right choice, but because their doctor has asked
them, and they believe their doctor has their best interests at heart. The extent
to which the patients consent could be called an informed choice is, there-
fore, brought into question. Again, we are faced with the possibility that
conducting research on humans is unethical, this time on the grounds that
the right of individuals to self-determination is not being honoured. Argu-
ably, then, research has not only brought into question the doctors duty of
care, but also the validity of patient consent.
Three areas of ethical concern in research: science, best
interests and autonomy
In the foregoing discussion of the sorts of moral concerns that issue from
research on humans, three broad areas have been looked at. The Wrst is in the
realm of what is necessary or valuable research, in terms both of its goals and
of whether its methods will achieve the goals reliably. The second area
concerns the doctors moral obligation to do the best for her patients,
understanding that to be not merely producing a textbook response but
tailoring treatments to particular individuals. The third area lies in the realm
of considering the wishes and needs of patients and potential research
participants, who have rationality to be respected and beneWts to gain or lose
by virtue of their participation in research, about which only they may know.
These are the three areas on which the ethics of a research project will
hinge. Therefore, if we want to be able to tell whether any given research
project is ethical or not, we need to be able to analyse it by reference to these
concerns. In order to do that rigorously, we have to understand why these are
the relevant areas of concern, so that we can have conWdence in our
conclusions. For the same reason, we also need to know how to conduct the
analysis.
The three areas of concern about research on humans relate to three
important moral theories. The concern with the scientiWc validity of a
research project is related to morality which is teleological (from the Greek
telos, meaning end, purpose). That is to say, it is predicated on the notion
that an action is justiWed by its results. The second concern, which is about
the duty to care for a research subjects welfare, is related to one kind of
deontological moral theory (from the Greek deon -ontos, meaning to be
necessary), which states that the moral agent owes people duties. The third
area, concerned with the research subjects autonomy, is related to another
kind of deontological theory, which states that the moral agent must respect
peoples right to self-determination.
Now some would propose that only one of these theories is necessary to
make moral decisions. It could be argued, for example, that all that is
necessary to decide whether an action is morally right is that its consequences
maximize happiness. Or one might decide that all that is needed is to obey
some sound principles of duty. Or one might think that, as long as my actions
accord with the wishes of those most aVected by those actions, then they are
morally acceptable. But we have already discovered that the moral questions
which arise in the context of researchcall upon all three of these theories. This
means that they all have to be taken seriously, and we have to be able to make
them work in combination, or know what to do if they come into conXict.
I therefore propose to investigate each of the moral approaches in turn as
they have been described by diVerent moral philosophers. We will Wnd that if
we try to make any one of thema comprehensive moral ideal, it will fall short.
Taken in combination with the two others, however, a more robust way of
making moral decisions emerges. But in so doing, it is necessary to distil from
each approach that in it which is useful for the purpose of analysing the ethics
of research on humans. I have given my distillation of the three approaches
titles which are borrowed from Ronald Dworkin (1977, pp. 16877), who
used them in respect of political philosophy. The titles were successfully
adapted to moral philosophy and medical ethics by Sophie Botros, who also
applied them to the ethics of research on humans (Botros, 1992).
(i) The goal-based approach, also known as consequentialism. Moral
theories which work from this perspective judge an actions moral
worth by its predicted or actual outcome. The goal at which the action is
aimed provides the moral determinant for the action itself. No con-
sideration is given to the question of whether the contents of the action
are morally right. Absolute consequentialismis inappropriate for ethical
research, but concern with the validity and importance of the research
11 Three areas of ethical concern
question, that is, its goals, is appropriate. I will call this part of the ethical
enquiry goal-based morality.
(ii) The duty-based deontological approach. Moral theories based on
notions of duty proVer rules of conduct related not to the goals at which
actions are directed, but to the nature of the actions themselves. An
example of such rules is the Ten Commandments (Exodus 20.1V). The
moral justiWcation for peoples actions is determined by the extent to
which they adhere to the rules. Duty-based deontological theories have
weaknesses, but they provide a counter-balance to goal-based argu-
ments that can sometimes be compelling. In considering the ethics of
research, the duty-based issues arise in relation to the way the research is
conducted, rather than what it is trying to achieve. Rules such as not
harming the research participant will apply. These sorts of consider-
ations can be called duty-based morality.
(iii) Right-based deontological moral thinking. There are many theories of
rights, and the language of rights has become common in medical ethics
as a counterbalance to overly paternalistic notions of the doctors duty.
For our purposes, the right with which those who conduct research on
humans should be concerned is the right to self-determination, or
autonomy. Its practical application is to ensure that research partici-
pants consent is sought and their conWdentiality respected. This ap-
proach can be called right-based morality.
This series of moral approaches will provide three distinct perspectives from
which to consider the ethics of research, which we will need to be able to
combine, for, in practice, every research project will need to be investigated in
the light of all three. The combination should produce a systematic approach
to deciding what makes research ethical. It should also produce conXict, in
that it is not always possible to give equal weight to all three. What to decide
in those circumstances is the principal challenge of the ethics of medical
research on humans, and forms the basis for the idea that some research is
simply unethical. Separating out the diVerent parts of the ethical assessment
should make it easier to establish which moral concerns need to take preced-
ence in individual research proposals.
Identifying these moral perspectives, and showing how to use them in the
context of research, will hopefully be of assistance in your own moral
thinking about research. The issue to be addressed is, ultimately, whether any
research project is ethical or not. But if it is to be addressed in a sound and
reasonable way, then a prior question has to be answered, which is: What
must I take into account if I am to come to a proper decision about whether
this research is ethical? This book should provide you with the answers to the
latter question, so that you can answer the former yourself.
aaaaaaa
13
2
Goal-based morality: scientic rigour in
research
The foundations of goal-based thinking
Research should aim to maximize health and minimize harm
A major issue to consider, if we want to know whether a research project
involving human subjects is ethical or not, is what the research is aiming to
achieve. If its outcomes are useful to medicine in that they maximize health
whilst minimizing any harm, and the research is capable of achieving those
outcomes, then, it could be argued, it is morally obligatory that it is under-
taken.
The philosophical idea that underpins this argument is that the proper aim
of humankind is to maximize happiness. The ideas classical formulation is
known as utilitarianism. This is an approach to decision-making that has a
place by right in our consideration of research, since researchis about seeking
and Wnding goals, or outcomes. Hence it is important to understand the
implications of trying to maximize happiness. In what follows we will
consider this approach to moral thinking, and discuss its origins, strengths
and weaknesses. In so doing we will see in which ways the approach can be
genuinely useful to the ethics of research, and in which ways detrimental.
Utilitarianisms strengths and weaknesses
Utilitarianism is a method of moral thinking which considers an actions
consequences to be the determinant of whether the action itself is right or
not. This method was Wrst applied to practical problems by Jeremy Bentham
(17481832). The foundation principle that Bentham described, and upon
which he built his moral philosophy, is this:
1. Nature has placed mankind under the governance of two sovereign masters, pain
and pleasure. It is for them alone to point out what we ought to do, as well as to
determine what we shall do. On the one hand the standard of right and wrong, on
the other the chain of causes and eVects, are fastened to their throne. They govern
us in all we do, in all we say, in all we think: every eVort we can make to throw oV
our subjection, will serve but to demonstrate and conWrm it. In words a man may
pretend to abjure their empire: but in reality he will remain subject to it all the
14 Goal-based morality
while. The principle of utility recognises this subjection, and assumes it for the
foundation of that system, the object of which is to rear the fabric of felicity by the
hands of reason and of law. Systems which attempt to question it, deal in sounds
instead of senses, in caprice instead of reason, in darkness instead of light.
But enough of such metaphor and declamation: it is not by such means that moral
science is to be improved.
2. The principle of utility is the foundation of the present work: it will be proper
therefore at the outset to give an explicit and determinate account of what is meant
by it. By the principle of utility is meant that principle which approves or
disapproves of every action whatsoever, according to the tendency which it appears
to have to augment or diminish the happiness of the party whose interest is in
question: or, what is the same thing in other words, to promote or to oppose that
happiness. I say of every action whatsoever; and therefore not only of every action
of a private individual, but of every measure of government.
3. By utility is meant that property in any object, whereby it tends to produce beneWt,
advantage, pleasure, good, or happiness, (all this in the present case comes to the
same thing) or (what comes again to the same thing) to prevent the happening of
mischief, pain, evil, or unhappiness to the party whose interest is considered: if that
party be the community in general, then the happiness of the community: if a
particular individual, then the happiness of that individual. (Bentham, 1789,
pp. 3334)
For Bentham, to determine whether actions are morally right, that is, whether
they honour the principle of utility, is a matter of simple mathematics. We
have to identify and count the number of individuals aVected by our actions,
calculate for each person whether the action would increase or decrease
his/her happiness, count up howmany individuals would be made happy and
how many made unhappy, and then act accordingly. Bentham listed what
needed to be taken into account when calculating levels of pleasure and pain,
namely, their intensity, duration, certainty, propinquity, fecundity and pu-
rity. Bentham also stated that calculations should be based upon each person
counting as one and no more than one. There were no gradations of
importance for diVerent people.
Utilitarianism is appealing: it requires no religious faith, no explicit moral
code, no general agreement of what rules must be obeyed, and only one law of
nature to accept, which is that everyone wants to experience pleasure and
avoid pain. Utilitarianismhas been called a minimumcommitment morality,
for to adopt it we need learn no detailed and demanding moral teaching. We
must want to act responsibly, but that is all. In an age of secular rationalism,
utilitarianism provides a common morality which does not ask us to ac-
knowledge God.
It could be argued that, as utilitarianism only considers consequences
and not the contents of actions, it could allow us to commit morally
abhorrent acts, providing that our actions maximize happiness overall. In
response to this, utilitarian thinkers after Bentham have suggested that a
15 The foundations
more sophisticated calculation be performed. Rather than working out which
is the morally right action in each case, we have to work out which are the
morally desirable rules which we should apply to our actions, according to
the utilitarian criterion that they must lead to the greatest happiness for the
greatest number. For the purpose of distinguishing the two forms, Benthams
utilitarianism has been called act utilitarianism, and this more sophisticated
form has been called rule utilitarianism.
The trouble with the compromise rule utilitarianism oVers is that by
means of it the simple and neutral grounds for calculation are lost. For rule
utilitarianism requires, Wrst, the identiWcation of those rules which we believe
would lead to the greatest happiness, and, then, to show that they do, in fact,
lead to the greatest happiness. The rules we identify will depend, inevitably,
upon what we believe a priori to be right. Bentham claimed that no beliefs of
any sort are required for moral decisions to be made: all we have to do is to
calculate whether our proposed actions will maximize happiness. The cir-
cumstances are given and real. For rule utilitarianism, the circumstances are
assumed and hypothetical. The realm into which we would reach for rules to
set the conditions for the greatest happiness of the greatest number will be the
realmof our own ideas and values, there by virtue of (at least) our upbringing
and personal experiences. How else did we form the idea that some actions
justiWed by act utilitarianism are morally abhorrent? Rules may, of course, be
selected as a result of observation and accumulation of actual consequences
of relevantly similar actions over a period of time. But such a selection is still
subject to the vagaries of an individuals interpretation. The utilitarian
thinker, R.M. Hare, argues that the Love Commandment (Love the Lord thy
God with all thy heart, and with all thy soul, and with all thy mind . . . and
love thy neighbour as thyself Matthew 22.37,39) is the rule which should
govern peoples actions, because it will maximize happiness (Hare, 1986).
Other rule utilitarians identify other rules. Bernard Williams (1993, p. 82V)
argues that utilitarianism, by adopting rules, has retreated into the realm of
received tradition and values of which it claimed it had no need.
The cases where a proposed action is justiWed by act utilitarianism, but is
also morally abhorrent, are the cases to which rule utilitarianism may be
applied. These are, Williams points out, the very cases that should be
supported in order to be a consistent utilitarian, since they are already
justiWed on utilitarian grounds. To take a step back into the realm of rules in
order to avoid a morally abhorrent act is inconsistent, since one is supposed
to be unaVected by the violation of rules which come from tradition or
religion. If an action will maximize happiness, then one should be prepared
to perform it, whatever it involves.
If this method of thinking were to be applied to medical research on
humans, we would Wnd that a research project has utilitarian moral justiWca-
tion if (i) it is likely to achieve results which can be translated into better
medical care for patients; and (ii) the number of researchparticipants is fewer
than the number of future beneWciaries; or (iii) that the beneWt to future
patients is of more magnitude than the harm to the research participants, if
the number of participants is greater. Chapter One mentioned the leukaemia
trials which have led to signiWcant improvements in healthcare. Benthams
criterion of increasing pleasure (good health) or decreasing pain (curing
disease or desisting from harmful treatments) is clearly met in the example.
The argument could be taken a step further to postulate that the need for
such research was and is paramount, and not morally optional. Without it
and research like it, medicine would remain in the dark ages, probably doing
more harm than good.
However, if that argument is accepted with no deontological qualiWcation,
we are also implicitly accepting that it is positively ethical to harm some
people for the sake of others. For if an action is performed that does this, we
can claim not only that the action was morally right, but that it was morally
required, because it maximized happiness. Not performing the action would
have minimized happiness. In the case of the leukaemia trials, utilitarianism
on its own would require those trials to be conducted even if some of the
children in the trials were disbeneWted, provided that the overall goal was a
success, in that treatment for childhood leukaemia in general improved.
In their defence at the Nuremberg Trials, the Nazi doctors used utilitarian
arguments to justify their experiments:
Out of 724 experimental persons, 154 died [in the typhus experiments]. But these 154
deaths have to be compared with the 15,000 who died of typhus every day in the
camps for Soviet prisoners of war, and the innumerable deaths from typhus among
the civilian population of the occupied eastern territories and the German troops.
This enormous number of deaths led to the absolute necessity of having eVective
vaccines against typhus in suYcient quantity. (From the Nazi doctors defence
speeches at the Nuremberg Trials, 1947 p. 1017)
The quotation shows what actions utilitarian-style arguments can be em-
ployed to justify. There is nothing in utilitarianism as such to prevent acts
which our intuition (or conscience or moral values) would regard as abhor-
rent. So utilitarianism would not forbid research which harmed its partici-
pants, providing its outcome led to more beneWt overall.
This is the point where a utilitarian thinker would argue that rule utilitar-
ianism should be adopted. Setting aside Williams logical quibbles with this
version of the theory for the moment, a utilitarian rule which goes like this
could be devised: research will be conducted which leads to better medical
care, but only if it does not harm the research participants. For this rule to
work under utilitarianism, it has to be justiWed not by any intuitive notion of
its being simply right, but according to the principle of utility; that is, its
application has itself to maximize happiness. Now, it could be argued that
17 The foundations
no-one would be happy to beneWt from research which has harmed others,
and on that basis this rule has utilitarian justiWcation. But how do we know
that that is the case? In theory, people might agree with the principle, but in
practice, if someone has a very ill close relative, would she reject a treatment
which might help her relative, but which had also been developed through
harming someone else? If she was happy to accept such a treatment, the rule
has failed its utilitarian test, and our moral thinking collapses back into act
utilitarianism. If the research which harmed some was already justiWed
according to act utilitarianism, and we are thoroughgoing utilitarians, not
harbouring latent deontological notions of what is or is not morally accept-
able, we should simply agree to the harm.
Williams made a telling observation of the inherent debasing pull of
utilitarian philosophy when he said that, being by nature pre-emptive, it is
subject to a kind of Greshams Law (an observation in economics that money
of a lower intrinsic value has a tendency to circulate more freely than money
of a higher intrinsic and equal nominal value). As Williams puts it:
There is a simple reason for this: a utilitarian must always be justiWed in doing the least
bad thing which is necessary to prevent the worst thing that would otherwise happen
in the circumstances . . . and what he is thus justiWed in doing may often be
something which, taken in itself, is fairly nasty. (Williams, 1993, p. 96)
An example of this kind of thinking is what happens if someone decides to
have an abortion. No one would argue that to have an abortion is in and of
itself a good action. It could only be found good in the sense that, in that
persons mind anyway, not to have an abortion would lead to worse conse-
quences than having an abortion. Similarly, and with apologies for the
somewhat insensitive comparison, the Nazi doctors did not try to show that
what they had done to their experimental subjects was intrinsically good, but
rather they tried to show that under the circumstances their research had to
happen, because the consequences of it not happening were much worse.
Goal-based moral thinking applied to medical research
Utilitarianism has the advantage of simplicity in that it only considers the
outcome of actions to matter morally. Inherent in this approach is the
disadvantage that it can justify or even require harm to some. Hence,
utilitarianism fails to oVer a suYcient moral basis for considering whether
any given research project involving human subjects is morally acceptable or
not. For research on humans to be ethical, not only must the consequences of
the research be considered, but what procedures and risks the research
subjects will be exposed to, and whether the subjects will want to be so used
must also be thought about. Utilitarianism cannot help us with these con-
cerns.
However, to dismiss utilitarianism would be to lose an important aspect of
moral thinking, which is that we should consider the consequences of our
actions before embarking on them. Although consequences should not be
regarded as our Wnal, and certainly not our only, justiWcation for deciding
whether the action itself is ethical, neither should we ignore them. This is as
true for research projects as it is for any other action. Hence, part of our
consideration of the ethics of research on humans is to identify what the goal
of the research is, and to judge whether it is morally right, not in order to
justify bad actions to achieve it, but because it is wrong to commence a
research project which has morally questionable aims. This, then, is how we
will use utilitarianism. It is a distinctly diVerent version of the usual theory,
because it will accept an intuitive notion, that some consequences are simply
wrong or, to put it another way, that some sorts of happiness are better than
others, and because it will not accept that maximizing happiness is the moral
trump card. I call this version goal-based morality.
When this approach to moral thinking is brought to the ethics of research
on humans, we are, in eVect, asking what use the research results will be to
medicine. This concern with the medical or scientiWc value of research also
obliges us to think about two other related issues. These are, Wrst, that the
research should be designed in such a way that the goals are achieved
satisfactorily, and second, that when they are achieved, they should be
disseminated. In a sense, these are technical questions, not ethical ones. But
they have serious ethical implications. Badly designed research gives rise to
inaccurate results which mislead and are potentially dangerous to patients.
Unpublished results, however accurate, help no one except the researcher.
The rest of this chapter will address these three areas of moral concern which
fall under the heading of goal-based issues.
The application of goal-based thinking
The goals of research
Immediate goals and general directions
In considering what value a research proposal will be to medicine, we Wnd we
can judge it in terms of both its immediate goals and the general direction of
which the individual project is a part. The immediate goal will present itself as
a question, such as: whether or not treatment A works safely and eYciently;
whether treatment B restores the balance of all the bodys functions; whether
situation C is the cause of condition D; whether patients prefer E to F.
These research questions are asked within larger contexts than their im-
mediate aims: within their own therapeutic Weld; within the Weld of medicine
and healthcare generally; within the direction of human endeavour; and
within the environment of the whole world. At each level, questions can be
19 The applications
asked about whether the research will have beneWcial eVects. For example, at
the level of the environment, the over-use of pharmaceuticals has had eVects
which potentially harm everyone, such as the presence of oestrogen in the
water supplies, or the growing resistance to antibiotics in populations.
Judging the morality of the goals of research proposals
These general directions to which individual projects contribute raise moral
questions, but they cannot be addressed at the level of the individual research
project, since it is collectively and cumulatively that they cause problems. If I
sit on a researchethics committee and consider a particular research project, I
can express qualms about the developments in medicine of which it is a part,
but I cannot really justify withholding ethical approval on those grounds
alone, for it would seem unfair to load the responsibility for the general
problem on to that one project. Moreover, judging the moral desirability of
the goal of any research proposal presents problems of what criteria it should
be judged against. The goals of maximizing health care whilst minimizing
harm can be retained as the ultimate aim, and some research will be clearly
moving in that direction. Other research proposals may not be capable of
being linked in the same way, and may seem trivial by comparison. They may
not be. Even with appropriate expertise amongst their membership, research
ethics committees may not understand the contribution of a particular
research proposal within the context of the development of the discipline of
which it is a part. Hence, it is diYcult for research ethics committees to
disapprove of a research project on the grounds of its goals alone. The
researcher will be in the best position to judge this, though even he may not
necessarily know where a particular investigation will lead. He is, neverthe-
less, morally obliged to consider the ethics of what he is proposing to do, in
general as well as in local terms.
Although it is hard to see how a research proposal could be found
unethical on its goals alone (unless the goal is manifestly bad, such as
developing biological weapons of mass destruction, in which case it will not
be conducted in the context of medicine), it still matters that some view is
taken of the degree of importance of a proposals goals. This is partly because
the relative importance will make a diVerence to the view that is taken of how
the research will be carried out. If it is trivial research, few concessions should
be made to allow it to happen; if essential, then other moral imperatives may
or may not be overruled. I discuss this balancing of diVerent moral claims in
much more detail when I come to consider duty- and right-based consider-
ations. It is also the case that establishing the importance of the research goal
matters in and of itself from a goal-based perspective, for the straightforward
reason that an agent should not commence an action without Wrst consider-
ing the morality of its consequences.
Researchers motive a guide: problems with paying researchers to
conduct research
Where the goal of a research proposal is not obviously related to immediate
improvements in healthcare it could be argued that the guide to its accepta-
bility lies in the motives of the researcher. He may wish to conduct his
research from genuine scientiWc curiosity, or he may have a strong sense that
a particular avenue of investigation will ultimately lead to real developments
in his Weld of medicine, or he may have an overriding concern with the
welfare of the patients in his discipline and want to help, albeit in a small way,
to improve their lot. All these motives could be said to be pure.
Suppose, however, that he is being paid to conduct the research? Tradi-
tionally, research ethics committees have been suspicious of payments going
into researchers pockets (as distinct, say, from departmental funds). Phar-
maceutical company fees for routine clinical research of drugs in develop-
ment will often be paid directly to the researcher. Sometimes these fees are
high. They are said to cover the researchers time and expenses for running
the trial. These costs are often considerable, and it is important that funds for
ordinary treatments do not get diverted into pharmaceutical company re-
search. But in the context of trying to decide what is important or necessary
research, and relying to some extent on the researchers motive as a guide for
this decision, the fact that he is being paid inevitably raises a question about
how important the research really is. (There are also questions about the
justice of the researcher being paid when the research subjects are not. In the
UK, for example, research subjects are only paid if the research is non-
therapeutic. Payment to research subjects is discussed in Chapter Eight.)
Even if research funds are not going to individuals, or are not from
commercial sources, there are still factors which can interfere with the
researchers motives for doing research. When research funds are applied for,
the applicant will tailor his research proposal to Wt the wishes of the grant-
giving body, in order to secure the grant. Hence, the researcher is not as free
as we might hope to act only for the best and purest motives. He has to do the
best he can in the circumstances, and those who judge him need to recognize
that.
Why researchers should always aim for the sky
The best treatment for any condition is the one that cures it completely and
does so with no harmful side eVects, either to the individual patient or to the
wider environment. Every researcher probably has such a grand hope lurking
somewhere at the back of his mind as he embarks on the long and laborious
process of designing, carrying out and disseminating his research pro-
gramme. It is a good hope to sustain and work towards, because it avoids the
tacit assumption that medicine and other treatments must of necessity do
harm, if good is also to be enjoyed. It might seem odd, to a lay person at any
21 The applications
rate, that such a notion exists, but it does. It has been argued that that which
could do no harmcannot, therefore and for that reason, eVect a cure either. A
researcher would not want to claim total success while further improvements
are still possible. Hence, the research should always aim at absolute good, not
because it will always achieve it, but because it is likely to get closer to it than
if a less absolute goal is sought. If we aim at the centre of a target our arrows
are more likely to hit it than if we only aim at the target.
Methods of research
Once a research goal has been identiWed we then have to work out how to
achieve it in a reliable way. DiVerent research questions require diVerent
research designs to answer them, if they are to be answered in a way that is
widely recognized and accepted. This is a matter of good science, but it is also
a matter of ethics, because research which is improperly designed is both a
waste of resources and a source of potential harm if its conclusions are wrong
and they are believed to be right. The choice of research method requires
technical knowledge. There is room here to describe only brieXy the basic
principles of diVerent research methods. I hope, however, to convey a sense
of how each method tries to answer questions, in order to show why diVerent
questions need diVerent ways of answering them. The three major types of
research methods are: randomized controlled trials; observational research;
and qualitative research. Some methods provide more objective means of
measuring outcomes; others are more subjective.
Randomized controlled trials
The randomized controlled trial is the most well-recognized method of
arriving at an objective answer to some sorts of research question. People
often wrongly assign causes to events because what they think they see is not
what is actually happening. The fact that a doctor gives a treatment to a
person with condition X, and condition X goes away after she has had the
treatment, does not mean that the treatment cures condition X. What the
doctor saw in that case may be due to the treatment really working, or it may
be due to a diVerent cause, such as the condition of the patient, faulty
measurements of the outcome, or chance. The randomized controlled trial
seeks to avoid wrongly attributing treatment success to the medicine being
tested, by factoring out anything that might give the wrong impression of the
medicine. The most robustly designed controlled trial is the one which has
factored out most causes of wrong attribution. This includes both patients
and doctors attitudes towards and beliefs about the eYcacy of a treatment.
The fundamental property of a controlled trial is that it makes a compari-
son, so that one group of patients receiving a newtreatment will be compared
to another group being given an existing, often standard treatment, a placebo
(dummy treatment) or no treatment at all. The patients receiving the stan-
dard treatment, the placebo or no treatment are known as the control group,
while those receiving the new treatment are the experimental group. Some-
times the groups change over after a deWned period, so that the control group
of patients become the experimental group and vice versa (crossover trial
design). Sometimes the trial compares diVerent doses of the same medicine.
If possible, the treatments are masked, so that neither the patient nor the
doctor knows which treatment the patient is receiving.
Patients will be randomly allocated to either of the two groups; no-one
chooses who goes into which group. Randomization is eVected by a method
equivalent to tossing a coin, usually by computer. There are eligibility criteria
for joining trials, not only for the safety of the patients, but also for ensuring
that the groups are appropriate. Bias in the group membership can give rise
to an apparent diVerence between the two treatments which would then be
wrongly attributed to the treatment.
Assessment or measurement bias is avoided by ensuring that the measure-
ments of results are objective, for example the patients serum cholesterol or
haemoglobin being tested by a machine in a laboratory. If, on the other hand,
the assessment has to be based upon the clinicians judgement of the success
of a treatment, or on the patients self-report, then the assessments them-
selves are masked. This is achieved by ensuring either that the clinician doing
the judging is not the clinician who administered the treatments, or, if it is
patient self-report, the clinician assessing the reports does not know which
patient said what.
To control the chances of errors in assigning causes, the sample size of
research subjects must be suYciently large, and the study must have an
acceptable p value and adequate power. The p value is the probability that a
diVerence at least as large as that seen in the data would occur by chance, if
the true diVerence between the treatments is zero (MacRae, 1996). Claiming
a treatment diVerence when there is none is called a Type One, or Alpha,
error. Conventionally, the hypothesis that the diVerence between the treat-
ments is zero is rejected if the p value is lower than 5%. When the p value is
lower than 5% the diVerence is referred to as being statistically signiWcant.
The power of a study is its probability of not missing a real diVerence
between the treatments. This is known as a Type Two, or Beta, error. The
power of the study needs to be high, traditionally 80% or above. A power of
80% means that there is an 80% chance that the study will produce a
statistically signiWcant diVerence if the treatments are diVerent from each
other.
MacRae notes that whilst the calculation of the p value and the statistical
power of the study requires mathematical expertise, the clinically relevant
diVerence between the two treatments being compared, that is, the goal of the
research, is a matter of value judgement and it is this which will determine the
23 The applications
sample size needed. The greater the diVerence looked for between two
treatments, the fewer the number of patients that are needed in the sample
size. The way that the question is approached is to state the null hypothesis
(the true diVerence between the two treatments is zero) and then try to
disprove it, which is much more scientiWcally sound than devising a positive
hypothesis and trying to prove it (Popper, 1959).
Observational or epidemiological research
Most observational or epidemiological research involves observing what is
happening without interfering, unlike experimental research, such as the
controlled trial which intervenes and measures the eVect of the intervention.
Examples of exceptions to the rule of not interfering are studies of the eVects
of placing Xuoride in the water supply to one area, or studying the eVect of
screening programmes for cancer in one area, where the Xuoridization or the
screening programme take place in order to be studied. Observational re-
search can be either descriptive or analytic. Descriptive studies only describe
phenomena without trying to identify cause and eVect. Some descriptive
studies are qualitative (see below), whilst others are quantitative, such as the
large-scale surveys conducted by government agencies to measure the atti-
tudes and habits of the population. Analytic studies do seek connections
betweencause and eVect, namely, exposure and outcome, and they are always
quantitative. There are three types of analytic observational research
methods:
(a) Cohort studies compare two groups, one of which is exposed to a risk
and one of which is not, and follow the two groups prospectively,
calculating the ratio of rates of disease. Importantly, each group has
chosen the exposure or lack of it (for example, smoking).
(b) Case-control studies start with the outcome and measure prior expo-
sure. In eVect, they seek to reconstruct a cohort study after the event, for
example to discover the causes of rare disease events.
(c) Cross-sectional studies collect data on individuals at one point in time,
for example the eVect on people of a large-scale disaster. These studies
cannot informthe researcher whether the observed eVects came before or
after the exposure, so they are not very reliable. (Thorogood, 1996)
Qualitative research
Qualitative research takes an interpretative approach to subject matter,
studying things in their natural settings and attempting to make sense of or
interpret phenomena in terms of the meanings that people bring to them.
Qualitative research recognizes that there are many diVerent ways of making
sense of the world, and it is, therefore, concerned with discovering the
meanings seen by the researched, rather than imposing the interpretation
of the researchers (Jones, 1995). Because of this concern not to impose
interpretative criteria on phenomena, the research methods are numerous
and some are very loose indeed. A few of the diVerent methods used by
qualitative researchers are given here.
(a) Observational methods involve the researcher systematically watching
people and events to Wnd out about behaviours and interactions in
normal settings (Mays and Pope, 1995).
(b) Qualitative interviews take many forms. There are three basic types: the
structured interview, with a structured questionnaire; the semi-struc-
tured interview, with open-ended questions; and the in-depth interview,
involving one or two issues covered in great detail, with the questions
based on what the interviewee says (Britten, 1995).
(c) Consensus methods provide another means of synthesizing information
than systematic review (see below), using a wider range of information
and harnessing the insights of appropriate experts to enable decisions to
be made. One commonly used consensus method is the Delphi process.
This proceeds in a series of rounds, in which a series of expert opinions is
generated and scored by experts until a consensus is reached. The other
commonly used method is the nominal group technique, otherwise
known as the expert panel (Jones and Hunter, 1995).
(d) Focus groups are a form of group interview that uses the communica-
tion between the members of the group to generate research data. Hence,
group interaction is speciWcally part of the method. The method is useful
for exploring peoples knowledge and experience, and for examining not
only what people think but why and how (Kitzinger, 1995).
(e) Case study evaluation involves studying the implementation of a policy
or other intervention empirically, by evaluating case studies of the
implementation. Qualitative or quantitative methods may be used. The
tradition of using case studies in medicine is a long one; this is an attempt
to systematize and extend that use (Keen and Packwood, 1995).
When to use which method
For a long time there have been fundamental diVerences of opinion about
which research methods are reliable and valid for the discovery of knowledge
which can then be believed by others. The randomized controlled trial,
because of its ability to measure objectively, has taken top place in the
hierarchy of evidence, and has provided the gold standard against which
other research methods are measured. Hence, the less able a study is to prove
something objectively, the less reliable and useful it is. We can consider how
valid this claim is when we look at disseminating the results of research.
However, I would point out here that to neglect other research methods is
inappropriate, since some aspects of medicine are not amenable to ran-
domized controlled trials. For example, if we wanted to know what the
physically harmful eVects of smoking were, we could not ethically answer that
25 The applications
question by a randomized controlled trial, because it would involve random-
ly allocating a group of people to smoke, which we know is bad for them. Or,
if we wanted to know about patients perceptions of a treatment we had
administered to them, a randomized controlled trial would not tell us,
because there is no prior hypothesis to determine the comparative groups to
which patients should be allocated.
The following comments are helpful in showing the usefulness of the
diVerent research methods:
Randomized controlled trials carried out in specialised units by expert care givers,
designed to determine whether an intervention does more good than harm under
ideal conditions, cannot tell us how experimental treatments will fare in general use,
nor can they identify rare side-eVects. Non-experimental epidemiology can Wll that
gap. Similarly, because the theoretical concerns about the confounding of treatment
with prognosis have been repeatedly conWrmed in empirical studies..., non-experi-
mental epidemiology cannot reliably distinguish false positive from true positive
conclusions about eYcacy. Randomized trials minimize the possibility of such error.
And neither randomized trials nor non-experimental epidemiology are the best
source of data on individuals values and experiences in health care; qualitative
research is essential. (Sackett and Wennberg, 1997)
Disseminating the results of research
Suppose that a researcher has found an ethical goal for his research, and
designed and executed the research in a scientiWcally valid manner. He still
needs to think about how to tell others about his results, and persuade them
to alter their clinical practice on the basis of what he have discovered that is
new. The diYculties that stand in his way arise from the unwillingness of
individuals to change, from publication bias and from bad indexing. There is,
in addition, a more fundamental question, which resonates with the dis-
cussion in Chapter One. Is evidence-based medicine, that is, medicine which
uses evidence from research, the best kind of medicine there is? Evidence-
based medicine has a signiWcant advantage over medicine practised solely
according to hunches and subjective interpretations of past experience, but
there are questions to be asked both about the underlying premise and about
the practical utility of evidence-based medicine.
The following paragraphs will consider: the challenge of encouraging
clinicians to put knowledge into practice; the challenge of bringing together
all the information available from research; and questions about the basic
idea of evidence-based medicine.
Encouraging clinicians to put knowledge into practice
There are numerous reasons for the diYculty of persuading clinicians to put
the results of research into practice (Lancet, 1993). It may be that results are
viewed with caution, with a desire to wait and see what happens when others
try using them. Many doctors will have had the experience of hearing about
promising new agents, backed by scientiWc studies, which have in the end
been found wanting. There may be a fear of potential adverse side eVects not
yet shown in the research, or confusion over indications and contra-indica-
tions. A clinician may have had a previous bad experience with the treatment
in question which overrides, in his mind, the less tangible beneWts shown by
the research. Cost eVectiveness and quality of life evaluations in speciWc areas
may make the application of the new treatment diYcult to use by an
individual doctor. This may be true if, for example, the new treatment
requires more knowledge or skill, and more equipment or personnel. The
clinician may simply not believe the research Wndings, either because he does
not share the assumptive basis of the researcher, or because he thinks the
measurements in the research are not rigorous enough.
The Lancet suggests ways in which doctors might be encouraged to put
advances in medical knowledge into practice. The primary way it suggests is
through education in journals and at postgraduate educational and scientiWc
meetings. However, a cautionary tale is oVered by researchers who conducted
a randomized controlled trial of educational visits to enhance the use of
systematic reviews in 25 obstetric units (Wyatt et al., 1998). They found that
their visits made little diVerence to the practice in these units.
The Lancet also suggests that meta-analyses are useful because the cumu-
lative evidence can be conveyed more quickly to a busy clinician. Doctors
who are well-acquainted with the pathophysiology and natural history of a
particular disorder are also more likely to put advances in knowledge into
practice. If the doctor has actually participated as an investigator in research
he will be more persuaded of a treatments eYcacy, having seen it for himself.
The lesson from this, the Lancet states, is that trials need to be planned with
the implementation of their results as part of the project.
Collecting the results of research
In 1991 Philippa Easterbrook and others discovered considerable publication
bias in clinical research (Easterbrook et al., 1991). Their survey found that
studies showing a statistically signiWcant diVerence between the groups of
patients receiving diVerent treatments were more likely to be published than
those Wnding no diVerence. Studies with signiWcant results were also likely to
lead to a greater number of publications and presentations, and to be
published in journals with a high citation impact factor (likelihood of being
quoted elsewhere). The bias was seen in all types of research, but especially in
observational studies. This means that if a researcher had completed some
research he would be far more likely to succeed in having it published if his
results showed a diVerence between treatments. If his results were negative,
that is, they showed that the experimental treatment was not, in fact, any
27 The applications
diVerent from its standard comparator, he would be less successful. The
implication is that important negative results have not seen the light of day.
Easterbrook concludes: These Wndings suggest that conclusions based only
on a review of published data should be interpreted cautiously, especially for
observational studies. Improved strategies are needed to identify the results
of unpublished as well as published studies.
In 1997 between 50 and 100 reputable medical journals called an amnesty
for unpublished trials. Investigators with unreported trial data were invited
to register their trials by completing an unreported trial registration form.
The aim, said Roberts (1998), was to tap the silent subterranean pool of
unpublished research and by bringing these data to the surface, to increase
the power of systematic reviews and reduce the eVects of publication bias. At
the time of writing, which was a year after the amnesty had been issued,
Roberts reported that only 165 such trials had been registered. There was
some involvement by the pharmaceutical industry, but government agencies
on both sides of the Atlantic did not give their support, arguing that
commercial conWdentiality could not allow the pharmaceutical industry to
tell all. Consequently, the amnesty was a Xop.
Herxheimer had earlier anticipated that commercial and political secrecy
would be an obstacle to the dissemination of research (Herxheimer, 1993).
He stated that pharmaceutical companies, which might want to lose results
which did not demonstrate beneWts of their treatments, as well as other
sponsors of research, who might want to suppress results which were not
politically desirable, thought that because they had funded and initiated the
research they owned the results. This ignores the rights of research subjects,
who participate because they think they are being useful to others. It also,
argues Herxheimer, ignores the rights of investigators, who take overall
moral responsibility for the proper conduct of their research, and the rights
of the institutions where they do their work, which support them.
Another problemwith relying on published researchis that if we were to go
to an index to Wnd the publications we want, we would miss an important
source of knowledge: letters published in journals following the publication
of research results (Bhopal and Tonks, 1994). The letters which follow
original papers are essential for correcting scientiWc error and contributing to
peer review. An example of their importance is the study of the Bristol Cancer
Help Centre. The Bristol Cancer Help Centre, as its name implies, oVers
treatment to people with cancer. Its methods involve no conventional or
allopathic medicine; only a strict diet and much support at the spiritual and
psychological level. The published study concluded that the Centres treat-
ments did not work. But there were a number of faults in the design of the
research, which numerous subsequent letters pointed out. One of the letters
was from the funding agency which withdrew its support from the main
conclusions of the research. Gaining a proper sense of the results of the
research would entail reading the letters as well as the original paper. Similar-
ly, a paper published in the British Medical Journal in 1992 suggested that
elective delivery by caesarean section of the breech fetus at term was far safer
than spontaneous vaginal delivery. Nineteen letters robustly challenged the
papers conclusions; subsequent articles cited the paper, not the comments
(Bhopal and Tonks, 1994). Internet indexing has improved this record.
Nevertheless, systematic reviews, as their name implies, are genuine at-
tempts to review research evidence which has taken steps to avoid bias and
the play of chance (Chalmers, personal communication, 2000). Despite the
challenges to the success of such projects, their contribution to the develop-
ment of sound medical practice should be celebrated and supported.
The limitations of evidence-based medicine
Kerridge has evaluated the ideas underlying evidence-based medicine in a
thoughtful article (Kerridge et al., 1998), and his observations on evidence-
based medicine are resonant both of goal-based morality and of research in
general. Kerridge agrees that evidence-based medicine is founded on a strong
ethical and clinical ideal: it allows the best evaluated methods of healthcare to
be identiWed and helps doctors and patients make more informed decisions.
However, it is also an expression of consequentialism, which argues that an
action can be assessed by measuring its consequences. Many important
outcomes cannot be deWned or measured. Whose interests should be taken
into account when determining outcomes? Should systematic reviews in-
clude research projects which were conducted unethically? And what if the
conclusions that a systematic review comes to are themselves intuitively
morally abhorrent?
In medicine, says Kerridge, intangible values such as quality of life and
justice are as important as quantiWable values such as mortality or cost. The
systematic reviews and meta-analyses, on which evidence-based medicine
relies, traditionally only recognize randomized controlled trials as substantial
enough to merit a place in their review. This is understandable if the aim of
systematic reviews is to provide objective knowledge, but this aim is inappro-
priate for medicine as a whole. Moreover, the problem of conclusions which
are intuitively abhorrent is a real one. For example, a systematic review of
treatments of the elderly may conclude that it is not economical to treat
anyone over the age of 65. Arguably, however, no-one should be chary of
discovering even unpalatable facts: the ethical questions arise over when and
how to use them.
Using or not using the results of research which was conducted unethically
presents problems in at least two respects for meta-analysis, and, indeed, for
the amnesty called by editors of journals in 1997. It may be that a journal
refused to publish a research paper because the way the research was conduc-
ted was unethical. This issue is discussed in Chapter Eight, citing a journal
29 The applications
which faced exactly this dilemma of either turning down an article because
the research it was based upon was unethical by 1990s standards, and losing
the useful results of the work, or publishing it and giving a voice and platform
to unethical practices, thereby encouraging them. Meta-analyses are only one
step removed fromthe publications they use. Should the Wndings of unethical
research be included in the reviews? Some answers to the question are looked
at in Chapter Eight when the particular case is discussed.
In evidence-based medicine, randomized controlled trials are at the top of
the hierarchy of reliable evidence. Controlled trials require the doctors to be
happy for their patients to go into any of the diVerent treatment groups for
them to be able, ethically, to enrol their patients. Where the treatments being
compared are roughly equivalent the doctor can do this. For interventions
which have great beneWts or which carry major risks, the doctor may well not
be happy to allow his patients to be denied beneWts or be exposed to risks.
Colebrooke faced precisely this dilemma with penicillin in the 1930s (see
Chapter Six). In such cases doctors may prefer to accrue evidence by report-
ing clinical observations or by using historical controls. For evidence-based
medicine and the systematic reviews which support it, this relegates the
evidence. The conXict between the need for scientiWcally reliable data and the
need not to harmresearch participants is discussed again in Chapter Seven, in
relation to placebo-controlled trials.
There are diYculties with using the evidence properly. Techniques for
accurate application of trial results are needed. The crude application of
results may, on average, do more good than harm, but they may none the less
disadvantage some patients. If evidence is used as a basis for allocating funds,
this will mean that only those treatments which have been tested by ran-
domized controlled trials will be funded, and this means that only quantiW-
able treatments will be considered, or that treatments not amenable to being
measured will none the less be subjected to inappropriate evaluations based
on quantity rather than quality.
Kerridge implies that those who seek to practise evidence-based medicine
are robotic and insensitive, which is unfair. Proponents of evidence-based
medicine have to counter what they see as the real danger of doctors
recommending treatments on the basis of experience and intuition when
these can be so misleading. As I have said, I would not want to discount such
practice, but I recognize the importance of balancing intuition with evidence,
and hence the need for doctors to remain open to changing their practice.
Whilst some research methods, mainly qualitative, try to measure systemati-
cally values other than quantiWable ones, in the hierarchy of evidence ran-
domized controlled trials come top because they are capable of providing
objective proof. Evidence-based medicine, then, relies most heavily on ran-
domized controlled trials, and so do systematic reviews. This means that if
anyone wants to use the results of research, and do so using systematic
reviews, as the Lancet suggested, they will be relying mainly on quantiWable
evidence, and many important aspects of medicine cannot be measured in
this way.
It would be a mistake to be disheartened by the extensive list of diYculties
with gathering and disseminating evidence. It is better to Wnd out systemati-
cally whether a treatment works, even if it is diYcult to persuade others of the
validity of the work. And although evidence-based medicine should not be
made into an absolute good, it is important for doctors to use evidence from
research as well as their own experience and intuition when treating patients.
Summary and concluding remarks
If goal-based morality is used as part of the overall consideration of what
makes a research project ethical, it has an important function, for it requires
that the consequences of research projects be considered as matters of ethical
concern. In other words, it brings what might be called the purely scientiWc
considerations into the moral arena where, in my view, they certainly belong.
The scientiWc questions relate to the goals of the research, the way the
research is designed, and the dissemination of its results. Hence, a researcher
needs to consider these issues in a morally responsible way. The most diYcult
issue is that of the goal of research. It is perfectly possible to give a scientiWc
justiWcation for conducting a research project in terms of its immediate goals
of Wnding out, for example, whether a particular drug does in fact treat a
particular symptom, or whether it is possible to grow human tissue from the
stem cells of cloned embryos. It is more diYcult to judge whether such
questions should be asked in the Wrst place. My own suggestions were that a
researcher should consider the eVects of his research at every level where it
can have an eVect, and that he should aim the research at the goal of Wnding
treatments which are harmless to all.
Once a researcher has decided on the acceptability of a research projects
goal, he then has to consider which of many research methods is the best to
answer his research question. DiVerent research questions require diVerent
research designs. Finally, he needs to disseminate the results. This too is a
practical matter necessary for ethical reasons. Not to publish results takes
away the overall goal-based justiWcation for the research, which is for the
greater good. Nevertheless, there are some real problems with dissemination
and with the way results are used.
Once we have addressed these ethical considerations as best we can, the
second third of the ethics of research on humans has to be addressed. This is
related to what the research subjects will have to undergo for the research to
take place. This ethical concern arises within the context of duty-based
thinking, which is the subject of the next chapter.
31
3
Duty-based morality: acting in the research
subjects best interests
The foundations of duty-based thinking
From goals to duties
Utilitarianism asks no more of us as moral agents than that we consider the
outcome of our actions. Provided that its outcome maximizes happiness, an
action is morally justiWed. We are not asked to give an account of the
rightness of the action itself. There are no principles to refer to for that sort of
judgement. The last chapter found that the goals of research are important
issues morally, but it is not enough to identify a good goal for research and
leave the moral consideration at that, even if the related issues of scientiWc
method and the dissemination of results are satisfactorily included. The
conduct of the research itself has to be subjected to moral constraints. The
example quoted from the Nuremberg Trials helped to show why. If some-
ones focus as a researcher is only on the outcome of her research, she is going
to miss the ethical implications of what she has to do to arrive at that result.
Goal-based morality cannot help us think about the ethics of the research
procedures, because its focus is always on the future. We need a duty-based
deontological moral approach rather than a teleological one to assist our
analysis at this stage.
Duty-based deontological morality concerns itself speciWcally with the
contents of an action rather than with its results. It considers actions in the
light of explicit moral principles, which have to be identiWed and adhered to.
In the context of research, some research which is aimed at highly desirable
outcomes, involving signiWcant steps forward in medical knowledge, may
nevertheless be wrong according to a duty-based deontological approach
because of what the research involves. For example, an experiment may
necessitate harm to its subjects in order to gain scientiWcally sound results.
The utilitarian may ask why the harmto a fewshould weigh suYciently in the
balance to prevent research whose goals will be signiWcantly useful. The
duty-based moralist needs to be able to say why it should, if indeed it should.
In short, duty-based deontological morality should be able to give some
substance to the intuitive notion that some actions are, like the Nazi doctors
experimentation, simply wrong. This chapter will consider the strengths
32 Duty-based morality
and weaknesses of the approach, and then see how to use it as part of a
moral analysis of research on humans. Two formulations of duty-based
ethics will be described and their practical use for our purposes will be
investigated. The Wrst formulation to be considered is the tradition of natural
law ethics. The second is that of Immanuel Kant (17241804), who proposed
an imperative moral law which was categorical, hence obligatory.
Natural law ethics
The moral principles of natural law ethics are derived from observable facts
about nature. The best way to explain howthis works is to give an example. In
his treatise on fornication and marriage, Thomas Aquinas (1225/61274)
derived a comprehensive package of moral teaching on the subject using
arguments from natural law. One of his derived moral rules is that the
marriage between a man and a woman should be for life. This is how he
argues the point:
Besides, there is in men a certain natural solicitude to know their oVspring. This is
necessary for this reason: the child requires the fathers direction for a long time. So,
whenever there are obstacles to the ascertaining of oVspring they are opposed to the
natural instinct of the human species. But, if a husband could put away his wife, or a
wife her husband, and have sexual relations with another person, certitude as to
oVspring would be precluded, for the wife would be united Wrst with one man and
later with another. So, it is contrary to the natural instinct of the human species for a
wife to be separated from her husband. And thus, the union of male and female in the
human species must be not only lasting, but also unbroken. (Aquinas, c. 1264 p. 147)
By identifying a natural desire in a man to know his oVspring, Aquinas Wnds
a basis for the moral rule that a man and a woman should stay together for
life. Where, however, does Aquinas derive his natural laws from? If they come
from simple, empirical observation, his laws could be contested on their own
grounds. For we could observe that a number of men do not, in fact, stay with
their oVspring, providing evidence of a lack of interest in them. If Aquinas
gains his laws of nature from his perception of what happens anyway, for
every perception he had from which he derived a moral rule, one could think
of a counter perception to undermine it. For his laws of nature, if they are to
give rise to rules which we are supposed to obey (and can, therefore, choose
not to obey), must be breakable. If he only identiWed laws of nature which we
could not dispute, such as the law of gravity, he would have no need to devise
a moral law from it, since the law of gravity is unbreakable. You must fall
through the air at great speed if you jump out of a window is not a moral law,
it is a statement of fact, and the word must means will, not ought to.
We might rejoin, however, that although the law of gravity is unbreakable,
it is possible to act as if it was. That is to say, you could decide to jump out of a
33 The foundations
window. The law of gravity does not remove your ability to do that. What it
does do is to ensure that if you do break the law, you will suVer the
consequences. But for a short time, say while you were sitting on the window
sill just before jumping, you might think that, in fact, you will not fall to the
ground but will Xy. Suppose that the desire of a man to know his oVspring is
equivalent to the law of gravity, and that his decision to leave his children is
equivalent to jumping out of a window. He might not experience the
consequences of his action to begin with, but he (and his children) might
suVer later, just as the person falling to the ground will suVer after a (short)
period has elapsed.
On this basis, we should be able to derive moral laws from observable facts
of nature. We would need to Wnd laws which are as unquestionable as the law
of gravity, and then draw up rules which are the equivalent of telling people
not to jump out of windows. It is worth noting that the reason for obeying
these kinds of moral laws would be prudential rather than virtuous. It would
not be in our interests to disobey them, if the consequences of doing so were
the equivalent of hitting the ground after falling from a great height. In fact, it
could be argued that these sorts of rules are not really moral at all. For
although it is unintelligent to jump out of a window, it could hardly be called
immoral to do so, or positively moral to refrain. Aquinas observed that the
proper function of semen is to eVect procreation, and for that reason it
should not be used for any other purpose, such as pleasure alone. No one
would dispute that semen fertilizes eggs, but many people do not think it
follows from that fact that it is immoral to have sexual intercourse using
contraception. Would it be immoral to break the law of gravity by Xying an
aeroplane?
What such a morality requires us to have is a belief that things have proper
functions and it is right for us to ensure that those proper functions are
fulWlled. If we want to know what is intrinsically right and wrong, we need to
look at the function of the thing we are dealing with. If an action helps it fulWl
its function, then the action is right. This sort of contention sounds trivial
when it is applied to an eye, or a kidney, for although those things have
deWnable functions it is hardly a matter of morality to ensure that they fulWl
them. But in the context of medicine the approach becomes more meaning-
ful. A doctor, by virtue of her professional title, has a particular role to play in
relation to patients, that is, she owes them a duty of care, usually expressed as
the duty to act in their best interests. This is her function, and it is unconten-
tious to assert that she ought to fulWl her function. She could choose not to,
but this would be called wrong, again, uncontentiously. Moreover, the sense
in which it is simply right for a doctor to act in her patients best interests is
more akin to the lawof gravity being simply right than, say, not taking illegal
drugs is simply right. That drug-taking is wrong is subject to the circum-
stances in which it happens. Smoking marijuana in order to heighten sensory
experiences might be wrong. Smoking it to alleviate the symptoms of
multiple sclerosis might not be. At least the diVerent circumstances merit a
discussion. The law of gravity cannot be contested in the same way, and
neither can the doctors duty to act in her patients best interests. This is
because the title doctor deWnes the function which the person who takes that
title has to perform. If she did not want to act in her patients best interests,
she would not call herself a doctor. Now, if she also wishes to conduct
research upon her patients, and hence add the title researcher to the title
doctor, a question arises as to whether it can still be taken as read that she
must act in her patients best interests. I would argue that such a function
remains so long as the title doctor remains, which it will do if there is
someone present calling himself a patient. It is even questionable whether
doctors can shed their role at any time (Walsh, 1998). The role of researcher
has its functions and associated duties, and a case would need to be made for
the doctors duty to care to override the researchers duty to extend medical
knowledge. The case can be made by considering the situation if the re-
searchers role was allowed to override that of the doctor. This would mean
that in a research project, were a participant to show signs of distress, but his
coming out of the trial would jeopardize the trial, the doctor/researcher
ought to keep him in the trial. On the contrary, the Wrst consideration should
be the well-being of the patient/participant, and the doctors duty to care
takes precedence.
Kants categorical imperative
A diVerent, though related, basis for duty-based deontological morality can
be found in the philosophy of Immanuel Kant. Kant argued that pure reason,
by its very existence, warranted moral behaviour. Human beings, having this
faculty of reason, are therefore bound by the moral law. His argument, as
described in his Fundamental Principles of the Metaphysic of Morals (1785),
develops as follows:
Reason, given to rational beings by Nature, is capable of recognizing
absolute truth, stripped of all empirical qualiWcations. Reason is, therefore,
autonomous, needing no external assistance to help it recognize what is true.
Reason directs the will, which, if obedient, overrides personal inclinations.
The will, properly directed, expresses itself as the motive of duty. Actions
which have moral worth are those which are performed solely from the
motive of duty and not for any other reason whatsoever. The motive of duty
is determined by the maxim or law according to which the action is per-
formed, which our reason recognizes and accepts. The absolute maxim or law
which governs our actions is called the categorical imperative.
The categorical imperative, if it is to be shown to be absolutely required,
will not have its truth demonstrated by experience or by reference to the
35 The foundations
empirical world. Indeed, we will not be able to say that we have ever seen an
act which has moral worth, because we will not know if personal inclination
had any part in its motivation. The example of others, reference to God, or
any other medley of external justiWcations or inspirations for moral behav-
iour, cannot provide the basis for our actions, for such justiWcations are
subject to change and are therefore unreliable. The categorical imperative has
to be demonstrated conceptually in order to be recognized by our auton-
omous reason. Kant identiWes the following categorical imperatives: (i) act in
such a way that the maxim governing your action can be a universal law of
nature; and (ii) act in such a way that you can at the same time will that the
maxim governing your action can be a universal law of nature. These, Kant
argues, are moral laws which apply in all circumstances and are absolutely
required of reason. It may be that in practice we do not follow the categorical
imperative, but this is not because it is not absolutely required of us as
rational beings. We fail to follow it because our inclinations are at odds with
our reason, and our will, which could override our inclinations, is not
obedient to our reason.
Kant oVers some examples of the application of the categorical imperative.
Here is one of them:
A man reduced to despair by a series of misfortunes feels wearied of life, but is still so
far in possession of his reason that he can ask himself whether it would not be contrary
to his duty to himself to take his own life. Now he inquires whether the maxim of his
action could become a universal law of nature. His maxim is: From self-love I adopt it
as a principle to shorten my life when its longer duration is likely to bring more evil
than satisfaction. It is asked then simply whether this principle founded on self-love
can become a universal law of nature. Now we see at once that a system of nature of
which it should be a law to destroy life by means of the very feeling whose special
nature it is to impel to the improvement of life would contradict itself, and therefore
could not exist as a system of nature; hence that maxim cannot possibly exist as a
universal law of nature, and consequently would be wholly inconsistent with the
supreme principle of duty. (Kant, 1785, p. 50)
The maxim has to be capable of being universalized, but it also has to be
consistent for the moral agent to will it to be universalized. Here is another
example:
A . . . [man], who is in prosperity, while he sees that others have to contend with great
wretchedness and that he could help them, thinks: What concern is it of mine? Let
everyone be as happy as heaven pleases, or as he can make himself; I will take nothing
from him nor even envy him, only I do not wish to contribute anything to his welfare
or his assistance in distress! Now no doubt if such a mode of thinking were a universal
law, the human race might very well subsist, and doubtless even better than in a state
in which everyone talks of sympathy and good-will, or even takes care occasionally to
put it into practice, but on the other side, also cheats when he can, betrays the rights of
men, or otherwise violates them. But although it is possible that a universal law of
nature might exist in accordance with that maxim, it is impossible to will that such a
principle should have the universal validity of a law of nature. For a will which
resolved this would contradict itself, inasmuch as many cases might occur in which
one would have need of the love and sympathy of others, and in which, by such a law
of nature, sprung from his own will, he would deprive himself of all hope of the aid he
desires. (Kant, 1785 p. 512)
Notice that Kant is concerned, not to persuade us that such actions are wrong
for consequentialist reasons, or that they are undesirable, but rather that they
are inconsistent. The categorical imperative does not ask, as rule utilitarian-
ism might ask, How would we like to live in a world where people did not
help each other? Rather it is asking whether it is logically possible for the
person who wants not to help others to will that that principle be followed
universally. Hence, for a rational being whose nature is to behave reasonably,
such actions are categorically forbidden. There can be no argument as to their
being wrong, for they go against the tenets of reason. It is rather like the
absolute requirement on a doctor to act in her patients best interests because
she is a doctor and for no other, external purpose. She does not do it because
it is a good thing to do, but because it is rational, if she calls herself a doctor,
to behave like one. In the same way, reason, by its own nature, accepts the
validity of the categorical imperative, because the categorical imperative is
rational. By virtue of being reasonable, we are obliged to obey it. We could
ignore it, but if we do so, then we can no longer claim to be reasonable
people. Following reason, or acting like a doctor if that person is one, has
good consequences because it makes us behave well, but we do not do it
because it is good, or because we feel altruistic, or because of some intuitive
idea of what is right, but because it is right in the same way that the law of
gravity is right. Here, then, is our robust deontological moral principle to
withstand the force of consequentialist arguments if we need to.
Duty-based moral thinking applied to medical research
The two examples of duty-based deontological ethics examined so far have
yielded some useful fruit. Natural law ethics suVers from the diYculty of
establishing facts of nature. However, it helps a doctor identify where her
duty lies. It can be postulated that, as a doctor, someones natural function is
to act in the best interests of her patients, and that it is as much her natural
function to do that as it is for bodies to fall through the air. Kant postulated
that reason, by its autonomous nature, is bound by the categorical imperative
absolutely, because the categorical imperative is reasonable in and of itself,
without any need for external justiWcation. By her nature as a rational being,
then, a doctor has no need to add a further justiWcation for acting in this way
according to the categorical imperative. Just as a rational being will naturally
37 The foundations
act according to reason, so a doctor will naturally owe a duty of care to her
patients.
From this natural law that a doctor must always act in her patients best
interests, we can derive a duty-based approach to medical research on
humans, in all cases where research is conducted by a doctor, or other
healthcare professional whose relationship with patients has similar priority
over other duties. Although the function of researcher has its own set of
important duties, these do not override those of the doctors function and
must necessarily give way should the need arise.
Although the duty to care of a doctor overrides the duty to do good-quality
research of a researcher, it need not follow that absolutely any risk of harm to
the patient/participant has to be removed before his doctor can rightly ask
him to take part in research. It might in any case be a strong wish of the
patient that he participate in a trial that exposes him to some risk but is of
great potential beneWt to future patients. Such a wish need not be ignored,
but the doctors duty to care entails that any risk should be minimized. A
view must be taken as to what constitutes too great a risk to ask a patient to
face. In coming to such a view, the doctor should not let the researchers zeal
for results overcome her primary concern for her patients well-being. It is a
matter of striking a balance, but the decision should err on the side of the
patients interests, not those of science and society (World Medical Associ-
ation, 1996).
In order to take a duty-based approach to the ethics of research we need to
know what the research is going to involve for the research participants. This
means that we need to complete the goal-based analysis Wrst, for that will tell
us what the research method is and, therefore, what procedures the partici-
pants will have to undergo. For example, a randomized controlled trial will
involve the blind allocation of diVerent treatments to patients. Outcome
measures like urine analysis or blood sugar levels will involve peeing in a
bottle or extra needles, and so on.
As I found in the earlier discussion, the duty-based principle which will
govern the analysis of a research projects procedures is derived from the
doctors duty to care for her patient, that is, to act in his best interests. As I
suggested, this duty is absolute, not contingent upon other factors being in
place. The doctors duty to care expresses itself in oVering to her patient those
treatments which are in his best interests, and avoiding doing anything to
him that is harmful. If, by virtue of being in a research project, he will be
exposed either to treatments which are of no beneWt, or to actions which may
cause him harm, then it is duty-based morality which will call the ethical
acceptability of the research into question. Goal-based morality will be
concerned with the potential good towards which the research is aiming.
Right-based morality will want to know if the participant is willing to be
harmed, and will not consider that it may be inappropriate to ask him in the
Wrst place. Hence duty-based morality plays an essential part in the analysis of
the ethics of research on humans.
It was said earlier that the duty-based principle to be applied to research on
humans is that of beneWting and not harming the research participants.
Applying the principle involves subjecting research procedures to scrutiny
against this yardstick. It is helpful, in doing this, to separate research into
therapeutic and non-therapeutic types, because diVerent levels of beneWt and
harm can be expected and justiWed in each. Therapeutic research is research
which is conducted in the context of clinical care: the participants in the
research project are patients expecting to be treated for their illness as well as
to help the researcher gain knowledge which can be generalized. In this
context some harm to the research subject might be expected, but it would be
experienced as a necessary part of the therapy, and would be outweighed in
the individual research subject by the therapeutic beneWts he may enjoy from
the treatment, be it experimental or standard. For example, treatment for
cancer is associated with risks of some very unpleasant side eVects, but the
risks are outweighed by the therapeutic beneWt, and the two are directly
related. If a research project involved giving chemotherapy to a cancer
suVerer, harm might be anticipated, but balancing beneWt would also be
anticipated, suYcient to render the harm acceptable.
By contrast, non-therapeutic research will oVer no treatment to research
participants; they are simply guinea pigs and can expect no therapeutic
beneWt from being in the research project. Strictly speaking, duty-based
morality demands that participants of non-therapeutic research be exposed
to no risk of harm, since there is no balancing beneWt to them. Weighing risk
of harm to individual research participants against beneWt to future other
people is the function of goal-based moral thinking, not that of duty-based
thinking. But it is impossible to remove absolutely all risk of harm from
non-therapeutic research.
The key to determining whether any given research project is therapeutic
or non-therapeutic is to look at the intention of the researcher. If she wishes
only to generate new knowledge then the research is non-therapeutic. If,
however, she wishes to treat the patient in front of her as well as gain further
knowledge from treating him, then the research is therapeutic. Typically,
therapeutic research involves randomized controlled trials, in which treat-
ments are allocated to groups of patients in a blinded fashion, so as to
discover, with as little bias as possible, which of the treatments being com-
pared is better. Non-therapeutic research can take many forms, from studies
of drugs being administered for the Wrst time to healthy humans, to question-
naire surveys of large populations.
39 The applications
The application of duty-based thinking
Therapeutic research
Therapeutic research involves the intention to treat as well as the intention to
gain knowledge which can be generalized. Hence, it will always take place in a
medical setting, and the participants will be patients as well as experimental
subjects. As was said earlier, therapeutic research usually takes the form of the
randomized controlled trial. There are two ways in which randomized con-
trolled trials can pose moral questions from a duty-based perspective. These
are, Wrst, that they may not oVer patients in the trial the best treatment
available, which it is the doctors duty to provide; and second, there may be
risks of harm which are not directly related to the therapeutic beneWt
anticipated from the treatments received in the trial. On the positive side,
therapeutic research can oVer beneWts to patients in the form of more care
and attention because of more resources available to the research team.
The reason a randomized controlled trial is set up is either because there is
some doubt about a treatment which has hitherto been oVered as a standard
therapy or, more commonly, because a new treatment is being developed
which is thought to be an improvement on any current treatment. The most
objective and reliable way to show that one is better than another, or at least
that one is not the same as another (Popper, 1959), is by gathering together a
group of patients who have presented with the condition which needs
treatment, and randomly allocating to each of them either the standard
treatment or the experimental treatment. Importantly, the allocation of
treatments is done not by preference of doctor or patient, but randomly, by
the equivalent of tossing a coin. In this way, if a diVerence is seen in the
outcome of treatment between the two groups of patients, it can be attributed
to diVerences in the treatments the patients were given, and not, for example,
because the patients, having chosen their treatment, feel more positive about
it and display improvements for that reason. This research method is good
for the goal-based purpose of answering the research question objectively. In
this chapter we have to decide whether it is good by duty-based standards as
well. Therefore, we will consider the design of randomized controlled trials
from the point of view of discovering whether the doctors duty to care Wrst
and foremost for her patient is able to be sustained in this speciWc context.
Denitions of equipoise
Equipoise is the name given to the attitude the doctor should have to the
diVerent treatments in a randomized controlled trial. By the nature of the
trial design, the doctor will have no say in which of the treatment options in
the trial her patients will receive. Froma duty-based perspective, she will need
to be entirely happy that such a situation is in her patients best interests. This
entails that she should have no preference for any of the treatments. She
should genuinely believe that each treatment option in the trial is as good as
any other. I shall call this attitude strong equipoise.
Kennedy and others argue that such a balanced perspective is unlikely:
Adoctor may be involved in a trial which is designed to compare two treatments, both
of which have their supporters. The doctor may have a preference for one form of
treatment, although he has no evidence of a scientiWc nature to justify his preference.
It may be based on experience, hunch, or some other reason; but whatever the reason,
he has a preference. (Kennedy, 1988, p. 219)
However, many would argue that such an attitude is not ethically incompat-
ible with becoming a researcher in a trial. Providing that the doctor remains
uncertain that her preference is correct, she is still in equipoise. The doctor
may have a preference, for after all she is unlikely to be proposing research
into a new treatment if she does not think that, in the end, such a treatment
will be better. I will call this form of equipoise weak equipoise. Some will go
further even than this, and, requiring no uncertainty fromindividual doctors,
will merely require that there is uncertainty or diVerences of opinion across
the relevant clinical community. Thus, individual doctors may strongly
prefer one treatment over another, but as a group there are roughly equal
divisions between them as to which is preferable. I shall call this community
equipoise.
From a goal-based perspective, both weak and community equipoise
require research to be conducted on the treatments in question, so that the
hunches and feelings of doctors can be tested objectively, and certainty
achieved as to which treatment is better. From a duty-based perspective,
however, it may be that only strong equipoise is suYcient for the doctors
duty to care to be honoured. Kennedy argues:
Now, if it is accepted that the therapeutic relationship between doctor and patient is
most successful when the doctor has faith in the treatment he is giving his patient, and
this faith is reXected also in the trust which the patient has in the doctor, it will be clear
that by adopting a form of treatment in which he has less conWdence, the doctor may
be aiding the pursuit of knowledge, but doing so by opting for a less than optimum
relationship with his patient . . . If the doctor genuinely has no preference, then the
problem does not arise, although I would submit that this is a comparatively rare
occurrence. (Kennedy, 1988, p. 219)
Others would say that the duty to care is still maintained in weak and
community equipoise because objectively there is equipoise, even if subjec-
tively there is not. But it is also true to say that the objective equipoise is only
there by virtue of (a collection of ) subjective attitudes, some of which may be
strongly held. It is the individual doctor who has to ask her patients if they
want to participate in research, and it is, therefore, she as an individual who
has to take responsibility for her decision whether or not to do so. Moreover,
41 The applications
her patients will want to know what she really thinks and believes, and they
are likely to be aVected by her preferences. Arguably, then, only strong
equipoise will do, and this, as has been indicated, is rare.
The diVerent views given here about the extent to which the setting of a
clinical trial is capable of allowing a doctor to act in her patients best interests
are crucial to the ethics of clinical research. If, by putting her patients into a
clinical trial, a doctor is acting in their best interests, then the therapeutic
research project should be considered in exactly the same way as ordinary
therapy. In the ordinary therapeutic setting, certain rules apply which are not
currently considered appropriate for research. For example, if a patient is
unable to consent (incompetent) and needs treatment, it is legally up to his
doctor to act in his best interests. If putting a patient into a clinical trial is the
same as treating him in the ordinary way, and a clinical trial is currently
underway involving patients with his condition, then that patient should be
put into the clinical trial, regardless of the fact that he is unable to give his
consent, for it is the doctors duty to enrol him. If, on the other hand, it is not
acting in her patients best interests to enrol them in a trial, the incompetent
patient will not automatically qualify.
I share Kennedys view of the likelihood of a doctor having preferences,
and the eVect of such preferences on her patients. However, it may be
objectively the case that a doctors preference is based on bad evidence. The
doctor may be part of a community in equipoise, that is, where there are
diVering views about which treatment is better. If she is positioned in this
way, she ought in fact to notice the current thinking about the treatment for
which she has a strong preference, educate herself out of her strong prefer-
ence and join the strong equipoise group instead. In that way, both objective-
ly and subjectively her duty to care is honoured if she then asks her patients to
take part in a trial of the treatments.
Can patients still think their doctors are giving them individualized
treatment if they are also participants in their trials?
Strong equipoise indicates that a doctor is convinced of her uncertainty
about the comparative merits of diVerent treatments. There is another aspect
of the doctors duty to care to which I referred in Chapter One, which is not
really covered by the doctor being in strong equipoise. This is the matter of
the patient receiving individualized care from his doctor. Even if the doctor is
impartial as to the treatment choices, does the very act of enrolling her
patient in a trial remove the possibility of her treating him as an individual in
his own right, with speciWc medical needs and peculiarities? There are two
ways in which a doctor can ensure that this loss of care does not occur. The
Wrst is that she should give consideration as to whether taking part in a trial is
in his best interests, not just as regards the diVerent treatments but also as
regards his personal well-being. She should be able to establish this by asking
him whether he would be happy to receive his treatment in the context of a
trial. Second, the doctor should ensure that the research is designed in such a
way that she or her research team can respond to particular medical needs as
they arise. The result of the research should not depend upon having to
ignore individual medical needs. Where needs cannot be anticipated or
woven into the research design, it must be accepted that a patient can
withdraw himself or be withdrawn from a trial should it be necessary.
Additional, non-therapeutic tests demanded by the research protocol
The second way in which therapeutic research can trouble duty-based moral-
ity is that participation in a clinical trial does not merely involve random
allocation to a treatment that a patient might have had anyway or to a novel
treatment. It also involves additional tests and measurements, such as urine
or blood tests, in order to gain the results of the research. Hence, more is
required of the patient in a trial than receiving treatment. It was suggested
earlier that patients do better in clinical trials for precisely this reason: they
receive far more attention, usually from more senior doctors, than they
would if they were normal patients. The additional tests may, however, be
irksome to patients as well as exposing them to further risks. Whichever is
true, and it probably depends much upon the individual patient, these extra
tests further emphasize the fact that being in a trial is not the same as
receiving treatment in the ordinary way.
Where do these two problems with therapeutic research leave the ethical
enquiry? From a duty-based perspective the focus has to return to the doctor
who is conducting the research. The arguments about just how uncertain she
has to be about the diVerent treatments in the trial for her duty of care to
remain intact are not objectively resolvable. It has, I submit, to remain a
matter for each doctors conscience and sense of duty to educate herself. For
the duty-based requirement to be met, she must be satisWed that the patients
whom she asks to enrol into the trial will receive as good a treatment as
possible, whichever arm of the trial they are allocated to. It has to remain the
doctors decision for it is her duty to care which is at stake. Similarly, she
must consider for whom of her patients the additional tests and measure-
ments that come as part of the trial would be burdensome, and for whom
they would not. Again, her duty will be to ensure that those for whom it
would be disadvantageous should be excluded. The somewhat paternalistic
ring to this suggestion will be considered below, when looking at non-
therapeutic research.
The use of placebo
It is worth glancing brieXy (this subject is considered in much more detail in
Chapter Seven) at the use of placebo (pretend medicine, from the Latin
placere, to please) in clinical trials, because it is a good example of where
43 The applications
goal-based and duty-based considerations can clash. The goal-based justiWca-
tion for a placebo armin a trial is considerable. It adds scientiWc weight to the
design and its inclusion as one of the arms of the study often means that
smaller numbers need to be studied, and therefore fewer patients are required
to be research subjects. Hence, on utilitarian grounds, the use of placebos is
justiWed. However, the duty-based requirement to ensure that each individ-
ual in a trial is properly treated means that there is no justiWcation for giving
some patients a worse treatment than others, simply in order for there to be
fewer subjects receiving the inferior treatment. The patient for whom the
doctor has a duty to care is the patient who is in front of her, not the
generality of patients. Placebo is only justiWed as part of a clinical trial if, as
described above, it can be oVered by a doctor to her patient in the certainty
that this treatment is at least as good as the other treatments. So, for
example, where there is no proven, available treatment (World Medical
Association, 1996), a placebo arm is justiWed. Sometimes, confusingly, there
are treatments which have become standard without ever having had their
eYcacy tested and hence are not proven. New treatments could sometimes
justiWably be compared to these and to placebo as well. Each case would need
to be tested on its own merits. But from a duty-based perspective the moral
test is whether each patient in the trial is being treated in his best interests. It
is not to ensure scientiWc validity, nor to keep the patient numbers down.
Non-therapeutic research
Non-therapeutic research consists of research which has no therapeutic
aspect to it for the research participants. The researcher does not intend to
treat them: they are in the true sense her guinea pigs, unlike therapeutic
research participants, who are still also patients being treated. Examples of
non-therapeutic research include Phase I, Wrst time in humans studies to
measure toxicity and pharmacokinetics (where the drug goes and how long it
spends in the human body before being broken down or evacuated); research
into new forms of diagnosis; survey-type research, etc.
Does it matter that the best interests test fails in non-therapeutic
research?
We could apply a strict interpretation of duty-based morality and argue that
non-therapeutic research is not morally acceptable under any circumstances,
because there is no therapeutic beneWt, and there may be some risk. But why
should a person be prevented from being altruistic and exposing himself to
risks for the sake of others, if he wants to? There are plenty of risks which
people willingly take for no good moral reason (like recreational scuba
diving); this is taking a risk for a good reason: the beneWt of others.
If we accept this argument about non-therapeutic research, however, we
have to take a consistent line with the acceptability of the extra tests thera-
peutic research would entail, which are also, taken on their own, non-
therapeutic. To put the decision of which of her patients would be burdened
by the extra tests in the hands of the doctor, is paternalistic. Surely the better
approach to the acceptability of non-therapeutic research, or the extra tests in
therapeutic research, is to ask people if they are willing to expose themselves
to risks, and only enrol them in the research if they are? Should duty-based
morality stand aside for this right-based view, or should it hold its ground?
The right-based view, which has intuitive merit, hangs upon the seeking and
obtaining of consent. The next chapter will consider this aspect in detail. At
this juncture, I shall Wnd suYcient problems with relying on consent as the
single moral determinant of whether research is too risky to support the
argument that doctors should not cease to give careful consideration to what
they ask people to consent to.
Should consent be the deciding factor?
The viewthat consent is what determines whether a non-therapeutic research
project goes ahead may be put as follows: if a person is willing to become a
research subject then he can be enrolled in the research, and if he is not, then
he should be excluded. But this simple view is based upon a number of
potentially false premises.
First, it relies on the belief that an action becomes right if a person wants it
to happen. On its own this is not enough, just as it is not enough that a person
should want someone to stab him to death to make the actual stabbing to
death morally justiWed. That persons views on the matter may be signiWcant,
but they are not all that is needed in the moral consideration. Similarly, a
research project does not become right just by virtue of the research partici-
pants wanting to be subjects.
Second, the view assumes that the elements of consent have been fulWlled,
namely, that adequate information has been successfully imparted to a
competent person whose subsequent agreement is voluntary. In particular,
the voluntary nature of the decision may not be valid. If a doctor asks a
patient to do something, or to take something, the patients natural assump-
tion will be that, if it does not beneWt him, at least it will not harm him, since
the function of doctors is to act in their patients best interests. The trust that
that is so is an essential part of the doctorpatient relationship. Once a
patient starts to believe that a doctor might not have his interests at heart, he
is simply not going to do what she says he needs to do in order to get better,
particularly if the treatment is unpleasant. In fact, he may stop seeing her. He
is likely, therefore, to agree, not because he has made an informed decision,
but because his doctor has asked him.
45 Summary and concluding remarks
Judgement of acceptability of risk remains with the researcher
To ask the participant to be the one who decides whether a risk in a
non-therapeutic research project is too great to justify the research is, frankly,
unfair. The onus remains on the doctor/researcher not to ask her patient to
do something that is bad for him. The question of what makes risks in a
non-therapeutic research project acceptable remains, then, with the re-
searcher. It is up to her only to ask her participants to undergo procedures
which are at an acceptable level of risk. Of course she must ask them before
enrolling them into the research (and in that way give the opportunity to
behave altruistically); that is not being questioned here. What is being
questioned is the extent to which the doctor must play a part in shielding her
patients from making risk assessments about factors that they should simply
not be asked to consider. I would argue that the doctor must decide what
risks are too great to be undertaken by the participants both in the thera-
peutic and the non-therapeutic context.
This chapter has described two foundations of duty-based morality which
give rise to the need to consider research from this perspective. These
foundations were, Wrst, that of natural law ethics, and second, the Kantian
categorical imperatives, absolutely required of all rational beings. Natural law
ethics provides principles of moral behaviour which are predicated on a belief
that everything has a function which it is right for it to fulWl. Kants
categorical imperatives demand that we act in such a way that we can
consistently will that the maxim governing our actions can become a univer-
sal law, and that the maximis capable of becoming a universal law. Both these
foundations point to a duty-based principle in the context of medicine,
which is that the doctor has a duty to care for her patient. If the doctor is also
a researcher, the doctors duty to care overrides the researchers duty to
achieve reliable research results. The duty lies within the function of doctor. If
a doctors patients are to be used as research participants then they should be
given the same level of care and consideration as they would be if they were
patients in the ordinary way.
In therapeutic research the doctor has to be entirely happy to allow her
patients to be randomly allocated to either or any of the treatments in a trial,
including placebo, if there is a placebo arm, before she can rightly ask them to
participate. This raises diYcult questions about the degree of uncertainty she
may have about the comparative eYcacy of the diVerent options.
In non-therapeutic research, the researcher must be satisWed that any risks
to research participants are acceptable for each of the participants, not by
being weighed against beneWts to future patients.
The duty-based perspective on a research proposal provides a balancing set
of moral concerns to those from the goal-based perspective. Duty-based
concerns demand that research is conducted in such a way that research
subjects are not harmed, or their exposure to harm is kept to a minimum in
the case of non-therapeutic research, and that their best interests are served in
therapeutic research. In this way, research subjects can be reassured that they
will each be treated as separately important individuals to whom duties are
owed, and not merely as instruments to a greater goal.
47
4
Right-based morality: respecting the
autonomy of research participants
The foundations of right-based thinking
From duties to rights
The last chapter distilled from deontological ethical theories a duty-based
approach to the ethics of research on humans, which stated that doctors owe
a duty of care to their patients, and must therefore act in their best interests.
This duty has traditionally bound doctors to behave ethically, but it has also
created an ethos of paternalism in which the doctor has been thought to
know best, gives orders, and is unassailably right in his decisions. Another
kind of deontological ethics not yet discussed is based on peoples rights.
Rights can be a useful counterbalance to the duty-based moralists tendency
to be paternalistic.
Ian Kennedy argues: Historically, those who enjoy (in all senses of the
term) power have insisted on the language of duty to express the relationship
betweenthe powerful and the rest (1988, p. 391). People who are responsible
for others have duties towards them. But although the language of duty can
be invoked positively, as a way of determining how that responsibility will be
carried out so that the good of all is served, it can also be used as a means of
retaining power to the detriment of those purportedly being served. Just
because a doctor believes he knows what is best for his patient, and just
because he feels that he is doing his duty by his patient, it does not follow that
what he proposes should happen. Kennedy would argue that unless some
restraint, outside his own control, is put upon the doctor, abuse of his power
can follow. Hence, some intrinsic powers need to be given to the weaker
party, and these can be called her rights. Put in the context of medical
research, this consists of only using a patient as a research subject if she wants
to be so used, not because it is the doctors self-appointed duty to consult his
patients wishes, but because the patient has an inherent right to self-
determination.
Arguably the doctors duty to consult and the patients right to be consul-
ted amount to the same thing. If the doctors duty is suYciently enforceable
so that he has no choice but to consult his patients wishes, what is the
diVerence between that and asserting that the patient has a right to be
48 Right-based morality
consulted? The answer, replies Waldron (1995), is the basis upon which such
a moral claimis made. If the reason we claim that a patients wishes should be
consulted is that it is her right, then we are asserting a right-based morality; if
it is on the basis of the doctors duty, then a duty-based morality. The reason
the distinction is important is that the focus of moral thinking shifts between
the agent of the action and the one most aVected by it. If our morality is
fundamentally duty-based, then the reason that we regard it as wrong to enter
patients into research projects without their consent is that the researcher is
violating his moral integrity by using people as a means to an end. If our
morality is right-based, then it is the denial of the right of a patient to exercise
her autonomy that is the fundamental problem. In the latter situation, it is
not the moral health of the researcher we are concernedwith, but the interests
or rights of the research subject.
Immanuel Kants theory of the autonomy of reason has often been invoked
by right-based moralists to support the existence of inherent rights. Kant
argues that rational beings should always be used as ends in themselves, and
never merely as a means to anothers end. But he argues this because it is
incumbent upon rational beings to do their duty according to the precepts of
reason, as said in the previous chapter. The duty of a rational being with
regard to other rational beings is to treat them as ends in themselves. This is
demanded by practical reason, and not by the construction or discovery of a
right inherent in the one done unto. However, for Kant, it is precisely the
prior existence of autonomous reason in a rational being that demands that
we should treat her as an end in herself. This is an important observation for
the way in which I am going to develop the use of right-based morality, so I
shall return to Kants moral theory later in the chapter.
If we are to assert that rights have a substantial existence which is capable
of practical recognition, we have Wrst to explain what is meant by someone
having a right to something. We then have to show that that right is valid.
Finally we have to show the practical application of the right-based approach
in the context of medical research on human beings.
Denitions of rights
The Hohenfeldian cluster
If the statement: P has a right to X is made, it can be meant in a number of
diVerent ways. For the following explanation I am grateful to Waldron for his
summary of Hohenfelds diVerentiation of meanings of rights (Waldron,
1995).
(i) P has no duty not to do X. For example, an ambulance driver has the
right to ignore road laws, by going through red traYc lights or driving
on the wrong side of the road. Hohenfeld calls this kind of right a
privilege. This right, to avoid the criticism of being selWsh, has to be
49 The foundations
enjoyed only in the context of a particular service someone needs to
perform. It has no merit if it is merely expressed in the childish assertion
that whilst others may be bound by certain conventions and manners, I
on the other hand amexempt, for no other reason than that I want to be.
There can be no intrinsic reason for a person to hold the right which is a
privilege: it must come with the function the person is called upon to
perform. An ambulance driver who switched on his vans siren and
broke road laws just because he was bored of waiting in traYc with the
rest of the population, and not in order to fulWl his duty, would be
abusing his privilege.
If this deWnition is turned around to state that P has no duty to do X,
then it could be argued that patients have a right not to take part in
research. Some, however, argue that no such right exists automatically,
because patients now enjoy the fruits of previous patients participation
in research, and they have a duty to oVer themselves as subjects in order
to give back similar beneWts to those they have received, for the sake of
future patients.
(ii) Everyone has a duty to let P do X. This duty, which is contingent on the
existence of certain rights, ranges from the negative duty not to impede
Ps action, to the positive duty to do what we can to make it possible for
P to achieve X. These rights are called claim-rights. Waldron notes that
some of them are absolute, such as the right not to be imprisoned
without due process, and some are contingent, such as the property
rights over an item which has been purchased. The diYculty with this
deWnition of a right lies in identifying what rights may be asserted in this
way, and in whether someone has passive or active duties in respect of
them. Even the passive duty of allowing us our freedomto act as we wish
is questionable in some cases. If a child, for example, wishes to exercise
her freedom in a way with which her parents are unhappy, do her
parents have to accept that she is a free agent and they have a duty not to
interfere, or does she not have that right? The age of the child will usually
be the signiWcant factor in this kind of decision. Another right that
might be asserted is the right of a woman to have children. If she does
have that right, is the contingent duty of everyone else simply not to
prevent her fromhaving a family, or is it our duty to take steps to ensure
that she can have a family? We might have no diYculty in accepting that
we owe other people the space and freedom to exercise their own wills
over their own lives, if they want. But the woman who is infertile cannot,
in most cases, administer fertility treatment to herself: she needs clinical
assistance.
The question of whether a right demands positive action arises in a
healthcare system such as the National Health Service in the UK, where
treatment is free at the point of delivery and paid for indirectly through
tax on income of the whole population. Decisions have to be made
about what sorts of treatments can validly be oVered and funded within
such a health service (and by extension what sorts of research projects),
and which treatments should be paid for at the point of delivery.
ArtiWcial reproduction is one such area; another is the operation to
change ones sex; another, the removal of tattoos; another, some forms
of plastic surgery. As skills to alter or restore the human condition
proliferate, so does this list of treatments which are not related to saving
lives but are related to well-being. The relevant question seems to be
whether the conditions which give rise to the search for treatment, such
as childlessness, discomfort with ones gender, or being ugly or unsight-
ly in ones own eyes, can really be called disease or illness. But since there
is so much disagreement over whether they are diseases or not, ident-
ifying that as the question does not help very much. It is easier to work
on the basis that everyone has a duty merely to let P do X, though, as in
the case of the errant child, there may be things we would feel it our duty
not to let P do. It is nearly impossible to identify those things which we
have a duty actively to assist P in doing, or even do for him. However,
this sort of laissez faire response to rights is not much use to us if we
cannot obtain what we want without help, such as having a child by
artiWcial means.
(iii) P has the power to do X. This sort of right is gained by physical
strength, or by powers conferred by the state, or other means. It could be
argued that these are the only rights which have any practical use, for
merely proclaiming rights, however logically valid they may be, does not
help if a more powerful entity decides to ignore or override them. For
example, to urge that we have a right to free speech, in a society whose
government practises censorship, has little practical use. What we need
is the means by which to exercise that right, that is, both the skill or
ability to do the thing, and the power or freedom to exercise it. In the
context of medical research, the signiWcant power a patient needs is
autonomy, that is, the ability to make decisions about and for herself.
The question of autonomy will be considered in more detail, taking into
account both the patient, who has to be competent to exercise her
autonomy, and the doctor, who has to respect her autonomy so that she
can exercise it.
(iv) P has an immunity with regard to X: We have no duty not to do X, and
moreover no-one has the power to alter that situation. This fourth
deWnition is meaningful in the legal context only: the constitution may
grant us an immunity and no-one can take that immunity from us.
Interest theory and choice theory
Rights theories can be further elaborated by dividing them into choice theory
and interest theory. Choice theory states that, apart from the right to liberty,
51 The foundations
we only have rights in so far as someone else has a duty towards us. Hence, Q
makes a promise to P, and therefore has a duty to keep that promise. Because
of the situation, P possesses a right, namely, the right to release Q from the
promise. Ps right is the correlative of Qs duty. However, as Waldron points
out, I consider that I have a duty not to harm or maim other creatures, but I
do not regard myself as released from that duty if the creatures tell me that I
am free to maim and harm them. This observation is relevant to the question
of whether or not it is ethical to conduct risky non-therapeutic research on
consenting people.
Interest theory, elaborated, as it happens, by Jeremy Bentham, states that a
right exists only where it is possible to say in advance who is the beneWciary of
the duty, or whose interests are directly aVected by the duty (Hart, 1973).
Hence, the Wrst example of the promise made by one person to another
confers a right upon the promisee, but the second example of not harming
other creatures confers no rights, the duties being too general and only
indirectly related to particular individuals. Waldron argues that this theory of
rights can be made to work if it is taken further than Bentham takes it. Rights
can be identiWed which have no speciWed duty or indeed person whose duty it
is, but which can nevertheless be acknowledged by a moral theory. For
example, we may recognize that a baby in Somalia has a right to be fed,
without knowing what the speciWc duty is that correlates with that right, nor
knowing how that right is to be recognized in practice (such as who is going
to meet the need). What such a recognition does is to identify towards whom
duties may lie. The explication of the right is the appropriate ground for the
assignment and allocation of duties. This sort of underlying right may incur
any or all of the Hohenfeldian cluster of deWnitions; the choice of which will
depend upon the right we are seeking to assert.
Choice theory is resonant of a kind of rights theory much in evidence
before this century and not so much thereafter. This theory states that the
only genuine right that exists is the right to liberty. The theory lent itself to
minimalist, laissez faire government, which was merely required not to stand
in the way of anyone exercising his liberty except in so far as it harmed others.
It required no action to ensure that rights were upheld. On this account,
socio-economic rights such as the right to free medical assistance, elementary
education and a basic standard of living were regarded as debasing the
language of rights and a category error. This century has seen a greater
recognition of these services as rights, following the extended interest theory
Waldron proposes. If we regard such services as medical assistance as a right,
then active duties are demanded, if not of us, then of someone. If, on the
other hand, we see only liberty as a right, no such activity is demanded, since
all we have to do is not to intervene. So the problemlies with deciding what is
ours by right, and what is not. If more rights than liberty are claimed, then
interest theory has to be employed. If only liberty, then choice theory will
suYce. One important consequence of this distinction is that under the
choice theory of rights, only autonomous beings can have rights, because
only autonomous beings are capable of exercising the liberty of choice that it
then becomes our duty not to obstruct. Interest theory confers rights upon
those without autonomy, such as the baby in Somalia. The task interest
theory has, however, is to state what rights people do have other than liberty,
and this, I admit, is impossible to do in any way which can be practical. This is
the predicament the NHS faces in trying to decide what treatments people
have a right to expect, free at the point of delivery. It is failing to suggest
answers which are capable of rational defence, and often resource allocation
decisions change with geography. Without a sound basis upon which to
determine the rights it is proper to defend, interest theory is in danger of
turning into Karl Marxs understanding of rights. He thought that rights
separated people from their communities and bound them in their own
private interests (Marx, 1843).
Choice theory of rights is capable of a more rational defence than interest
theory. The mutual recognition of the right to liberty would go a long way
towards according the respect that human beings should receive from each
other, further, perhaps, than is clear at Wrst sight, which sees only the laissez
faire nature of the duties this right demands. If each one of our actions has an
impact, the extent of which cannot be predicted, but which we assume goes
further than we are able to see, respecting the liberty of (all) others demands
unselWsh behaviour even when we are alone. For example, taking more
than we need of anything has an ultimate impact on the environment and
ecological balance of the world. Demanding Wrst place necessitates others
taking second place.
Does the choice theory of rights have an answer to the problem of the
Somalian baby? Interest theory gave us a reason for the duty to provide her
with nourishment and shelter. It could be argued that the right to liberty is
also suYcient for these purposes, if it is the taking of what is hers through
selWshness and greed that has meant she cannot eat or Wnd shelter, or rather
her parents cannot. They have no diYculty in doing so, if given the basic tools
(Yunus, 1998). The tools of subsistence need to remain with the babys
parents and not to be taken away by others for their own purposes, such as
civil war or greedy Western markets. Her parents need this freedom to
express their own potential in their own way, rather than patronage. In order
to exercise their freedom, assistance may be needed and positive duties may
therefore follow, such as providing the means to overcome the eVects of
natural disasters or civil war. But this can be seen as a duty to enable the right
to freedom, not a duty to feed and clothe, even if food and shelter are oVered
as part of the assistance.
Of course, the right to liberty can be abused. It only works if we use it as a
basis for respecting anothers right to liberty, rather than trying to exercise
our own. If the right is understood in the Marxist sense of giving us freedom
53 The foundations
to do what we want, it becomes licence and there is no constraint on the
choices made. But if it is taken as a basis upon which to recognize the essential
reasonableness and freedom in each person with whom we come into
contact, and indeed those with whom we do not come into contact (who may
nevertheless be aVected by our actions), it forms a useful guide to conduct,
which is what ethics is about. That is to say, we should not claim rights for
ourselves, only respect them in others. This, in my view, is the ethical value of
the language of rights.
The practical value of right-based deontological moral theory for the
analysis of the ethics of research on humans, is that we should always consult
the judgement of those who are aVected by our actions. This speciWc version
of rights theory I shall call right-based morality. In research on humans,
right-based morality translates into the requirement to consult potential
research participants before enrolling them into studies.
The validity of rights
The right that is in question as regards research on humans can be called the
right to self-determination. That is, a person has the right to withhold
consent from taking part in a research project, if she so wishes. Using the Wrst
deWnition of a right according to Hohenfeld, we might say that patients and
people in general have a right not to take part in research. For this right to be
defended, people may not only not become research participants if they do
not want to, but also they may not become research participants by default,
that is, their express permission needs to be sought before they are enrolled.
Such is the current climate of opinion, particularly amongst research ethics
committees, that this right tends to go unquestioned. But this was not always
the fashionable view, and indeed is not a view shared by all researchers even
today. If it is to be upheld, its validity needs to be demonstrated so that it can
be defended even when the assumption of the primacy of autonomy goes out
of fashion, as it may do in respect of, inter alia, organ donation.
In order to identify what underpins the right of a person to refuse to take
part in research, we will consider the circumstances of a research project
which is ethically acceptable according to goal-based and duty-based criteria.
So, the research is aiming to answer an important question which can
demonstrably be seen to be useful for future patients. It is therapeutic, and the
risks to the research participants related to the taking of the experimental or
control treatments are adequately balanced by their anticipated therapeutic
beneWts from it. Other risks from extra research procedures are minimal. The
clinician conducting the research is certain that he is doing the right thing by
his patient in putting her into the trial, because he genuinely does not know
which of the treatments in the trial is better. Is anything ethical being left out
of the decision whether or not to enrol his patient? The answer might be, that
what is left out is any consultation of the wishes of the patient in the matter.
But suppose the clinician thinks that it would unnecessarily distress and
confuse his patient to be given a choice whether or not to take part in the
research? The patient trusts the doctor; the doctor believes the trial is the best
place for the patient; the trial is aiming at a beneWcial goal. Is the doctrine of
respect for autonomy suYciently important to override even the duty-based
concern that its application would possibly cause harm? The answer, I
submit, is that autonomy is important enough. I return to the moral theory of
Kant to explain why, and relate his philosophy to the choice theory of rights.
Kants autonomy of reason
Like the Renaissance humanists, Kant equated reason and freedom, by which
he meant that reason was free from natural grounds, sensuous impulses
and inclinations, and free from empirical conditions. Distinguishing it from
the act of understanding, which is tied to the conditions of a possible
experience, reason does not follow the order of things as they present
themselves in appearance, but frames for itself with perfect spontaneity an
order of its own according to ideas, to which it adapts the empirical condi-
tions, and according to which it declares actions to be necessary, even though
they have never taken place, and perhaps never will take place. (Kant, 1787,
p. 576). It is to this essential reasonableness in each human which respect for
autonomy should rightly direct itself. Unless we give others the opportunity
to exercise their own reason in coming to understand what they ought to do,
we deny what is essentially human in them. This, together with the capacity
to express their intelligence through the power of speech, is what distin-
guishes human beings from animals. Notice, however, that it is Kants
understanding of reason that is invoked as that which must be honoured
through respect for autonomy. His understanding of reason is that it is
capable of discovering what ought to happen. It is not an expression of
personal desire, which Kant carefully separates from the exercise of pure
reason. The right that is being defended in the context of research on humans
is not the right to do whatever we please, it is the right demanded absolutely
by the exercise of pure reason. The existence of this sublime quality permit-
ted, the choice theory of rights is suYcient to support peoples right to
consent to participate in research. This is because if we did not seek their
consent we would be denying them the liberty to decide to act according to
their own reason, which, by its nature, Kant asserts, wills what is good.
This is an altogether higher notion of a right than has elsewhere been
suggested. Often, as seen at the beginning of the chapter, rights are invoked as
a way of protecting the weak against the strong, a notion investigated shortly.
What I would suggest is that the purpose of seeking consent from people is
not to protect them against stronger forces, but to give them an opportunity
to exercise what makes them essentially human, pure reason, which is
55 The foundations
precisely not weak. It is asking, in our context, the clinician to respect that
which is Wnest and most noble in human nature.
Kennedys rights for the weaker party
Kants language, and my own in the previous section, is high-Xown. We may
feel that discussions of pure reason are simply too idealistic and impractical.
Ian Kennedys analysis of patient rights is more down to earth. He has always
argued that the appropriate basis for medical lawis human rights, not just the
rights which are declared in international and national charters and conven-
tions, such as the Nuremberg Code, the Declaration of Helsinki and the
International Declaration of Human Rights, but also the inchoate rights
which are the product of reasoned moral analysis (Kennedy, 1988, p. 386).
Kennedy notes that talk of rights in the medical context has often provoked
hostility from the medical profession or its spokespeople.
This is because the rights discussed are asserted as the rights of patients. Such
assertions do not sit well with a profession which is, quite properly, educated to think
of itself as concerned with the diYcult task of caring for the sick and seeking to
overcome illness. Talk of rights is often represented . . . as if it inevitably involves . . .
strict or rigid restrictions on what the doctor may do. (Kennedy, 1988, p. 386)
Recalling Hohenfelds deWnitions of rights, it can be recognized that what is
under threat is the privilege right which doctors have traditionally enjoyed,
and Kennedy points out that, after all, clinical freedom is itself a right. But
there is a counterpart to the doctors right to clinical freedom, argues
Kennedy, which is the patients right. It is the patients right which sets the
framework within which the doctors discretion may be exercised. This right
has a further contingent importance because:
As between the doctor and the patient there is an inevitable imbalance or disequilib-
rium of power. The doctor has information and skill which the patient, who lacks
these, wishes to employ for his beneWt. (1988, p. 387)
These powers possessed by the doctor include the privilege to touch and
invade the body of the patient (p. 387). Hence, insisting upon rights that
accrue to the patient introduces a check on the privilege rights of the doctor, a
check which comes from outside his own moral and professional framework,
and which can, therefore, be judged independently of him on behalf of the
patient.
Kennedy argues that the right that must be asserted above all is the right to
respect for autonomy. The patient may not have the same power as the
doctor by virtue of knowledge and experience, and she may be in the hands of
the doctor as regards the receiving of the right treatment for her condition,
but she does, or should, have the power to agree to what is to happen to her.
We could adopt the choice theory of rights, which was the most minimal of
right-based theories, asking only that we respect anothers liberty, to be the
basis for patient autonomy. Or we could prefer the Kantian notion of the
autonomy of reason as the appropriate basis. Or we could adopt Kennedys
view that it is to redress the imbalance of power between doctor and patient
that requires this respect. In each case, I will conclude that autonomy should
be respected, and that it can be respected by always seeking the patients
permission before doing anything to her.
Scarmans test
The Kantian deWnition of reason, namely, that in a person who is capable of
making a good and right decision, is that which should be respected and
referred to when seeking consent. It is not at all the same as consulting the
personal inclination of a patient, which may be, for example, to avoid
treatment which she ought to undergo for her own well being. What if,
however, the reasoning of the patient does not concur with the reasoning of
the doctor? If it always did, there wouldbe noneedtoconsult the patient, since
she would always agree anyway. Respecting a persons autonomy requires
consulting her, which implies that the patient may, for her own good reasons,
disagree. Suppose thenthat the reasons are not, in the viewof the doctor, good
ones? Does that mean that the patient is not actually exercising her reason, but
rather referring to something else, like her inclinations; or worse still, does it
mean that the patient is not, in fact, competent to make a decision? Suppose
the doctor was convinced that his patients refusal, say, to receive a particular
treatment, was based on Xawed judgement? Should he conclude that his
patients autonomy is suspect, and override her? Or should he accept that
there may be aspects to what is proposed which are unwanted by the patient
for her own good reasons, which he is unaware of or does not understand? I
suggest the doctor should take the latter course. Otherwise the principle of
respect for autonomy becomes meaningless, since it only includes the exercis-
ing of that autonomy which furnishes similar views to those of the doctor.
The distinction between what the patient might regard as reasonable and
right and what his doctor thinks was recognized by Lord Scarman in his
judgement in the Sidaway case (1985). The question under discussion then
was what level of risk it was right to inform the patient about, to which
question I shall return in a moment. What is signiWcant for the present
argument is Lord Scarmans statement, that the determination of what
should be told to the patient depended not upon what the doctor thought
would be appropriate, but rather a reasonable person in the patients position.
Hence the focus was on the patients view, not the doctors. The judgement
resonates with Kants account of rational beings and their proper treatment,
though it should be noted that this prudent patient test does not focus on
what the individual patients view might be, but, rather, what a reasonable
person would want.
57 The applications
However, to impose upon the doctor the rule of accepting his patients
views even when they disagree with his own may put impossible strains on his
integrity, if the disagreement lies in situations which may require the doctor
to do what he really believes to be wrong. The adoption of the choice rather
than the interest theory of rights helps to distinguish between those circum-
stances where the doctor must accept the views of his patient, and those where
he need not. It would be wrong to override the wishes of a patient when they
consist merely in the refusal of a treatment the doctor thinks is right. But to
respect the wishes of a patient who is asking for something the doctor believes
is wrong is taking the rights of the patient too far, not least because the
consequence of respecting that right would be to oblige the doctor positively
to act in a way contrary to his own reason. This distinction is made, for
example, in the explicit recognition of a doctors right to refuse to help to
procure, or to carry out, abortions. This right was reiterated at the 1999
British Medical Association Conference.
The applications of right-based thinking
In the context of medical research we are not faced with the situation of a
patient asking for a treatment. We are concerned with cases where patients,
or healthy people, will be asked to become research subjects. The right-based
morality that I am going to adopt for the purpose of analysing the ethics of
research on humans consists, then, in this: that we should consult the wishes
of those most aVected by our actions. I am also going to include the right of
patients to have their conWdentiality respected. This is a narrow deWnition of
rights, but it is suYcient when taken together with the goal- and duty-based
approaches to complete the analysis of the ethics of medical research on
humans.
The consent procedure
For consent to be valid, it must be (i) given by a competent person; (ii)
adequately informed; and (iii) voluntary.
Competence and the use of vulnerable groups in research
There will be times when potential research subjects have no competence,
such as when they are unconscious, mentally incapacitated (though this is
contentious), or too young to make a reasonable decision. Such cases cannot
in fact be dealt with under our understanding of right-based morality since
these people do not have the ability to express choice. However, these cases
raise the consent issue in respect of who should be consulted if the patient
cannot be.
Determining a childs competence
Children gain competence to diVerent degrees and about diVerent things
over time. Some very young children, who suVer chronic medical conditions,
may have a complete understanding of their situation, and be able to
participate fully in any consent procedure. Others will be less comprehend-
ing. In the UK, the law recognizes this diVerentiation and oVers what is
known as the Gillick competence test to establish whether the child or her
parents should be the ones to consent to treatment. The Gillick test arose out
of a famous case (Gillick v Norfolk and Wisbech Health Authority, 1986) in
which a mother took a Health Authority to court over the fact that a General
Practitioner had prescribed contraceptives to her 15-year-old daughter with-
out informing her. She eventually lost the case, which went all the way to the
House of Lords, and the Gillick test emerged. It states that if, in the view of
the clinician, the child is capable of understanding the nature and conse-
quences of her treatment, that child is deemed competent and her consent
overrides that of her parents. This ruling does not mention research, and I
imagine that a court would expect a greater level of competence in order to
validate a childs consent to research than to treatment. She might need to be
able to understand randomization, altruism, or blind allocation of treat-
ments, to name but a few of the concepts that arise in the research as distinct
from the treatment context.
Decisions on behalf of an incompetent child or adult
Where a child is clearly incompetent, such as a baby, then the people in loco
parentis are the ones whose consent has to be sought. But their consent can
only be given to that which is in the best interests of the child. This, of course,
is precisely what becomes the most relevant question in relation to any
incompetent potential research subject. Interestingly, however, while it is the
parent whose consent must be sought on behalf of an incompetent child, in
the UK the law decrees that no one may consent on behalf of another adult. It
is up to the persons doctor to decide what is in her best interests. As with
parents deciding on behalf of their children, this process is not at all the same
as the process of obtaining consent from the person who is the potential
research subject. Right-based morality is about respecting the autonomy of
the interested parties. It is about precisely not doing what we think is the best
for another, but eliciting from her what she believes to be right for herself.
Whether it is the researcher/doctor or a parent who has to make the decision
on behalf of the incompetent patient, that decision will be made on her
behalf, using best interest criteria. This is a duty-based moral decision, not a
right-based one. The use of advance directives, made out before incompe-
tence sets in, indicating patients wishes, has increased the possibility of using
right-based criteria to determine courses of treatment. Even advance direc-
tives may be disregarded, however, if it is the doctors view that its requests
are not in the patients best interests.
59 The applications
When an incompetent adult or a child clearly needs treatment, it is given
without consent, and its moral justiWcation is duty-based: that is, it is in the
patients best interests that the treatment be given. If research is proposed
involving incompetent subjects, the duty-based issue of best interests be-
comes the important deciding factor. In therapeutic research there is the
equipoise test together with the consideration of whether the interests of the
individual concerned would be best served by receiving treatment as a
participant in a trial. Some would regard these criteria as suYcient to make it
ethical to enrol an incompetent patient into a therapeutic trial. Others
consider that any randomized controlled trial cannot, by deWnition, be said
to be in each individual patients best interests, since the treatment is not
being tailored precisely to each individual. Non-therapeutic research on
people who cannot consent is more diYcult to justify, for non-therapeutic
researchcannot be said to be in the best interests of a research subject. If she is
unable to consent to it on her own behalf then no moral justiWcation for the
research is left except the goal-based one of the desirable consequences. It can
be the case that the only way important developments in some areas of
medicine can be made is if non-therapeutic research on incompetent subjects
takes place, for example, research into the causes of Alzheimers disease.
These can be the most diYcult research projects about which to make ethical
decisions, particularly when the goals seem very important. Examples of this
will be discussed in Chapter Seven, where it emerges as a duty-based issue.
SuYce it to say here that allowing such research to happen, even when the
goals are extremely signiWcant and the risks minimal, is no light matter, for
the principles of behaviour demanded by duty- and right-based morality
have perforce been compromised. It was the compromising of precisely these
principles that made possible some of last centurys worst excesses, starting
with the notion that (some sorts of ) people were expendable.
Competence in the ordinary adult
In the UK, as in many other countries, the law regards everyone over the age
of 18 as competent unless there are compelling reasons not to, for example if
the person is unconscious. In practice it is more accurate to say that there are
degrees of competence even amongst those whom the law deems fully
competent. In medicine, and medical research, patients are frequently not
fully competent. This may be because of their illness, their medication, their
physical circumstances or their emotional state. These factors encroach on a
persons competence and cannot be ignored if the consent procedure is to be
taken seriously. At the same time, the laws clear-cut approach reXects a
Kantian notion of personal responsibility: that rational beings are answerable
for their actions, so that even if someone does not feel very competent she
should still try to make right decisions if called upon to do so, accept
responsibility for what happens to her, and be willing to face the conse-
quences of what she does decide.
Adequacy of information
In order for a patient to make a valid decision about whether or not to take
part in research, she needs to comprehend information whose context is
medicine. It takes several years for an individual to be trained suYciently to
call himself a doctor. Medical researchis by its nature often at the vanguard of
medical practice and so specialized that even qualiWed doctors can Wnd it
diYcult to understand the nature and purpose of diVerent research projects.
There is evidence that concepts which are not found solely in the medical
context, such as the concept of randomization, are not easily explained and
understood (Snowdon et al., 1997). It is unlikely, then, that someone with no
medical training is going to understand the detail of what is proposed. We
can imagine taking time to inform a single individual whom we know well,
tailoring the information-giving process to her particular needs and ability,
but the research context demands giving information to hundreds, some-
times thousands, of people. I would not be wholly pessimistic about this
process, provided those tasked with communication are aware of the chal-
lenges. The best teachers are the ones who, though experts in their subjects,
are nevertheless able to convey the principles of their subject to schoolchil-
dren. It has to be acknowledged, however, that the physical circumstances of
many healthcare settings are not conducive to a genuinely informative
exchange between a patient and her doctor.
Voluntary nature of the decision
Freedom to choose is the ultimate expression of autonomy. However, for the
patient being asked to take part in research there are numerous inXuences
aVecting her freedom, related to: her condition, which may have an eVect on
his mental state; her location, which may be in the somewhat depersonalizing
and dependent atmosphere of a hospital bed, in which, despite everyones
best eVorts, she still feels she has to ask permission to do anything; and the
all-important relationship with her doctor. It is this last factor which seems to
me to count most against a free decision being made, for reasons which are
not all bad. The doctors powerful position in relation to his patient can give
rise, not only to fear that the doctor will not be pleased if the patient does not
agree to his requests, but also to trust, in that she will assume that the doctor
will never ask her to do anything that is bad for her. She might, therefore,
believe that there is no reason to refuse any request he might make. There-
fore, the doctor should never ask his patient to do something which is
harmful, whether it is in the interests of future patients or not. This would be
counter to his raison detre, which is, after all, to heal, not to harm, but it does
detract from the voluntary nature of any decision the patient may make.
61 The applications
Concluding remarks on consent
Although there are many obstacles to the success of the consent procedure,
the need for it is still paramount, as the discussion in the Wrst half of the
chapter showed. Not respecting a patients autonomy is equivalent to treating
her as expendable. In Kants terms, she should be treated as an end in herself,
and never merely as a means to an end. According to Kennedy, not seeking
consent leaves all the power in the hands of the doctor. He already has the
advantage over his patient, in that he has more knowledge than his patient.
Giving the doctor the freedom to act as he thinks best, without consulting his
patients wishes, doubly disadvantages the patient. The choice theory of
rights, which is the most minimalist of rights theories, nevertheless places an
obligation on the doctor to seek his patients consent, because otherwise he
will not be respecting her right to liberty.
For these reasons, we should not discard the consent procedure despite its
practical diYculties. Rather, we should Wnd ways to overcome them. It
should be recognized, however, that in the medical context, there are some
things which are too risky for doctors to ask their patients to face, even if,
particularly if, the patient might agree if she was asked.
Condentiality
ConWdentiality is a right-based issue because it involves respecting that which
is private to a person, which, if it cannot be said to belong to her, certainly
does not belong to anyone else. This is information of a personal nature,
which is disclosed only on the understanding that it will be kept conWdential,
as is the case in the medical encounter between patient and doctor. There are
numerous cases of research where conWdentiality is threatened or compro-
mised, one of which will be discussed in Chapter Eight. The principle is
simple: if a patient speaks to her doctor about her private medical life she
does so on the express understanding that her doctor, if he does pass the
information on, will only do so to those who have a direct responsibility for
her healthcare.
Consent to the disclosure of private information to those with direct
responsibility for the patients healthcare may be inferred and need not be
explicitly sought. However, a doctor cannot infer his patients consent to her
identiWable medical information being made available to researchers, who do
not have a direct interest in her medical care. If such disclosure were to be
made, it would be a violation of the right-based demand to respect a persons
privacy. It would not be justiWable on duty-based grounds either, since it is
not in a persons best interests that her records are disclosed. The reasons for
doing it (and they are often compelling) would be goal-based. For example,
the national incidence of a particular disease plays an important part in the
planning and delivery of the right kinds of health care. The information
about this is gathered by epidemiologists, who will search patient records for
it. Sometimes thousands of records are involved. Seeking consent from each
patient before looking non-therapeutically at their records would honour the
right-based principle of respecting autonomy, but it would also render the
research impossible logistically and Wnancially. So, currently, such research is
conducted without patient consent. In the UK, as elsewhere, researchers are
bound not to disclose medical information to any other source (Department
of Health, 1996), but they themselves have access to it. Another example is of
researchers seeking a particular patient population for the purposes of their
research, for example, all people in a particular area who are recently be-
reaved for research into bereavement recovery rates. These researchers may
Wnd out lists of such patients from GPs and then approach the potential
recruits directly, potentially causing tremendous upset. Also, there are an
increasing number of disease registers from which researchers can obtain
information about particular patient groups with a given disease, thus avoid-
ing asking GPs. If those patients whose details are on such registers do not
know they are listed there, there was a breach of conWdence when the
patients name and details were entered on the register (not at the point when
the researcher gains access to the information). One of the reasons for
creating disease registers is to make such information easily accessible for
valid purposes, and no doubt patients with conditions which would beneWt
from more research would welcome better organization of such information,
were they to be asked. The ethical problem arises because of doing it without
asking Wrst.
These examples of breaches of conWdentiality are instances where goal-
based and right-based claims to determine what is morally right come into
conXict with each other. As with the dilemma over non-therapeutic research
on incompetent subjects, it is essential that researchers, if they choose to go
ahead with such research, and research ethics committees, if they choose to
approve the research, recognize the cost of achieving the goal in terms of the
loss of deontological principle. On the other hand, the harm to patients
whose records are used in this way is negligible. It could even be argued that
there is no harm, if the patients never know that researchers have been
allowed to see their records. Meanwhile, the advantages of records-based
research are numerous. This sort of research sets goal-based moral values at
loggerheads with duty- and right-based ones. Solutions, therefore, will always
be unsatisfactory to some people.
This chapter has considered four diVerent accounts of rights: the Hohenfel-
dian cluster of rights, which helped to clarify what the word right can mean
in diVerent contexts; Waldrons distinction between choice and interest
theories of rights, where interest theory demanded positive duties and choice
theory demanded respect for peoples liberty; Kants elucidation of the
autonomy of reason and its consequent requirement that all rational beings
be treated as ends in themselves; and Wnally, Ian Kennedys assertion that
when there is an imbalance of power in a relationship the weaker party must
be granted rights which place strict constraints on what the stronger party
may do. In the context of medical research on humans, the practical recogni-
tion of rights lies in the seeking of consent from any potential research
participant, and also in respecting her conWdentiality.
The three factors that make up the consent procedure are that the person
whose consent is sought be competent, adequately informed, and able to
make her choice voluntarily. Competence is absent in some people, in which
cases the duty-based demand that they are treated in their best interests takes
precedence. If a child, her parents must consent or refuse according to, again,
her best interests. Competence may be compromised by states of health, but
it has to be assumed to be there in people over the age of 18, and sometimes
younger, for the purposes of the consent procedure. Information can be
diYcult to convey. Finally, the choice needs to be voluntarily made, and this
may be aVected by the way the patient relates to her doctor.
These obstacles to success do not mean that consent should not be sought.
Rather, they indicate that it is unwise to rely upon the consent procedure to
validate the ethics of a research project on its own. Hence, the right-based
approach to moral thinking is, like goal- and duty-based thinking, necessary
but not suYcient for the purpose of ensuring that research is ethical. Never-
theless, even if the seeking of consent is not the only important ethical
principle to uphold in research on humans it is still crucially important, as
the earlier part of this chapter showed. The analysis of the process of
obtaining consent emphasised the point that a researcher should only seek
consent to do things which are right, as far as he is able to determine. Hence,
the goal-based demand that the research be aimed at something good and
important must be met Wrst, then the duty-based demand that no inappro-
priate harm should befall the research participants who take part should be
met, and then Wnally, the patients wishes must be consulted. These three
components, based on the three approaches to morality, all play their part in
providing a framework for determining the morality of a research project
involving human participants.
There are some sorts of research projects which violate right-based moral
claims. Research on incompetent participants, and research which involves
breaching conWdentiality, are both examples of this. In these situations, even
deciding not to conduct the research on the grounds of caution is morally
problematic, for it means that goal-based moral claims are ignored. In the
second part of the book cases will be encountered where the three approaches
come into conXict with each other, and we will see whether it will be possible
to come to resolutions that do suYcient justice to all three approaches.
65
5
From principles to practice
Introduction
This chapter acts as the link between the three theoretical chapters and the
three practical chapters. Until now the three approaches to moral questions
have been discussed in the context of human participant research at the level
of ideas and theoretical applications. In the next three chapters cases will be
analysed using the three approaches. The same order of presentation will be
followed. Each chapter will focus on one of the three approaches, and the
practical issues discussed will mirror those raised within each of the theoreti-
cal chapters. This introductory chapter will summarize the philosophical
foundations of each of the three approaches, and introduce the cases to be
discussed in the second half of the book.
Goal-based morality
Goal-based moralitys theoretical basis summarized
Chapter Two introduced goal-based morality, describing the theories of
some of its more famous utilitarian forbears. Utilitarianism considers the
outcome of an action to be its moral determinant, rather than the content of
the action itself. In simple terms, this means working out the extent to which
an action maximizes happiness. Jeremy Bentham thought that each person
aVected by an action should count as one, so that the more people who were
made happy by an action, the more justiWed it was. This simple theory,
although attractive because it requires no complex moral code or belief in a
higher authority, nevertheless countenances and even justiWes harm to some.
It does not seek to condemn any action whatsoever, if it maximizes happi-
ness.
In response to this accusation of being open to justifying bad acts for
predicted good consequences, rule utilitarianism was proposed. In this the-
ory, rules are applied which are utilitarianly justiWed: if followed, they will
lead to the greatest happiness for the greatest number. Although this is a
more acceptable moral theory because it opens the way to outlawing certain
morally abhorrent acts, it has lost the simplicity that had made utilitarianism
attractive in the Wrst place. For the rules have to be devised prior to particular
66 From principles to practice
actions being proposed. With act utilitarianism, as Benthams utilitarianism
is known, the situation is real and the consequences of an action easier to
foresee. With rule utilitarianism, the situations are anticipated and their
consequences hypothetical. Hence an authority has to be recognized. Where
are the rules to come from, if not from tradition, belief and even religion?
R.M. Hare, for example, a rule utilitarian, Wnds his rules in the Christian
commandment to love God and neighbour.
In any case, if we are hard-headed utilitarians, are we not required to
perform that act which maximizes happiness, regardless of the content of the
act, because we are not supposed to mind if moral rules from tradition or
religion are broken? Utilitarianism has an inherent debasing pull, as Williams
has pointed out, in that, being by nature pre-emptive, it countenances actions
which, taken in themselves, are fairly nasty, in order to avoid what are
perceived as worse consequences. In medical research, the more important
the goal, the more the claims of duty- and right-based morality will be under
threat. However, not to consider the consequences of a research project as
part of the ethical analysis would be literally nonsensical, since research is
always undertaken to achieve a goal.
Goals of research in theory and practice
It is, therefore, important to consider whether the consequences of a research
project are good and desirable, even if they do not, Wnally, justify the action.
This moral approach, adopted as just one part of the overall moral analysis, I
called goal-based morality. In research, this involves considering the goals of
the research in relation to their contribution to health care. Once it is clear
what the research is aiming to do, the way in which it is carried out can be
described, and then subjected to the scrutiny of duty- and right-based moral
questions. Deciding what is a proper goal for research is not easy. How
should research priorities be set? The possible goals a researcher might
embark upon a research project to achieve are many, ranging from open-
ended scientiWc curiosity to the promise of Wnancial gain, with the gamut of
answers to a variety of research questions in between. The challenge of
goal-based morality is to Wnd a reference point for the moral validity of the
research goal. This attempt raises fundamental questions of what medicine
should really be trying to achieve. I suggested that the ideal that researchers
should hang on to is that of discovering complete cures in their medical Welds
which have no side eVects. This provides an ultimate goal towards which all
research can ultimately aim, even though the everyday reality is that research
takes steps on the way to it.
In Chapter Six, where actual cases are discussed by way of illustration of
issues in goal-based morality, organ transplantation is taken as a case study.
The procedure for organ transplantation involves harm in several diVerent
forms, which research is continually seeking to remedy with improvements
and reWnements. In particular, diVerent sources for organs are being inves-
tigated. So far, no method has been achieved which does not convey harm in
one form or another. But the future holds new possibilities for safer, better
treatment.
Research method summarized
Once the goal of the research has been identiWed, the appropriate method for
achieving it has to be chosen. As described in Chapter Two, this might be a
randomized controlled trial, in which groups of patients are randomly
allocated the treatments which are being tested, and both researchers and
participants are masked to avoid bias. The method chosen might be observa-
tional, with no experimental intervention. Or it might be qualitative, seeking
to elicit the views of the researched with no imposition of the researchers
beliefs on the study or its results.
The practical implications of the choice of research method
Once the appropriate method has been chosen, it is important to let the
science speak and not to superimpose beliefs, however rational they might
seem, on to the evidence. The importance of this discipline is demonstrated
in the two cases related to research method that are discussed in Chapter Six.
One, a historical example, is the discovery of penicillin. This story shows the
consequences of holding particular notions of how the world works. Penicil-
lin was not properly discovered until old ideas about the immune system and
the eVectiveness of drugs had been dropped. A modern example of research
into homeopathic medicines is then considered. The literature indicates that
similar prejudices are standing in the way of clinicians being able to see the
implications of the results of the research.
Summary of the problems with disseminating the results of research
Chapter Two said that disseminating the results of research is an intrinsic
part of the research process. It then considered the obstacles which lay in the
path of such dissemination. These were, Wrst, that encouraging clinicians to
put evidence into practice is not easy, because of ignorance of the research,
unreasonable intransigence, or because the evidence is not suYciently per-
suasive. Second, it is diYcult to collect the right evidence together, because of
publication bias (which might include not wanting to publish because the
research was conducted unethically), because of researchers and their spon-
sors wish not to share results, and because the Wndings in published original
papers needs to be supplemented by the correspondence which follows, and
this is not often cited. Third, dissemination of results is part of the wider aim
of encouraging evidence-based medicine, which often only deals with quan-
tiWable results, and therefore might miss important, non-quantiWable aspects
of medical practice.
Introduction to practical examples of dissemination problems
Chapter Six discusses, by way of illustration, research into so-called futile
treatment, the results of which are to many minds morally unacceptable. It
then brieXy discusses pharmaceutical company research which found itself
the result it wanted by dint of extended Wshing. Then some research into
fetal monitoring during labour is mentioned, which produced results of
statistical signiWcance which were of no use to women in labour, nor to their
babies.
This third aspect of goal-based morality in research seems to undermine
the Wrst two, for what is the point of believing the research evidence if it is
beset with problems? However, the examples are meant to be cautionary, not
discouraging. They emphasize the need to ensure that the goal with which the
research endeavour begins is really the one that is called for by reference to
patients needs and not serving some smaller aim. The problemwith the three
cases discussed under dissemination of results in Chapter Six was that the
research question with which they began was misguided.
Duty-based morality
Duty-based moralitys theoretical basis summarized
Chapter Three described two formulations of duty-based deontological mor-
ality. One is the tradition of natural law ethics, and the other is Kants
autonomy of reason giving rise to the categorical imperative. Natural law
seeks to give an account of the world which provides a basis upon which to
establish moral behaviour. The thesis is that everything has a natural func-
tion, and works according to that function. The natural way to live is the right
way to live, and this can be discovered by investigating the way the natural
world works. The diYculties with this thesis are that all deWnitions of
natural are open to question, and that equating what is natural with what is
right can be nonsensical or illogical. However, the example given was that of
the law of gravity. We cannot break the law, but we can ignore it. In ignoring
it we suVer the consequences. That a doctor has a duty to act in her patients
best interests is true in the same way that the law of gravity is true. Natural
and right are uncontentiously identical in this case. Moreover, the eVect of
following that natural law is good. The other formulation of duty-based
morality is from Kant, who argues that, nature having endowed us with
reason, we are bound to live dutifully, according to the moral law, rather than
seek our own happiness. Reason is autonomous in that it can determine what
is absolutely true by its own light, and not by virtue of some other agency
which will be unreliable because of being external. Reason seeks to direct the
will by means of the moral law. The moral law is formulated by Kant as the
categorical imperative, that is, it must be obeyed by the precepts of reason.
The categorical imperative is to act only according the maxims that can
become universal laws, and only if we can consistently will that those maxims
governing our actions should become universal laws.
The practical implications of duty-based morality
The application of duty-based morality for the purposes of medical research
on humans is to be found in the way doctors/researchers should treat their
patients/participants. As doctors, they must act in their patients best inter-
ests. In therapeutic research, which is conducted in the context of medical
care, the patients who are the research participants must be treated in the trial
at least as well as they would be treated outside the trial. For this to happen,
doctors must be in what I have called strong equipoise. Strong equipoise is
the state of mind whereby doctors do not know which of two or more
treatments being oVered in a trial is best for their patients, and are therefore
happy that their patients best interests are served if they are randomly
allocated to any one of them. Duty-based morality would also require that
doctors should consider each of their patients as individuals with speciWc
needs when deciding whether or not to enrol them in their research, and that
the research is designed in such a way that researching doctors can continue
to treat each participant in his best interests.
Introduction to examples
If a trial has amongst its treatment options a placebo arm, that is, the
possibility that a participant may receive no treatment, that must also be
accounted for in the doctors state of equipoise. So they must be happy that
their patients might receive only dummy treatments. This can present prob-
lems, discussed in theory in Chapter Three. The inclusion of a placebo arm in
a trial is scientiWcally very important, as it adds to the rigour of the trial and
makes the data easier to interpret, which is a goal-based imperative. And yet it
can be directly counter to the doctors duty of care. Chapter Seven discusses
the issue of placebo-controlled trials in considerable detail, using several
examples, for it is an issue that is likely to continue to vex researchers and
research ethics committees for as long as trials using randomization con-
tinue. It is a good example of where duty-based and goal-based moral
requirements come into conXict, where diVerent people will be arguing an
ethical line and yet disagreeing with each other.
Another situation in which duty-based morality comes into conXict with
goal-based morality is in non-therapeutic research, in which there is no
intention to beneWt the research participants. In this situation of no beneWt,
there may still be risk of harm. Medical science may need to Wnd an answer to
certain questions, and future patients may beneWt from the answer being
known, and yet this cannot, as mentioned in the discussion of goal-based
morality, be the sole justiWcation for using people as research participants,
especially as they may be harmed. The general rule given in most guidelines
for research (Foster, 1997) is that non-therapeutic research should expose
participants to no greater than minimal risk. This is something of an arbitrary
measure, and of course diVerent views will be held about what constitutes
minimal risk. Chapter Seven discusses a controversial trial in which babies
were placed in tents with 15% oxygen, to mimic the atmospheric conditions
of an intercontinental Xight, in an attempt to investigate a possible cause of
sudden infant death syndrome. The research was thoroughly debated froman
ethical point of view before it commenced, and its publication was hedged
about by ethical commentaries. Nevertheless, the research went directly
against the tenets of duty-based morality.
Duty-based morality also clashes with right-based morality when research
participants, fully competent and healthy, agree to take part in research
projects which will lead to harm, sometimes fatally so. This can and has been
the case with healthy volunteer research, which is also discussed in Chapter
Seven. Healthy volunteer research usually takes the form of Phase I studies of
novel chemical agents. These agents, not yet called medicines, are given for
the Wrst time to humans to measure toxicity and pharmacokinetics, before
being tried out in the patients for whom they are ultimately intended.
Researchers perform medical examinations on the often young and impe-
cunious students who volunteer to participate, in order to ensure that they
are in a Wt state to be research participants. However, not everything can be
discovered by means of medical examinations, and the researchers rely on
truthful accounts by the volunteers of their alcohol and drug use, and other
potentially dangerous activities or states of health. The right-based view is
that these volunteers, being competent and informed, can and should be
relied upon to take care of their own welfare. The cases discussed in Chapter
Seven indicate otherwise, oVering a useful illustration of the conXict between
duty-based and right-based morality which can arise in healthy volunteer
research.
The principle that the doctor/researcher has a duty to beneWt and not to
harm her patients is what underlies the concern to uphold duty-based
morality. Like the tremendous importance of the goal of the research being
chosen rightly, so the doctors duty to care needs to be retained at all times. If
a research method demands a compromise of the doctors duty, it is the
science which should change, not the duty.
Right-based morality
Right-based moralitys theoretical basis summarized
Chapter Four contained an account of some diVerent theories of rights which
have been used in moral philosophy, to underpin a discussion of the place of
rights in medical research on humans. Several accounts of rights were
considered. Choice Theory, the articulation of the right to freedom, was
taken as one helpful way to think about rights for our purposes, together with
Kants description of human reason. In these terms, seeking consent to
participate in research is an appeal to the rational aspect of a person, who will
thereby make the best decision for all concerned. It is not a consultation of
mere whims and fancies. Another way of thinking about rights which was
discussed in Chapter Four is that proposed by Ian Kennedy, in which the
articulation of the patients rights takes the position of power away from the
doctor and equalizes the relationship between doctor and patient. Expressed
as the patients consent to treatment (research, for our purposes), rights give
patients the Wnal say as to what is to happen to their body. Their consent to
participation is the gatekeeper of the actions of the doctor.
Introduction to examples
However, there are diYculties with the process of obtaining consent, which
were identiWed in Chapter Four. Chapter Eight gives those diYculties sub-
stance by reference to a number of empirical studies, such as the one which
showed considerable confusion about the concept of randomization amongst
parents of neonates in a research project. Other empirical studies discussed
investigate whether seeking consent gives rise to greater anxiety, which if true
would be a duty-based reason for being wary of seeking consent. Yet more
studies indicate that the consent process reduces recruitment rates of subjects
for research, which threatens research. This, of course, is a goal-based con-
cern, and provides a goal-based reason for being reluctant to take the consent
process seriously. These threats to the claim of right-based morality, that
seeking consent is of paramount importance, are not inconsiderable, which is
why it is necessary to be able to give substance and weight to the principle that
is elucidated in Chapter Four.
Chapter Eight also discusses research which cannot take place if consent is
sought, when the outcome of the research would be aVected if patients knew
they were research participants. This represents a clash between goal-based
and right-based morality. The case used for illustration was published in the
British Medical Journal. In passing, the issue of whether research which has
been conducted unethically should be published is discussed. Publishing it
arguably provides an opportunity for something good to come out of
something bad, but it also tacitly encourages bad practice. The discussion on
consent is Wnished by ruminating on the status of written consent, and some
empirical evidence for the way it is perceived by patients is oVered.
ConWdentiality is treated as a right-based issue in this analysis, and its
importance and philosophical underpinning were described in Chapter Four.
Chapter Eight discusses the case of records-basedresearch, in which compell-
ing goal-based arguments for conducting this sort of research without con-
sent compete with right-based arguments for respecting patient conWdential-
ity.
Does the three-approaches framework succeed?
Chapter Nine contains a discussion of the success of the three-approaches
model for considering the ethics of medical research on humans. In this
chapter I admit that the approach is by no means fool-proof. The role of
research ethics committees is also discussed, as is the importance of re-
searchers taking full moral responsibility for the research they propose to
conduct. To demonstrate this point, the story of the way research has been
governed in this country fromthe Nuremberg Code of 1947 until nowis told.
Step by inevitable step, the governance of research on humans has become
increasingly regulated and restricted, bound by that voracious question, Who
guards the guardians? Given that we cannot ultimately rely on others to guard
and protect our own morality, there remains an imperative for each individ-
ual to take to heart, which is to take full moral responsibility for his own
actions. The book ends with a plea and a guide for this.
73
6
Case studies of goal-based issues
Introduction
This chapter will cover speciWc cases of research whose goals, scientiWc
methods and dissemination of results create particular and interesting ethical
problems. As an example of ethical problems in research goals we will look at
the aims of diVerent branches of organ transplantation research. In seeking to
improve the chances of people who are in need of neworgans, tissue and skin,
research moves in such ethically sensitive areas as elective ventilation, xeno-
transplantation and the use of embryos as a therapeutic tool. When we come,
secondly, to look at issues in research method, two examples will be con-
sidered. The Wrst is the discovery of penicillin, and the second is research
into homeopathic treatments. The two accounts have some similarities.
Thirdly and Wnally the dissemination of results will be discussed, looking
speciWcally at cases which highlight the diYculties in making evidence-based
medicine a reality. These are: research into administering cardio-pulmonary
resuscitation to neonates to determine whether and when that treatment is
futile; research into uses for interferon; and research into fetal monitoring
during labour.
Goals of research
When this issue was discussed in Chapter Two, it was suggested that re-
searchers should aim at discovering treatments that cure without harm to
individual patients or to the wider environment. The development of treat-
ments for those whose organs are in disrepair and need replacing will now be
taken as a case study of research goals.
Organ transplantation
Initial fears about Frankensteins monster were overcome as, in the case of
some organs at least, successful transplantation and the good quality of life of
the recipients were demonstrated many times over. The outcomes of organ
transplant operations continue to improve as more is understood of the
treatment of what is called graft versus host disease, which is the problem of
74 Case studies of goal-based issues
the recipients immune system developing resistance to the donors organ. A
balance has to be struck between the immuno-suppressive drugs the host
must take, to help her body accept the alien organ, and the need for the hosts
immune system to be allowed to continue to work against other disease
insults. This is more-or-less successful in diVerent bodies and with diVerent
organs. The receipt of an organ makes the diVerence between a person having
a disease which is incompatible with life, and one which is compatible with
life. That is to say, organ transplantation is not a cure even though it is
life-saving. Hence, one of the qualiWcations for receiving an organ is that the
patient must agree to adhere to the drugs regimen and life-style requirements
that go with having someone elses organ in her body. Research in all the
diVerent aspects of organ transplantation is geared towards the successful
receipt and carrying of a foreign organ. Both in terms of the overall goal and
the individual parts of the diVerent research projects that serve that goal,
there remains little that is thought to be ethically questionable here.
Two areas remain of concern in organ donation from dead bodies. The
Wrst is the respect that is owed to the dead. It might seemirrational to think in
terms of harm when removing organs from cadavers, but certain other
practices raise questions about it. When we honour the dead, at a funeral or
cremation, or at a service of commemoration, such as of those who died in a
war, the focus of our attention is the body which is committed for burial or
cremation, or the graveyard where those who fell were buried. Even now, the
crematoria of the dead of World War One in Northern France and Belgium
are tended with supreme care, and still visited by thousands who, to judge by
the comments they write in the crematorium registers, are profoundly af-
fected by what they Wnd there. The commemoration is not, of course, of dead
bodies; it is of the acts and qualities of the people who inhabited them.
Nevertheless, respecting the physical remains is a way of showing respect for
the person whose body it was. In the light of this, it must remain at least a
concern that transplant surgeons have to harvest organs from bodies for
therapeutic purposes. The furore caused when it becomes widely known that
organs and other body parts have been taken without permission is sympto-
matic of the concern and sensitivity of the issue.
The second area of concern is the shortage of organs for donation. As the
techniques for organ transplantation reWne and the number of patients who
qualify for transplantation increases, so the waiting lists for donors grow
longer. The terrible choice has to be made as to who should qualify for an
organ. Intelligent systems for keeping the allocation of organs fair have been
devised. For example, in the UK, hospitals where organ transplantation is
carried out are maintained on a rotating list. Great Ormond Street Hospital
for Sick Children, where paediatric transplants are carried out, always re-
mains in second place. So the availability of organs is based on chance, but it
is a fair chance, and children are given high priority. The decision about
75 Goals of research
which patients names should be put on waiting lists requires a deliberate
decision, however. Clinicians worry about the criteria they use for inclusion
on the waiting lists. Should psychological factors play a part? What if a
patient is eager for the transplant and yet has a history of non-adherence to
medical regimens? What if her family does not seem suYciently enthusiastic
to provide the support she is going to need after surgery? Attempts to
rationalize criteria for inclusion have failed at a national level, though they
are made in individual hospitals, or by individual surgeons. Some surgeons,
for example, believe that people with Downs Syndrome should not be
oVered the heartlung transplant they often need, for the consequentialist
reasons that they are going to die (relatively) young anyway, and that the
prolonged life a transplant gives them only stretches out the time society has
to support them, as they cannot be independent. Moreover, it is argued, their
quality of life diminishes rapidly, for social rather than medical reasons, once
they leave the protected environment of school and enter the adult world.
Where a heartlung transplant is needed and is not given, survival beyond
school age is unlikely. This harsh view is counterbalanced by numerous
successful projects of self-supporting communities of people with Downs
Syndrome. Every criterion for exclusion can have such a countervailing
argument.
I would argue that these sorts of decisions are essentially qualitative and
subjective. They are not amenable to checklists, however much fairer that
may seem. Each individual coming for consideration will have her own
unique set of circumstances. DiYcult as it is, it is far better that the decision is
made on the basis of the person who is actually standing in front of the
clinician, to whom he brings his experience and judgement, rather than by
reference to some pre-arranged code.
Needless to say, constant eVorts are made to increase the number of organs
available for donation. The number of card-carrying donors remains small;
the debate about whether to create an opt-out system of donation, whereby a
persons organs are available unless that person has speciWcally requested
otherwise, is alive and Xourishing. It is a goal-based answer to the problem,
which does not recognize the duty- and right-based moral obligations not to
take organs from cadavers unless permission is speciWcally given.
Maintaining the condition of donor organs
Organs to be used for transplant need to be in a healthy condition. From the
moment blood ceases to circulate in an organ it begins to deteriorate rapidly.
If a brain-dead patient is kept on life support, then the organs remain
vascularized as the blood continues to circulate around her body. It eVective-
ly means that a body (person?) is kept breathing and its or her heart beating
for the purposes of harvesting good-quality organs. Although the utilitarian
justiWcation for it is considerable, the action taken on its own is fairly nasty
(Williams, 1993). An account by a medical student who witnessed the process
performed on a young boy who had died in an accident is oVered here. The
process took place when the student was not yet used to such procedures and
therefore saw it as a lay person would in the same circumstances.
Finally we were ready. After a few hours of work, the key moment was at hand. The
anaesthetist turned oV the ventilator and removed the machines supplying the heart
with medications that were helping it to beat. The surgeon instructed Paul [another
medical student] and me to feel the dying heart. At the time, it felt no diVerent from
any other beating heart; it looked strong and healthy and had a regular motion. Then
the arterial clamps were placed, and the vessels that connected the liver to the rest of
the body were severed. The surgeon lifted the liver out of the body and rushed it to the
waiting basin of an ice-cold solution especially designed to maintain it for as long as
possible.
I marvelled at the powerful actions taking place before my eyes. The surgeon had
taken a knife to another persons skin and was hoisting the internal organs into the air
for all of us to see. I was reminded of an ancient religious ritual in which one life was
sacriWced for the good of the rest of the community. Such patterns were repeating
themselves here in the modern era.
Instead of following the surgeons hands as she moved to take the kidneys, I was
drawn back to the heart. In the minute after the machines were turned oV, the boys
heart rate had slowed substantially. I could not help staring at the fading, barely
beating muscle.
The Xurry of activity continued as the kidneys were placed in their proper cold
preservative. The surgeons attention moved to the pancreas. At this point, the
anaesthetist politely and uncomfortably excused herself after verifying that her tasks
were Wnished. She could do nothing else for this patient. As she stood up to leave, I
found my eyes again drawn to the heart. Now each beat was followed by a number of
seconds of stillness. The boy had been dead for 16 hours, but all of my instincts were
telling me that this was the moment, this was when Mark was really dying. I was
standing next to the cloth that separated the surgical team from Marks face. Because
of the apparatus keeping the boys chest open, the drape over his face was very close to
the surgical Weld in which all of the action was taking place. When I looked to the head
of the operating table, I could make out the outline of the boys face under the cloth.
From there, my eyes were again drawn towards the open chest, and I saw Marks
heart beat in its Wnal contraction. I found myself crying uncontrollably, my tears
hidden from the people around me by my protective goggles and mask. The surgeons
were too busy organising and removing tissue to notice or acknowledge my reaction. I
felt as if I was the only one at the table who had witnessed Marks passing; the others
were caught up in the removal of organ after organ. The heart was Wnally taken from
the chest as the person from the eye bank came to procure the eyeballs. (Rothstein,
1995).
I should point out that there are very few medical or surgical procedures that
have no gruesome aspects to them. People tolerate them, both in themselves
and in society as a whole, because of the hoped-for beneWts. In most surgery,
of course, the beneWts accrue to the one whose body is being invaded. In the
case of elective ventilation and organ harvesting, the beneWts accrue to the
recipients of the organs. This does not constitute an argument against the
practice. The circumstances are that there is a shortage of organs, and though
the dead body can be disrespected, the person who formerly occupied it
cannot be harmed. That makes it better, on balance, to harvest organs from
dead bodies rather than not to. But it is not the ideal solution even on its own
terms, because of graft versus host disease, and because there are never
enough organs to go around.
Xenotransplantation
The animal kingdomprovides a further source of organs. The goal of research
in xenotransplantation is to provide, successfully, animal cells, tissues and
organs for humans. Research is currently focused on the use of pig organs,
although the organs of old world monkeys and apes would be better tolerated
by humans. The medical problems and risks of xenotransplantation, to which
I shall brieXy give some attention, are of three kinds: physiological; immu-
nological; and microbiological (Weiss, 1998).
Physiological obstacles to the success of xenotransplantation as far as pig
hearts are concerned are not considerable. The pig heart should be capable of
functioning in a human body, even though it will have to pump blood
upwards, to which it will be unaccustomed. But pig kidneys have a diVerent
structure and function from human ones, and may not function properly in
the human body, chieXy because porcine erythropoietin, a hormone syn-
thesized by the kidneys which is essential for regulating the production and
maturation of red blood cells, does not act like human erythropoietin. There
may, thinks Weiss, be many other physiological discordances which compli-
cate the acceptance of the porcine kidney in the human body.
Immunological complications between pig and human are greater than
those seen in the transfer of human organs into human bodies, because the
graft of the animal organ can be destroyed in several diVerent ways, all of
which have to be prevented if the organ is to succeed in its new environment.
The length of time and extent to which recipients of animal organs will have
to remain on immuno-suppressive drugs to avoid rejection is not known.
Such drugs have a deleterious eVect on the bodys capacity to remain
protected from other insults, and the balance between taking suYcient
immuno-suppressive drugs to prevent the organ being rejected, and taking
doses which are small enough to ensure that the body is able to ward oV other
infections, is diYcult enough to achieve with human organ donation. Por-
cine organs may need a much greater suppression of the immune system.
Microbiological risks from animal transplantation could be great. The risk
of viruses lethal to humans being introduced into the human environment is
present. Once the physical barriers to animal viruses are broken by the
placing of animal tissue in the human body, there is a danger that the
immuno-suppressive drugs, taken by the human to prevent rejection of the
organ, will help a virus to propagate and adapt to its newhost. There may also
be viruses which are found to be lethal to humans which would have been
quite harmless to the pig whose organs have been used. Herpes virus B of
macaques only gives the monkeys cold sores but causes fatal encephalitis in
humans. The same pattern may be seen in the transfer of porcine viruses. Of
course, if the pigs could be reared in environments which keep them free
from all viruses, this last risk should not be present.
There are, clearly, serious medical problems which stand in the way of
successful xenotransplantation and much more research needs to be done to
ensure that the risk to the human recipients is minimized. There are also
wider implications of risk to the community in releasing unknown diseases.
Weiss argues that the decision to receive an animal organ is based on a
riskbeneWt analysis. The risk is not just to the individual receiving the organ,
who might be very willing to be so exposed if the alternative is death, but also
to the community which may then become susceptible to hitherto unknown
diseases. Weiss says this is just like the problem of taking antibiotics, the
extensive use of which has led to substantial resistance in some bacteria,
which has meant that the population is now exposed to more serious risk of
infection than before. Individuals needing antibiotics to cure an infection
may be willing to take the risk of future resistance, but they have something to
gain from taking the antibiotics. The general population faces the same risks
but not the immediate beneWt. In this respect, xenotransplantation research
is an example of a research goal that may cause harm at a more general level.
This is not a suYciently strong argument to suggest that the research should
not take place, only that it is not a Wnal answer.
Another, wider ethical question about xenotransplantation relates to the
fact that the source of the tissue for transplantation is a living creature, which
humans as a species have considerable responsibility not to harm. The
argument for using pigs instead of apes and old world monkeys is that the
primates would have to be born by caesarean section and brought up entirely
apart from other monkeys, a condition which would, to a primate, be very
cruel. Pigs, it is thought, perhaps mistakenly, can be reared in such conditions
without the same cruelty being wrought. Although it is clear that diVerent
animals have diVerent requirements for their welfare, and those of the
primate are more complex than those of the pig, nevertheless it seems to me
that the diVerence is only a matter of degree. Either way, the animal is being
produced as a crop for harvesting. This dilemma is no diVerent from the one
in which animals are grown in order to be killed and eaten; which is, after all,
the speciWc reason pigs are reared. There is also a question about whether
xenotransplantation is just too unnatural, and that because of that, it is
bound, in the end, to fail. Although I do not propose to defend or attack these
lines of argument, I recognize that they introduce a question mark over the
research goals in xenotransplantation. Therefore, again, if it were possible to
Wnd a diVerent way of helping people whose organs are in need of repair,
which did not involve such calculated growing and harvesting of organs in
pigs, it would be better to use it.
Autografting using cloned embryos
One such way that is currently being researched is the development of
autografting techniques, whereby a person, in eVect, provides her own tissue
and organ replacements. For this to be possible, however, the stemcells which
grow in the early stages of embryo development have to be gathered. It is
these cells which can be manipulated to grow into diVerent tissues and
organs. In order to grow and harvest stem cells, embryos have to be created.
The technique for doing this is identical to the cloning process. The nucleus
of a somatic adult cell of the human whose tissue is to be grown is trans-
planted into an unfertilized egg whose own nucleus has been removed. To
put a quite remarkable process in extremely simple terms, the transfer of a
nucleus, together with the introduction of some electricity, tricks the egg into
thinking it has been fertilized. It then grows as an embryo from which stem
cells can be harvested.
Research in this area is in its very early stages. The UK government asked
the Human Fertilization and Embryology Authority and the Human Gen-
etics Advisory Committee to submit to it a report on cell nuclear transfer and
human cloning generally. The report (HFEA/HGAC, 1998) argued that
whilst human reproductive cloning should remain, as it is currently, illegal,
the technique of cloning for the purpose of harvesting stem cells was morally
acceptable and should be brought within the terms of the Human Fertiliz-
ation and Embryology Act (1990). This Act permits the creation of embryos
for research into infertility, stipulating that they should be destroyed after 14
days, which is the point at which the embryo would normally have implanted
in the womb of the mother. The main ethical issue that arises in this
technique is that the embryo is used for purposes other than reproduction,
and is subsequently destroyed. If a life begins when fertilization takes place,
then life is being destroyed for the sake of organ and tissue growth. A
secondary issue is that the self-replication involved in cell nuclear transfer
requires the availability of unfertilized eggs. If the person needing the organ
or tissues is a woman she may be able to provide her own eggs, which would
have to be removed surgically. If she is unable to provide her own eggs, or if
the person is a man, the eggs would have to be donated, and there is already a
general shortage of eggs for donation to women who are seeking in vitro
fertilization anyway.
The fact that embryos are regarded by law as disposable items for 14 days
does not mean that it is right that they are so regarded, and again, it is worth
noting that if we were able to come up with a way of growing our own tissue
without having to create embryos I am sure we would prefer to use such a
method. Weiss thinks that eventually it may be possible.
Organ transplantation is a matter of life or death, and therefore research
aimed at improving treatments in this area has to be supported. But in each
of the techniques for producing new tissue so far discussed, harm is involved.
The practices can be and have been justiWed by consequentialist arguments of
need, and by riskbeneWt analyses, where the beneWts unquestionably out-
weigh the risks. But imagine being able to produce new tissue for patients
without such a cost. Weiss thinks it is possible that we may eventually be able
to regenerate organs from somatic, not embryonic cells, which would be
preferable. It is less unnatural than taking foreign organs. If the research goals
for organ transplantation were, in addition to saving life by these means, to
Wnd methods which harmed no-one, and which were as natural as possible,
more eVort might go into improving autografting techniques from somatic
cells. I submit that no other technique will stand the test of time, because all
the others require drastic interventions either to the host, or to the recipient,
or both.
Methods of research
When designing a research project, it is essential that the researcher is clear
what his hypothesis is. Poppers contribution to method in scientiWc dis-
covery is immensely important, since by doing his best to disprove his
hypothesis, rather than to prove it, the researcher is protected fromseeing the
evidence for his hypothesis wherever he looks, whilst ignoring anything else
(Popper, 1959). But even with this safeguard, the fact remains that the
hypothesis as it is fashioned will reXect what is normative for the researcher,
and will determine everything about the way the research is conducted and
the results that are achieved. By its very nature, then, planned research cannot
produce another penicillin. Such discoveries can happen serendipitously, but
that requires the researcher to remain open to developments that are wholly
outside his normal experience. By the same token, results from other medical
traditions, however unexpected they may be, should not be dismissed simply
because they do not seem rational by an individual researchers own stan-
dards.
81 Methods of research
The discovery of penicillin was due to luck and hard work, but not
planning
Steven Lehrers book Explorers of the Body (1979) is the source for this story,
which begins with the work of Sir Almroth Wright (Sir Almost Right, as his
detractors called him). He was the professor of pathology at the Army
Medical School at Netley on Southampton Water in 1892. Very soon after his
appointment there he produced a typhoid vaccine from killed typhoid
bacillae. The vaccine attracted attention and interest because it was badly
needed: many died from typhoid in England, and soldiers sent to hot climates
in parts of the Empire were particularly susceptible. Encouraging early
clinical trials led Wright to try to convince the Army authorities to vaccinate
all the soldiers (numbering 320 000) going to the Boer War. The authorities
remained dubious of the vaccines eYcacy, not least because Wright had
broken with the Pasteur tradition by using killed vaccines. In the end only
16 000 soldiers were vaccinated. This could, in scientiWc terms, have been a
perfect controlled trial, but the follow-up was somewhat messy. Apparently,
for example, one medical orderly documented every man who contracted
typhoid disease as having been vaccinated, believing that the fact they had the
disease proved as much.
Wright was encouraged by such results as there were; others, however,
remained sceptical. The continuing doubt in his Wndings culminated in
Wright furiously resigning his post in the Army Medical School. Securing
another professorship in pathology at St Marys Hospital, Paddington,
Wright continued to promote his vaccine. He won the favour of Lord
Haldane, the Secretary of State for War, who said that the soldiers must be
inoculated. The army establishment, however, remained unconvinced. Hal-
dane said that Wright would have to become a public Wgure in order for his
work to be recognized by the establishment; accordingly, he had him
knighted. Wright won the argument over the typhoid vaccine, and on
account of his resolution in this matter his reputation grew, together with his
conWdence in the theory of immunity which he developed on the basis of
these early successes. This theory, which he dubbed the opsonin theory, was a
hybrid of two rival theories of immunity: the Wrst being that factors in the
blood fought oV germs, and that vaccination produced these factors; the
second, borrowed from another researcher called MetchnikoV, that cells will
defend an organism against intruders. MetchnikoV called this action of the
cells phagocytosis. Wright argued that the factors produced by vaccination in
the circulating blood facilitated phagocytosis, thus far pre-empting modern
theories of collaboration between antibodies and phagocytes. However,
Wright further argued that the blood factors worked by making the germs
more appetizing for the phagocytes. Wright coined the term opsonin from
the Greek opsono, meaning, I prepare food for.
Wright distilled his discoveries into an essential phrase: mobilize the
immunological garrison. This belief was partnered with the view that drugs
are a delusion, because they could only repress symptoms, not provide a
cure. This attitude pervaded the inoculation department at St Marys Hospi-
tal when, in 1928, Alexander Fleming, a staV biologist, noted some strange
behaviour in bacteria growing on a dish contaminated by mould.
In 1914, Fleming had conducted experiments as part of the War eVort
showing that antiseptic lotions applied to wounds did no good. They killed
oV the phagocytes, thus giving the germs the upper hand. Fleming became
convinced that microbes, once in the body, could not be attacked. This belief
seemed to be supported by other experiments conducted in 1922 in which he
identiWed a substance called lysozyme.
The timing of the 1928 experiments was unbelievably lucky. When they
were repeated, much later, it was found that the destructive eVect of the
mould on bacteria could not be produced at temperatures below 68 or above
90 degrees Fahrenheit. When Fleming performed his experiment there had
been a heatwave followed by nine days of cold, when the temperature rose
above 68 degrees only twice. Not only did Flemings plate become con-
taminated with the one mould out of thousands that makes penicillin, it also
happened when the ambient temperature was just right. A third piece of luck
was observed by Dr Ernst Chaim in 1971. Fleming had thought he was
looking at the eVect that is the basis of penicillins action against animal
infections. In fact he was observing a special penicillin eVect, seen in only a
few bacterial species at just the right age and state of development.
But Fleming was too sceptical to believe he had found the long-sought-
after drug capable of curing bacterial infection. His own war work and
Wrights philosophy that drugs are a delusion weighed too heavily against
the signiWcance of his Wndings. His simple, factual paper, with the crucial
observation that the mould did not harm animals or humans, which later
won him the Nobel Prize in medicine, went unnoticed by both fellow
scientists and by pharmaceutical companies at the time. He was, by all
accounts, very bad at communicating. Added to which, he was Wrights
protege and this made him unpopular with other bacteriologists at St Marys,
who by this time hated Wright in general because he was a despotic head of
department, but in particular because he had put his own name to a paper
which described work done by another (J.B. Freeman). The medical commu-
nity outside St Marys was hostile to Wright and his arrogance.
In his Nobel Prize speech Fleming justiWed himself thus:
My only merit is that I did not neglect the observation, and that I pursued it as a
bacteriologist. The Wrst practical use was to diVerentiate between diVerent bacteria.
We tried to concentrate penicillin but found, as others did later, that it was easily
destroyed, and so, to all intents and purposes, we failed. Had I been an active clinician
I would doubtless have used it more extensively. (Lehrer, 1979, p. 307)
Privately he sounded more bitter:
I would have produced penicillin in 1929 if I had had the luck to have had a tame
refugee chemist at my right hand. I had to stop where I did. (Lehrer, 1979, p. 307)
So Fleming had no chemist to help him. Wright, who believed chemists had
not enough of the humanist in them to make them suitable colleagues in the
pathology department, would not have them on his staV. The story of
Flemings two assistants, Craddock and Ridley, who were bacteriologists and
therefore had to read up on chemistry as they went, is a tribute to the capacity
to keep going in the most inauspicious circumstances. They gave up just one
hurdle short of victory, though they were not to know that, having managed,
in the most frightful conditions, to isolate the active residue of penicillin.
Craddock later said:
We could not know at the time we had only one more hurdle to cross. We had been so
often discouraged. We thought we had got the Thing. We put it into the refrigerator
only to Wnd, after a week, that it had begun to vanish. Had an experienced chemist
come on the scene I think we could have got across that last hurdle. Then we could
have published our results. But the expert did not materialise. (Lehrer, 1979, p. 311)
In 1929, just six months after Fleming had had his Wrst glimpse of the activity
of penicillin on a plate, the attempts to isolate the substance he had found
came to an end. Ridley had boils and went on a cruise for his health.
Craddock had just been married and secured a better position at the Well-
come Research Laboratories. Fleming made no immediate eVort to Wnd
anyone to continue the work. He was so inXuenced by Wrights philosophy,
which also included the view that animals were not appropriate models for
humans, that he did not inject the extract into bacterially infected mice, an
obvious next step to todays scientiWc mind.
In 1931 Raistrick, a chemist, took the experiments further but was, like
Fleming, still unaware of the therapeutic implications of penicillin. He was
able to make the substance dissolve in ether, which Craddock and Ridley had
been unable to do. However, when he found that it disappeared along with
the ether when the ether was evaporated to extract the penicillin, he gave up
and turned his mind to other experiments. Like Fleming, he lost his assist-
ants. Like Fleming, he did not replace them because he did not appreciate
how important the work was.
Lewis Holt picked up the pieces where Raistrick left oV, but he was only
shown Raistricks assistants work, not that of Craddock and Ridley. Lewis
found a way of preserving the penicillin whilst evaporating the ether but he
was thwarted by penicillins instability and gave up. Fleming did not encour-
age him to continue trying. Meanwhile, Dr C.G. Paine in 1931 employed
penicillin therapeutically with dramatic results; he however began to Wnd
diYculty producing more penicillin in that some was highly active and some
not at all active. He too gave up, later saying that he was transferred to a
diVerent line of work which, together with the failure to produce more
penicillin, made him feel it was the right decision at the time. His experi-
ments were never published in scientiWc journals.
Although four articles were written on penicillin between 1928 and 1935,
no one believed a substance produced in a petri dish could treat animal
infection. German chemists at the soon-to-be notorious I.G. Farben Indus-
trie (which later produced Zyklon B gas), were working during the 1930s on
Wnding a more penetrative and hence more eVective dye. One such new dye
was called Prontosil. A Dr Domagh, Director of research at the German Bayer
company, a subsidiary of Farben, tested Prontosil as part of a mass screening
programme to detect chemicals that might be eVective in treating infections.
He was using animals for experimental purposes. He had already discovered
that some chemicals, which showed no activity in test tubes, were active in
animals. Prontosil showed dramatic eVects in animals, and then in humans,
the Wrst of whom to be given the substance was Domaghs daughter, who was
dying of septicaemia caused by pricking herself with a knitting needle. She,
and 1500 treated patients after her, survived and recovered.
There followed a three year hiatus. Domagh was arranging iron-clad
patents. But research on Prontosil in France pre-empted him. The French
researchers isolated sulfanilamide as the active molecule of the two into
which Prontosil split when it entered the body. Because of earlier, though
abandoned, experiments with this substance, it was not possible to patent it.
Colebrooke at St Marys, experimenting on all subjects (having decided it
would be unethical to treat only half of them, though this would have been
more rigorous scientiWcally), showed remarkable results.
However, sulfanilamide was active against some bacteria but not others.
Medical research continued to look for more eVective anti-bacterial com-
pounds, and the interest was now keen, for the prejudice had Wnally broken
down against systemic antibacterial therapy.
Howard Walter Florey from Adelaide in South Australia was the one who
Wnally brought the project to a successful Wnale. When he was appointed to
the Chair of Pathology at Oxford he determined to create what we would now
call a multi-disciplinary department, with integrated bacteriology, chemical
pathology and biochemistry. Funds were very short; Oxford University and
the Medical Research Council refused support, and the bright stars Florey
had attracted to his department were threatening to leave. The Rockefeller
Foundation stepped in and it was Floreys department, free from prejudice
and with diVerent disciplines playing their rightful parts, which succeeded in
producing penicillin. On Saturday May 25th 1940, experiments in mice were
totally successful.
By 1943 American and British army doctors were administering penicillin,
dramatically increasing the survival of soldiers suVering from wound infec-
tions. After the War, the availability of penicillin completely changed the
course of bacterial diseases which had hitherto been killers.
Of course, penicillin is not the holy grail it appeared to be when its
characteristics were Wnally recognized and put to use. Its extensive use not
only produces (albeit minimally harmful and still outweighed by the beneWts)
side eVects in the individual, it also gives rise to problems of resistance
generally. The purpose of telling the story, however, is not to celebrate the
drugs discovery as an unqualiWed success, but rather to consider the implica-
tions of how it was discovered. It was a case not so much of seeing is
believing as of needing to believe in order to see, or at least to be open to the
possibility of believing before being able to see. Once the belief was there, and
the evidence began to support it, energy and resources began to be made
available for further research and exploration.
Alternative and complementary therapy research needs open minds
Some publications about trials in homeopathy show an uncanny resemblance
to the penicillin story. Stacey (1991) says it is a failure of Western scientiWc
models to understand anything of alternative medicine, not just individual
myopia, that gives rise to the refusal to accept that alternative medicine
works.
For some time now alternative approaches to medicine have become
increasingly popular, and numerous traditional methods are reappearing:
All these ways of healing have burgeoned forth, like a stream re-appearing
from underground in limestone country, a stream which had been eclipsed
but was not eliminated (Stacey, 1991). Brewin (1993), in a critical article,
puts the popularity down to a series of factors. First, patients desire more
time and attention from doctors, including more sympathetic understanding
and, in some cases, hope. Then, they desire to be given causes and explana-
tions. In the case of serious illness, the desire to try anything and everything is
very strong, often accompanied with a desire to feel in control of the
situation, instead of accepting circumstances and facing the unknown. Final-
ly, alternative, or fringe as Brewin calls it, medicine appeals to that side of our
nature which prefers magic to logic, which he thinks is evidence of a basic
instinct we all have that is trying to Wnd other outlets since the decline of
religious aYliation. He observes that fringe medicine is not logical, since it
conducts very little testing of remedies, very little self-criticism or learning
from mistakes, and is usually based on beliefs without adequate supporting
evidence. He identiWes fringe medicines use of bad magic in its assumption
that human life was healthier when it was more natural and less civilized,
even though, as he points out, today we generally enjoy a safer life, a better
quality of life, and a longer life than our ancestors (at any rate in the West).
Fringe medicine suVers from the belief that a healthy mind protects us from
all ills, and it also takes the view that mainstream medicine has failed if it has
not cured. In fairness, Brewin observes that mainstream medicine is as prone
to believing in magic as fringe medicine is, so that claims by the World Health
Organization to bring about health for all by the year 2000 seem as far
fetched as the claim in the Oxford Handbook of Complementary Medicine that
homeopathy can be successful in all diseases. He runs through a series of
propositions that mainstream medicine has failed to take account of in the
past, leading to some really useless or downright harmful treatments being
perpetuated in the Wrmbelief that they work. He observes, for example, that if
treatment A has better results than treatment B, then it may be that treatment
A is useless and treatment B is harmful; also, if an assortment of widely
diVerent treatments all give broadly the same result, they may all be useless.
The only way that mainstream medicine is to rid itself of its tendency to
believe in magic is by carefully controlled comparison trials, starting with
pilot trials. Brewin wonders why nobody thought of comparative trials before
now. He throws down the gauntlet to fringe medicine to subject itself to just
such comparisons; he asks, for example, why the Bristol Cancer Help Centre
did not subject its very exacting special diet to a comparison trial, whereby all
the participants received the same treatment from the Centre, but the
treatment group received the special diet and the control group received an
ordinary well-balanced diet. Would it be the case, though, that had such a
trial taken place and the conclusion been that the diet did work, that that
conclusion would have been discounted by those who start with an assump-
tion such as that food is not medicine? This is the kind of obstacle that those
who champion evidence-based medicine are up against.
Trials in homeopathy
One area of fringe or alternative medicine that has taken the gauntlet of
comparison trials is homeopathy, not least because of all the varieties of
alternative treatment this is the most amenable to such trials. In 1994, D.
Reilly and others published a paper in the Lancet giving the results of a trial
which sought to reproduce the evidence of two previous trials that home-
opathy diVers from placebo. The test model was homeopathic im-
munotherapy. The summary stated:
Twenty-eight patients with allergic asthma, most of them sensitive to house-dust
mite, were randomly allocated to receive either oral homeopathic immunotherapy to
their principal allergen or identical placebo. The test treatments were given as a
complement to the unaltered conventional care. A daily visual analogue scale of
overall symptom intensity was the outcome measure. A diVerence in visual analogue
score in favour of homeopathic immunotherapy appeared within one week of starting
treatment and persisted for up to eight weeks (p =0.003). There were similar trends in
respiratory function and bronchial reactivity tests.
A meta-analysis of all three trials strengthened the evidence that homeopathy does
more than placebo (p =0.0004). Is the reproducibility of evidence in favour of
homeopathy proof of its activity or proof of the clinical trials capacity to produce
false-positive results? (Reilly et al., 1994).
In 1997 The Lancet published a meta-analysis on homeopathic trials
together with two unfavourable commentaries (Langman, 1997; Linde et al.,
1997; Vandenbroucke, 1997). Interestingly, the authors of both commenta-
ries Wnd themselves unable to believe the results, not because the evidence is
unconvincing, but because they simply cannot believe that a treatment which
is the result of a series of dilutions, such that there are no molecules left at all
of the original substance, can work. Jan P. Vandenbroucke states: Bayesian
investigators will remain unimpressed by the results of homeopathy trials;
when there is no convincing theory underlying a trial, the results will remain
uninterpretable. He acknowledges that this has implications for the status of
evidence for trials in allopathic medicine, as Reilly anticipated. Vanden-
broucke writes: It might be impossible to identify false-positive Wndings in
trials of allopathic medicines, because our belief in the proposed mechanisms
could blind us to the possibility that the trial results are wrong (Vanden-
broucke, 1997). The other commentator concludes:
Their [the authors of the meta-analysis] careful analyses expose sources of bias, but
there is enough in the study to give sound reasoning for asking for good controlled
trials. However, the scientist must question whether the diversion of signiWcant
resources to support these trials can be justiWed when a rational basis for choice of
homeopathy, or any particular modality of it, is lacking. (Langman, 1997)
The failure of these two commentators to accept evidence, or to agree that if
this evidence is not sound enough that further research is justiWed, is based
upon their inability to believe that homeopathy can work. Although such a
failure of belief is understandable (how can something work that is not
there?), these commentators demonstrate the same sort of lack of logic that
Brewin so decried in his earlier article. It resonates with the prejudices of the
staV biologists in Sir Almroth Wrights department at St Marys, Paddington.
It was their Wrm belief that microbes, once in the body, could not be attacked,
and that nothing that was prepared in a petri dish could treat animal
infection, that prevented Fleming from seeing what he was looking at, as it
were.
Reilly writes:
The usual response to the possibility that homeopathic treatments are eVective is to
call for a mechanismof action asking how? before asking if? is a bad basis for good
science when dealing empirically with things that may as yet evade explanation.
(Reilly et al., 1994).
He then goes on to speculate as to how it is that homeopathy works, as
follows:
Opposite action at low dose a poison becoming an aid is not unfamiliar in
conventional science and is the subject of much debate . . . For todays science,
however, the main barrier to acceptance of homeopathy is the issue of serially vibrated
dilutions that lack any molecules at all of the original substance. Can water or alcohol
of Wxed biochemical composition encode diVering biological information? Using
current metaphors, does the chaos-inducing vibration, central to the production of a
homeopathic dilution, encourage biophysically diVerent fractal-like patterns of the
diluent, critically dependent upon the starting conditions? Theoretical physicists seem
more at ease with such ideas than pharmacologists, considering the possibilities of
isotopic stereodiversity, clathrates, or resonance and coherence within water as
possible modes of transmission, while other workers are exploring the idea of
electromagnetic changes. Nuclear magnetic resonance changes in homeopathic dilu-
tions have been reported and, if reproducible, may be oVering us a glimpse of a future
territory.
Dissemination of the results of research
Chapter Two discussed the numerous problems which arise from trying to
disseminate the results of research, despite the importance of doing so. The
cases discussed in this section illustrate some of the problems. In particular,
they demonstrate the inXuence of the motives of the researchers in their
choice of research question and inevitably in the results they generate. These
results can be useless or even seriously misleading to those who do not share
the views of the researchers, which is often the position of the patients whom
the research was supposed to help.
Results of research into futile treatment depend on what is understood
by futile
Harper (1998) wrote about the issue of research into futile treatment. His
study is fascinating because of the light it sheds upon the relationship
between the intentions and beliefs of the researcher, the research question,
the method used in the research, and the answer gained. He agreed that it was
not good that doctors should oVer experimental treatments more than a very
few times to their patients at the end of their lives, on the shaky grounds that
it might help and nothing else at this stage has. Clearly, any treatment that
might be thought to help should be subjected to proper and controlled
testing to see whether the hunch of the doctor was right or not. But the
research question that is put to such a treatment is all important, not only for
the design and conduct of the trial but also for the use to which the trial data
89 Dissemination of the results of research
can then be put once they have been generated. William Harper cites a
number of trials on treatments routinely given in intensive care, which
concluded that such treatments were futile. Judgements of medical futility
are, he argues, value laden and diVer amongst medical professionals even
when there is agreement on the facts of the case. For example, he asks, is the
administering of cardio-pulmonary resuscitation (CPR), which oVers a child
only three more days of life, still dependent upon intensive care, futile or not?
In his paper, Harper cites a study by Lantos, which concluded that CPR
should not be oVered to low-birth-weight babies routinely because it was in
many cases futile; the outcome measurement in Lantos trial was survival to
hospital discharge. As Harper points out, such a question will be felt to be
more-or-less relevant depending on the individual concerned. Some will feel
that it is appropriate whilst others might feel that even a few days of life are
better than none, and worth the treatment. Harpers concern is not whether
those who would choose to have the extra time are wrong in their views, but
rather that Lantos study could not provide them with information they
would want to know in order to decide whether to allow CPR for their child
or not, unless they studied his method. Lantos study would be helpful only to
those who, eVectively, agree with his initial premise: that treatment which
does not give survival to hospital discharge is futile.
Results of pharmaceutical company research are always commercially
favourable
A similar issue arises within pharmaceutical company research, when the
clear aim of a company is to show that its drugs successfully treat certain
conditions. A trial can be designed to do this, providing the question is right.
Toine Pieters showed that the three pharmaceutical companies which had
most heavily invested in interferon in the 1970s managed, by dint of much
Wshing, to Wnd the right answers to the use of the drug (namely, within a
multitreatment framework). As Pieters put it:
They tried diVerent combinations and diVerent routes and durations of administra-
tion. In doing so, they ultimately tinkered towards success in terms of establishing
newtherapeutic drug practices for interferon and actively working on the treatments
eVectiveness . . . In positioning interferon as a helpful neighbour, compatible with
and supportive of existing treatment practices, the pharmaceutical companies suc-
ceeded in having interferon relatively quickly absorbed into the medical infrastruc-
ture. (Pieters, 1998)
In his discussion of research into futile treatment, Harper identiWes three
reasons why research should not be narrowly focused. First, the decision
about whether or not to let a patient die is one over which informed and
well-intentioned people form diVerent moral judgements. Moreover, it is a
deeply personal decision, which needs adequate information, and the
research done so far does not oVer such information. If the gathering and
dissemination of information is limited to the personal concerns and beliefs
of the researcher, this imposes his judgement as to when a life is worth living
on to those who use the results of his research. In such cases as these, the
pressure to use evidence-based medicine can be positively counter to pa-
tients interests. A second reason for not narrowly deWning research goals is
that a researcher with unpopular value judgements may be frozen out of
funding. The spate of trials into futile treatment in intensive care in the UK
was a response to the Governments concern about the cost of such treatment
in the state-funded health service. If the sponsors of research want a particu-
lar answer to a question, it may well be a matter of simply making sure that
the question is right, then the desired answer may be gained. The third reason
Harper identiWes is that medicine is a social enterprise, and research should
be aimed at meeting public needs, not the wishes of individuals.
Results should ultimately meet public need
Research should not be narrowly focused because medicine is a social enter-
prise, existing at the service of the public. Hence, allowing the personal values
of researchers to choke oV (Harper, 1998) the information the public needs
to make responsible decisions about its own health care is inappropriate. This
third reason has given the spur to the considerable eVorts being made to
include consumer groups in the initiation and design of research protocols.
Chalmers (1995), in particular, has argued that including such groups makes
a big diVerence to the research questions being asked. He cites, as an example
of the usefulness of consumer perspective, the trial on the eYcacy of intensive
monitoring of fetuses during labour. The trial showed that this monitoring
decreased the chances of a baby having seizures. The diVerence the monitor-
ing made was statistically signiWcant, but, to women going through labour,
the encumbrance of being connected to fetal monitoring equipment during
labour was not worth the potential beneWt for the diVerence in the number
of babies with seizures dropped from 998 in one thousand to 996. This study
is a good example of MacRaes point that statistics are simply mathematics,
but the clinically relevant diVerence a treatment makes is a value judgement
which should not be made by a statistician, but by those whose interests are
aVected by the outcome.
This chapter has considered the practical implications of the goal-based
approach to moral thinking under three headings: the goals of research; the
design of research; and the dissemination of the results of research, following
the pattern of the discussion in Chapter Two. SpeciWc examples of questions
and problems in these three areas were looked at. Under goals of research
organ transplantation was considered. The diVerent sorts of investigations
and treatments that have been, and are being, developed for this medical need
were looked at. At each stage, the treatments involve actions which involve
some risk of harm. The assumption that this had to be so was challenged,
wondering whether Weiss view that it should be possible to grow autografts
without creating human embryos could not be realized.
The section on methods in research described the story of the discovery of
penicillin, which had to wait upon the attitudes and beliefs of the clinicians of
the time before its properties were realized. Similar reactions to research into
complementary medicine were then described. The research and the treat-
ments may in fact be unsuccessful, but mere prejudice will not prove that any
more than it will recognize success.
Finally, in discussing the dissemination of results of research, research into
so-called futile treatments, pharmaceutical company research, and research
into fetal monitoring during labour were looked at. In all three cases, the
intentions of the researchers or the ones commissioning the research were
out of line with those of the people who would make use of the results of the
research.
All these examples of issues which arise in the context of goal-based
thinking raise serious questions about how research goals are set and met.
The goals of research arise out of the motives of researchers or their sponsors.
The motivation for undertaking research is varied: there may be simple
curiosity about how something works or whether it is possible to achieve
something, or there may be a clear recognition of a medical need and
potential way of meeting it; there may be commercial motivations, or selWsh
ones. This chapter suggested that the goals of research are lent moral validity
by their being placed within a wider goal of complete cures with no side
eVects. In serving these goals, there need be no fundamental assumptions
about what will and what will not work. There certainly need not be the belief
that there are no medical beneWts without associated harms, even if the
evidence appears that there is no gain without pain. Setting aside assump-
tions of this sort opens the possibility of bold, even quantum leaps in our
knowledge and understanding of medical care. It also facilitates the work of
those who seek to educate the medical profession into using evidence-based
medicine. The intuition and experience which it was argued were an essential
part of a good doctors approach to his patients should, therefore, always be
subject to the possibility of change.
aaaaaa
93
7
Case studies of duty-based issues
Introduction
Chapter Three sought to establish the validity of the duty-based perspective
in moral thinking by showing some of its foundations, namely the tradition
of natural lawethics, and the Kantian categorical imperative. It suggested that
the practical implication for the ethics of research on humans of the duty-
based moral perspective is that, by virtue of their function, the doctor/
researchers have a duty to care for each of their patient/participants, which
they exercise by acting only in their best interests. This duty has no contin-
gent justiWcation. It is absolute, inherent in the function of the doctor. Hence,
a research project might be aiming at a goal which is good, and it might have
the agreement of the research participants who are to help achieve the goal,
but it still needs to be justiWed according to duty-based criteria, namely that it
is in the patients best interests that they participate.
The issues that arise in the context of duty-based morality in research do so
because of clashes with competing goal- and right-based moral claims. The
examples investigated in this chapter will show how this happens. Clashes
with goal-based morality happen frequently in therapeutic research with the
question of whether or not there should be a placebo control with which to
compare the experimental treatment. Goal-based morality, expressed as
scientiWc rigour, demands that new treatments are compared with placebo
unless there are compelling reasons to the contrary. Duty-based morality,
expressed as a patients best interest, demands that no placebo arm be
included if a standard treatment is available. Both points of view are based on
ethical thinking but they do not accord with each other.
In some trials, placebos were omitted on ethical grounds. This is illogical because
studies destined to produce unreliable results should themselves be considered un-
ethical. (Tramer et al., 1998).
This statement demonstrates a muddle between two competing ethical prin-
ciples. One principle is that it is wrong to harmresearch subjects by oVering a
placebo when an active treatment is available (duty-based); the other is that it
is wrong to conduct research which produces uninterpretable results, which
the omission of a placebo arm can do (goal-based). Both principles are valid,
but the two do not necessarily work together. It is not, therefore, illogical to
94 Case studies of duty-based issues
say that placebo controlled trials needed for scientiWc purposes are neverthe-
less unethical, they are just unethical for diVerent reasons from scientiWc
ones. Resolution of the dilemma, if it is possible, will only happen if the two
needs are clearly seen, together with the extent of their claim on our moral
attention. Hence, the arguments for and against the use of placebo will be
considered in some depth, using as examples trials of folic acid in pregnancy,
trials of treatments of peptic ulcer disease, and trials of ondansetron for
post-operative nausea and vomiting.
In non-therapeutic research the clash between goal-based scientiWc con-
cerns and duty-based concerns for adequate care of individuals can arise even
more acutely. By way of demonstration a trial which involved placing babies
in chambers with reduced oxygen to investigate a possible cause of sudden
infant death syndrome will be considered.
The clash between duty-based and right-based moral perspectives shows
itself when potential research subjects are willing to participate in research, so
fulWlling the right-based requirement to consult those most aVected by the
action, but who, according to a duty-based view, should not have been asked
in the Wrst place, or whose request to participate should have been refused,
because of the risks involved. The example considered to illustrate this
diYculty is non-therapeutic research on competent subjects, namely, Phase I,
healthy volunteer pharmaceutical trials, when drugs are given for the Wrst
time to human beings. Here, the research subjects consent is carefully sought
before they are enrolled. The adequacy of the consent procedure is not in
question here, so it is a good example to use to demonstrate the tension that
can genuinely arise between duty-based and right-based morality.
Therapeutic research
Duty to care versus scientic goals: placebo controls in therapeutic
research
As discussed in Chapter Three, the researcher conducting therapeutic re-
search needs to be in strong equipoise for duty-based morality to be honour-
ed. This means her duty to care for each of her patients must be active: she
must act in her patients best interests, and so she must be entirely happy that
the treatment that each of her patients receives in the trial is the best as far as
she knows. It follows that each of the arms of a trial to which her patient may
be allocated must be as good as the others, as far as she can tell. A placebo
arm, where no treatment is on oVer, can only ethically (in the duty-based
sense) be included if no proven, available treatment exists (World Medical
Association, 1996). If a treatment did exist, then the doctor is morally obliged
to oVer it. Not to do so would be equivalent to oVering her patient, who
95 Therapeutic research
presents with an illness in ordinary circumstances, a pink pill instead of a
proper drug, when the proper drug is available. In some cases, the novel and
untested treatment is so likely to be of some use that it is unethical to deny it
to anyone. This was the position of Colebrook with penicillin. In other cases,
patients may have such life-threatening conditions that denying them treat-
ment would be unacceptable.
The goal-based argument for including placebo arms except in unavoid-
able cases, such as where the patients condition is life threatening, is that
without them most trial results are uninterpretable. It is not worth conduc-
ting randomized controlled trials of new treatments if there cannot be a
placebo arm, because the researcher can never otherwise know the extent to
which a new treatment is genuinely causal of an improvement in a condition,
or whether the improvement would have happened without any active
intervention. Active drug comparators cannot provide this information. In
any case they themselves may never have been proved against placebo. If
there is more than one possible active comparator, there is no reliable way of
establishing which of them, and at which dose, can be the gold standard
comparator for the treatment being tested.
Trials of folic acid in pregnancy
The example of the trials into folic acid in pregnancy illustrate the puzzle of
whether or not to use placebo controls (Scott et al., 1994). If, in the early days
of pregnancy, the neural plate of the fetus does not close to form the neural
tube which encompasses the spine and the brain, either spina biWda results, if
the spine is not properly enclosed, or anencephaly, if the brain is not properly
enclosed. It was thought that if the mother took folic acid she could help
prevent the tragedy of neural tube defects. A research project was proposed to
test folic acid against placebo to see if the hypothesis was correct. But the
research ethics committee which had to give the study its approval decided
that it was unethical to deny folic acid to anyone, because there were many
indications that folic acid did indeed make a considerable diVerence to the
development of the fetus. In the event, the control group consisted of those
mothers not in the trial. Interpretation of the results was so diYcult that
another, placebo controlled, trial was set up and conducted by the Medical
Research Council (MRC) to resolve the question. That trial found that if
mothers took 4 mg per day of folic acid the chances of neural tube defects
reduced by three quarters.
It could be argued, from a duty-based perspective, that the fears of the
earlier research ethics committee were realized, for in the MRC trial the
women randomized to the placebo arm were denied treatment which was
found to be extremely eVective. From a goal-based perspective, however, the
MRC research was justiWed by its result, because it demonstrated that folic
acid did indeed make a diVerence, something the earlier, non-randomized
trial had not been able to demonstrate. All the research ethics committee had
done, on this argument, was to delay the truth being discovered, and hence to
expose numerous women, more than the number who were allocated placebo
in the MRC trial, to the risk of pregnancy without folic acid. Moreover, the
value of folic acid is found only if it is taken around the time of conception
and then in the early days of pregnancy (because neural tube defects occur
early). Since numerous pregnancies are not planned, or, if they are planned,
conception does not take place as soon as contraception is withdrawn, it is
diYcult to persuade women of child-bearing potential to take folic acid on a
semi-permanent basis, just in case they fall pregnant. The results of the
research needed to be absolutely certain for the advice to take folic acid to be
worth the eVort. The goal-based justiWcation for the placebo controlled trials
is cast iron. But the duty-based concern is also genuine. There will have been
women whose babies were born with spina biWda or anencephaly, whose
researching clinicians will have been aware of a way, unsubstantiated but
likely to be eVective, of lessening the risk of that happening, but who denied
them the treatment. The researcher in a person might be able to turn a blind
eye to the individual womans plight, in the certainty of the importance of the
research for millions of (other) women in the future. The doctor in the same
person might not be capable of such scientiWc sang froid, unless he was
convinced that folic acid might not do any good at all. The case demonstrates
the conXict.
Trials for treatments of peptic ulcer disease
The debate betweenthose who favour placebo controlled trials and those who
do not is a debate between goal-based thinkers who would always use a
placebo unless there were compelling reasons not to, and duty-based thinkers
who would only use a placebo if there was no alternative. Writing in favour of
the use of placebo controls for trials into treatments of peptic ulcer disease,
Ciociola and others (1996) provide some powerful arguments on the goal-
based side. There are plenty of proven, available treatments for peptic ulcer
disease. Hence, from a duty-based perspective, no trial of new treatments
should have placebo comparators, because to deny people with the disease a
treatment which was available would be to act against their best interests. But,
argues Ciociola, placebo therapy is eVective in 48% to 58% of cases of peptic
ulcer disease. What that means is that giving dummy substances to people
with peptic ulcers has resulted in about half of them getting better. In a trial
without a placebo arm, it will not be possible to know if the eVects that are
seen from the treatments given in the trial are from the content of the pills, or
just fromthe act of taking the pills. It is necessary, then, to have a placebo arm
to establish the baseline remission rate of the disease. Also, the safety data for
the new chemical entity being researched will not be adequate until the
research has taken place. It is ethical, argues Ciociola, to oVer the drug with
its unknown side eVects to as few patients as possible until more is known.
Placebo controlled trials require far fewer participants than active compara-
tor trials. Type I and Type II errors are inversely proportional to the number
of participants needed in the trial.
As explained in Chapter Two, a Type I, or Alpha error, is the probability of
falsely rejecting the null hypothesis, which means that the treatments appear
to be diVerent, but in fact they are the same. For a trial to be considered valid,
it must have a p or probability value of less than 5%. A Type II, or Beta error,
is the probability of failing to reject the null hypothesis when in fact there is a
diVerence between the two treatments. This is regarded as less serious than a
Type I error, and hence the probability is allowed to be greater, 20%, or to put
it another way, a probability of rejecting the null hypothesis when the null
hypothesis is false of 80%.
In a theoretical example, Ciociola shows that a placebo controlled study
needs only 28 patients in each of the active and placebo arms, or on a 2:1
ratio, which is statistically acceptable, 42 patients in the active arm and 21 in
the placebo arm, for the study to avoid both Alpha and Beta errors. If the
study were to be conducted as an active drug comparator study, with no
placebo arm, 408 patients would be needed in each arm to avoid Alpha and
Beta errors.
Ciociola concludes that placebo controlled trials should be conducted for
new peptic ulcer treatments, as long as high-risk patients are eliminated and
supplemental antacids are available to the placebo arm. We can appreciate
that from his perspective, which recognizes the goal-based need for the
research to be done scientiWcally, this is the most ethical approach. But it
could equally well be argued, from a duty-based perspective, that it is worse
to deny treatment to patients than it is to conduct scientiWcally valid studies;
that if it is wrong to give an untested treatment whose safety data is not
adequate to many people, then it is equally wrong to give it even to one
person. If a project needs 816 rather than 56 or 63 patients to be valid then it
should recruit that number, rather than compromise duty-based principles.
Scientic arguments against the use of placebo
The problemin seeking to ensure that ethical researchis conducted is that the
goal-based and duty-based perspectives often do not speak in a mutually
meaningful language to each other in order to Wnd a resolution when they
clash. Goal-based morality uses the language of science; duty-based morality
uses the language of doctors duties. Freedmans analysis of the arguments
for placebo controlled trials (Freedman et al., 1997a, b) may be useful
in this respect. He concludes that placebo arms are justiWed in only four
circumstances, so taking the duty-based line that reasons have to be found for
using placebo controls rather than for not using them. But many of his
arguments are scientiWc. Thus, he meets goal-based thinkers on their own
territory. He thinks that the belief that having a placebo arm provides a
reliable baseline against which to test whether a new treatment works is too
simplistic. It presumes the existence of two sets of causes for a drugs
eVectiveness psychological and biological that act independently, so that
by subtracting one of these sets fromthe groups results, the other remains (as
in Ciociolas justiWcation for a placebo arm to factor out the placebo, or
psychological, eVect). This notion ignores the possible synergistic actions of
the two causes. For example, the psychological expectation of beneWt might
be reinforced by a subjects perception of improvement caused by biological
factors. In this model, the placebo eVect of active drugs would be greater than
the placebo eVect of placebos.
The placebo eVect is well-documented, says Freedman. DiVerent coloured
placebos, for example, have had diVerent eVects on those who take them.
What sort of placebo should be used for a baseline? He suggests that a trial of
a drug which suppresses the appetite which compares the (blue) active drug
with an identical placebo might show the active drug to be superior. But
suppose the active and placebo tablets had been red? It may be that in a red
tablet the placebo eVect would be potentiated and the diVerence between the
drug and the placebo eliminated or rendered non-signiWcant. Freedman cites
Philips paradox: that placebo controlled trials are only successful when they
reveal no advantage for the experimental intervention over placebo; that is,
positive results of placebo controlled trials are uninterpretable because the
more eVective an experimental treatment is, the more likely it is to become
unblinded (obvious to all concerned that it is the active drug) during the
study, rendering it impossible to state whether the observed advantage of the
experimental arm is due to its inherent activity or to the expectancy eVect it
has generated.
The FDAs arguments for requiring placebo
The United States Food and Drug Administration (FDA), which is respon-
sible for licensing all newdrugs for marketing in the US, has stated that it will
require placebo controlled trials to demonstrate a new drugs eYcacy in all
but trials of treatments for life-threatening conditions. This is signiWcant for
medicines research around the world, for all pharmaceutical companies wish
to sell their drugs in the lucrative United States market. Hence, the stipula-
tions of the FDA set the international standard for medicines research. The
FDArequires placebo controlled trials to be the norm. The 1962 amendments
to the Food and Drug Act (Food, Drug and Cosmetic Act, 1938) expanded
the FDAs role to include an evaluation of a drugs eVectiveness by requiring
substantial evidence that the drug will have the eVect it purports. Substan-
tial evidence is deWned in the Act as follows:
Adequate and well-controlled investigations, including clinical investigations, by
experts qualiWed by scientiWc training and experience to evaluate the eVectiveness of
the drug involved, on the basis of which it could fairly and responsibly be concluded
by such experts that the drug will have the eVect it purports or is represented to have
under the conditions of use prescribed, recommended, or suggested in the labelling or
proposed labelling thereof.
The way this requirement has been interpreted is that absolute eVectiveness
has to be demonstrated, that is, that a new drug is better than nothing; rather
than relative eVectiveness, that is, that a new drug is comparable to, or better
than, the standard drugs already available. Many would agree that placebo
controlled trials are justiWable if the standard treatments which would other-
wise be the comparator have not themselves been proved against placebo, but
even when standard treatments have been so proved, the FDA insists that the
comparator for any new drug should still be placebo. Freedman responds
passionately:
To understand eVectiveness in this (absolute) way is absurd; by this measure, Xinging
a glass of water at a burning building is an eVective Wre-Wghting strategy because it is
better than nothing. The public knows that FDA represents a massive, expensive
bureaucracy, needing to fulWl its role as a gatekeeper of new drugs. But how many
health care professionals or lay members of the public are aware that because of its
reliance on placebo controls, the normal operation of FDA may result in the licensure
of a new drug inferior in every way to presently available treatment? And how many
are aware that to fulWl this mandate, FDA will not licence new drugs unless the
manufacturer runs trials that are contrary to numerous authoritative pronounce-
ments on research ethics, both national and international? (Freedman et al., 1997b).
Robert Temple (1996) has been inXuential in maintaining the view within the
FDA that the goal-based scientiWc arguments for placebo controlled trials are
pre-eminent. Like others, he argues that active control trials (where the
comparator is the standard available treatment rather than placebo) present
insuperable statistical diYculties. Freedman disagrees, presenting an alterna-
tive trial design which aims to show active control equivalence. In this design,
two treatments are compared, not to test if one is superior, but to test
whether they are equivalent. The null hypothesis is that the treatments diVer
by more than some deWned amount; the alternate hypothesis is that the two
treatments diVer by less than that deWned amount. The hypotheses are thus a
reversal of the hypotheses in placebo controlled trials. They can be designed
around conventional values of Type I and Type II errors. The sample sizes of
these studies fall between the large numbers needed for active control studies,
and the small numbers needed for placebo controlled studies. Others have
oVered advice on ways to make these trials as reliable as possible (Jones et al.,
1996). However, it must be recognized that such trials are not as statistically
rigorous as placebo controlled trials, and if the goal-based moralist is not
inclined to give any concessions to the duty-based perspective (except to
recognize extreme cases of risk), she is not going to give up her own trial
design in favour of an inferior one. For even though Freedman uses scientiWc
language, he makes allowances for the need for scientiWc rigour only after the
duty-based demand to give patients treatment in their best interests has been
satisWed. His arguments start from a duty-based perspective, whereas those of
Temple and the FDA start with the goal-based need for rigorous science,
making allowances for duty-based ethical demands to treat patients accord-
ing to their best interests. Thus, the two approaches do not really meet in the
middle, despite Freedmans attempts to make them do so.
Meta-analysis of trials of ondansetron
The trials of ondansetronare a practical demonstration of the diYculty as it is
currently being faced by researchers and research ethics committees. Ondan-
setron is a treatment for postoperative nausea and vomiting. When ondan-
setron trials began to take place during the early 1990s, considerable concern
was expressed over whether these trials should be placebo controlled or not,
because of the existence of so many other treatments, none of which,
however, had been established as best practice. As a consequence, of the many
trials of ondansetron which took place, some of them were placebo control-
led, some with and some without an active comparator arm, and some were
active comparator only. Aspinall and Goodman (1995) were concerned not
only that as many as 2620 patients were denied existing drugs in the trials
they reviewed from the published literature, but also that trials which only
showed that ondansetron was better than placebo provided clinicians with
insuYcient information as to whether or not to use ondansetron, since they
were not thereby informed of other drugs to treat the same condition, nor of
their relative eYcacy against ondansetron.
Three years later, Tramer et al. (1998) were to use the trials in ondansetron
to demonstrate that placebo controlled trials are essential. They chose studies
of this drug because placebo controlled trials had been repeatedly questioned,
and because they already had good estimates of ondansetrons anti-emetic
eYcacy and harm postoperatively, from a systematic review of the trials,
against which to compare the active controlled trials.
A review of all the active controlled trials was conducted. Many diVerent
regimens of ondansetron had been compared with many diVerent anti-
emetic controls, demonstrating how much uncertainty there was about
which regimen of ondansetron is the best and which active comparator
should be the gold standard active control. Of the placebo controlled trials,
whatever the dose of ondansetron used, ondansetron showed superiority
over placebo. But of the placebo group, variations of nausea and vomiting
ranged from 10% to 96% up to 48 hours postoperatively. The signiWcance
this result presented for the meaningful interpretation of results was that if
some patients do not vomit postoperatively anyway, then prophylactic anti-
emetic eYcacy cannot be shown; if all patients vomit then it will be exag-
gerated.
Many of the trials showed no diVerence between ondansetron and the
active comparator. This failure to show a diVerence does not necessarily
mean equivalence, however, despite Freedmans claims. In an equivalence
trial, Tramer argues, we need to know the extent of the placebo response, and
that it does not vary. Otherwise, two active treatments showing the same
result could simply mean that both are equally ineVective. One trial showed a
remarkable result: ondansetron was equivalent to placebo but better than the
active drug. This was evidence of internal insensitivity of the trial, and this
was only detectable because there was a placebo arm.
The scattergram Tramer produced, which compared ondansetron with
other drugs, suggested that ondansetron was no better than any of the other
drugs, except possibly metoclopramide. However, because the optimal dose
of metoclopramide is not known, it is not known if patients in the trial were
just given too little of the drug for it to take eVect. The absence of placebo in
the active control trials meant that they could not inform as to the rates of
nausea and vomiting without anti-emetic prophylaxis in these study popula-
tions, nor did they have an index of internal sensitivity. Hence, argues
Tramer, their results cannot be interpreted. The conclusion of these authors
was that placebo is justiWed in ondansetron trials because of the underlying
variation in likelihood of nausea and vomiting and because of the lack of a
gold standard comparator.
These two reasons are goal-based, however. Fromthat perspective, they are
justiWed, adequate reasons for including a placebo arm. But from a duty-
based perspective, a placebo controlled trial of ondansetron means a group of
patients will not receive treatment for postoperative nausea and vomiting
when many commonly-used treatments exist. Many patients suVer from
nausea and vomiting. From the duty-based perspective, perfectly good treat-
ments are being denied to patients in the interests of science and society, as
Aspinall and Goodman pointed out. The goal-based response to this might
be to underline the fact that postoperative nausea and vomiting are not
life-threatening conditions, and the importance of establishing the truth
about the eYcacy of treatments for the condition outweighs the discomfort
that the patients may feel by being denied treatment. The duty-based moralist
will not accept such a view, because it keeps the goal of the research in a
superior position to the individual interests of the research participants.
Indeed, the assertion that it is important to establish the truth about ondan-
setron gives support to the duty-based approach, for it means the condition
which it is hoped to treat is one that should be treated. So, to emphasize the
point, some patients in the trial are being denied treatment for a condition
which should be treated, and which may not kill them but is horrible to
experience, for the sake of science and society.
Concluding remarks
Fromthe range of views identiWed above, we can see that to decide whether or
not a trial should be placebo controlled rests on whether we tend to be more
goal-based or more duty-based in our thinking. The goal-based approach is
to conduct placebo controlled trials wherever possible, only avoiding them
when there is likelihood of real harm. The assumption is that the placebo
controlled trial is the better trial scientiWcally. Arguments which have been
raised against that view are (i) that this entails a too-simplistic interpretation
of the placebo eVect (Freedman, 1997a, b); and (ii) that a placebo controlled
trial only shows a treatments eYcacy against nothing, it does not shows its
merits relative to other treatments (Freedman, 1997a, b; Aspinall and Good-
man, 1995). The latter problem can be overcome by including active arms in
a placebo controlled trial.
The duty-based approach, which will more often conclude that a trial
should not be placebo controlled, would not do so for scientiWc reasons.
Rather, the view is that it is simply wrong to deny treatments, if there are any,
to patients who after all expect to be treated. For it is one thing to establish
that, for scientiWc reasons, a trial needs a placebo arm. It is another for an
individual researcher to be in equipoise about that placebo arm. For the
duty-based approach to be satisWed, the researcher needs to be entirely happy
that her patient will receive a placebo instead of an active treatment. This is
not just a matter for science. The doctor has a series of responses to her
patients which are neither mechanical nor capable of being anticipated. It
may be, in some trials and for some patients, that a doctor will feel herself to
be in equipoise about placebo. That is, she may feel it is as good for her
patient to receive placebo as to receive an active treatment. There is a
detectable placebo eVect, and it may be that by avoiding the active arm the
patient may also be avoiding noxious side eVects. It is certainly true that
patients do get better without active assistance from time to time. But there
may be other cases where the doctor wants to take what Freedmancalls a soft
interpretation of proven, available treatment. Then she may feel that a
treatment which has become standard, even without having been proven
against placebo, should be available to her patients in the trial who do not
receive the experimental active treatment. In this case, active control equival-
ence studies would have to be designed. Collier (1995) argues that the gold
standard against placebo should be established for every new drug, and the
second generation of trials should no longer include placebo, but adopt the
103 Non-therapeutic research
active control equivalence model. These are all possible answers to the
fundamental challenge of therapeutic research on humans, but none is Wnal.
My own conclusion is that duty-based morality, in the form of acting in a
patients best interests, should take priority, because the doctors duty of care
applies in every circumstance, whereas science ought to serve only ethical
goals, and should compromise its rigour to that end. Otherwise, people
become the servants of science inappropriately. The situation where the
ultimate truth has been Wnally established about treatments for any condition
is unlikely to be reached, if not ever, then certainly not for a very long time.
Improvements will, indeed should, always be sought. If it is accepted as the
norm that patients can be denied existing treatment for the sake of discove-
ring the truth about the experimental treatment, that has to be recognized as
a permanent feature of medical research. That, surely, is asking too much of
patients? However, goal-based concerns, as they have shown themselves so
far in this chapter, need to be taken seriously. The moral outlook of the
doctor/researcher should always start with the duty to care for her patients,
but it should stay open to the needs of all future patients too. She should not
be prepared to compromise her own patients needs but neither should she be
ignorant of the need for rigorous research into new treatments. Finally,
deWnitions of harm should guard against being over-cautious.
Non-therapeutic research
Duty to care versus scientic goals: potential risks in non-therapeutic
research
The diYculties of balancing duty to care against science are challenging
enough in therapeutic research, where patients are in trials with therapeutic
intention and can expect their researching doctors to treat them as well as if
they were patients in the ordinary way. Non-therapeutic research provides
even more scope for having to make diYcult decisions.
The question of whether or not to use placebo was a diYcult one because
of the duty-based need to ensure that patients in research projects were
treated as well as possible. In non-therapeutic research, there is no intended
therapeutic beneWt, and there is always some risk of harm. The most diYcult
of all to justify are those studies where the subjects are not able to consent to
take part. When a justiWcation is found, it has to mean that the duty-based
concern to do no harm, and the right-based concern only to act if the person
aVected agrees, have both been subsumed under the goal-based need to
conduct the research. But the goal-based needs are considerable. Research
into conditions which aVect those who cannot consent needs to be done if
treatments are to be developed. Some of that research will, of necessity, be
non-therapeutic.
Research in a vulnerable group: trial into the causes of sudden infant
death syndrome
A controversial study reported in 1998 highlights the moral diYculties
(Parkins et al., 1998a). Sudden infant death syndrome is harrowing and one
of its worst aspects is that its cause is unknown. Researchers at an academic
department of paediatrics became interested in a possible cause when two
sets of parents attended their outpatients clinic reporting a child dying of
sudden infant death syndrome shortly after having been on an intercontinen-
tal Xight. The clinicians thought that a reduction in the partial pressure of
inspired oxygen may increase the risk of apparent life-threatening events and
sudden death in infancy. Airway hypoxia may be caused by respiratory tract
infection or by a change to a higher altitude, such as air travel. In adults acute
airway hypoxia has pronounced eVects on the control of breathing during
sleep. Very little is known about the eVect of airway hypoxia on respiratory
function in infants, but it is thought that the function would be more
vulnerable to harm than adults.
The study exposed 34 infants to 15% oxygen in nitrogen (similar to the air
in the cabin of a commercial aircraft) to discover the eVects of airway hypoxia
on arterial oxygenation and on the frequency of isolated and periodic apnoeic
pauses. Twenty-one of the infants were from an obstetric unit run by a GP
practice (mean age 3.1 months), and 13 were siblings of infants who had died
from sudden infant death syndrome (mean age 1.8 months). Eligibility
criteria included that the infants must have been born at term, and have had
no history of respiratory distress or congenital anomalies.
The infants had an overnight recording of their oxygen saturation and
respiratory variables in their own homes at room temperature and sea level
air pressure. In terms of procedures the infants were put through, this meant
a pulse oximeter (a cuV that is placed over a Wnger) and a volume expansion
capsule placed on the abdominal wall (a tampon strapped to the babys
body). The recordings were analysed to ensure that the infants heart rate and
lung function were within normal limits before the infant was exposed to
15% oxygen. This took place on the second night in an oxygen tent at the
paediatric unit. During this time the baby had the same measurements taken
with the same instruments, and in addition, to monitor respired oxygen and
carbon dioxide, a cannula was placed on the upper lip, which would have
been Wxed in place with a clip. A paediatrician remained with the infants
during the period of exposure to 15% oxygen, and if the infants oxygen
saturation dropped to less than or equal to 80% for more than one minute,
exposure ceased. Where this did happen, parents were advised not to take
their baby on an aeroplane until it was at least 12 months old.
The result of the study showed a highly variable response to acute airway
hypoxia. Some infants became so hypoxic they had to be withdrawn from
exposure. It was not possible, however, from the previous measurements in
room air at sea level air pressure, to predict which of the babies would suVer
this eVect. There may, then, be a relationship between hypoxic conditions
and sudden infant death syndrome. The clinical implications of the study
were recorded by the authors as follows:
We have shown that a small number of infants may become hypoxaemic during
several hours of exposure to a fraction of inspired oxygen of 0.15 to 0.16. We could
not, for ethical and humanitarian reasons, determine whether this would have
progressed to clinically apparent cyanotic episodes [infant turning a greenish-blue] if
exposure had continued. Unfortunately, there was no physiological or clinical variable
in this study which would help identify infants who might develop clinically import-
ant hypoxaemia during later exposure to airway hypoxia. We believe that additional
research is urgently needed into the eVects on infants of prolonged airline Xights or
holidays at high altitude. Our Wndings may contribute to an understanding of the
possible relation between respiratory infection with resulting airway hypoxia and
some sudden deaths in infancy.
The outcome of the research looks important. This study showed reliably that
there was reason to be worried about the eVect of air travel, and laid the
foundations for future research. A goal-based analysis would Wnd that it was
necessary and morally justiWed. From a duty-based perspective, the research
was designed so that the moment any distress was caused to the infant by
virtue of its exposure, it was removed from the exposure. To that extent it
honoured duty-based moral demands. But it is also true to say that it was not
during the airline Xight but afterwards that the infants whose untimely death
had given rise to the researchers hypothesis had in fact died, a point not
picked up in the paper itself but noted by Savulescu who commented on it
(Savulescu, 1998). So although the infants were carefully protected from
harm during the study itself, they could not be protected from the risk of the
consequences of what was done to them. For this important reason, the
duty-based moralist will assert that the research should not have happened,
or at any rate not using an experimental trial design.
It is often argued that the only way to research hypotheses about the bad
consequences of risky behaviour is to use observational methods of research,
for example, into the hypothesis that living near a busy traYc junction causes
asthma in young children, or that smoking cigarettes causes lung cancer,
because it would not be ethical to set up an experimental trial by, say, moving
some children to live near a busy junction and seeing whether they contracted
asthma or not. The fact that there is a concern with the risk of the exposure
means that it can only be measured in an observational model. By contrast,
the babies deliberate exposure to 15% oxygen was an attempt to create the
environment which the researchers believed may have been the cause of
sudden infant death syndrome. The infants were placed in that environment
in order to see what the eVect was.
The goal-based justiWcation for the research is not trivial. Sudden infant
death syndrome remains a mystery and this research has provided a clue
which may help to solve it. Such right-based concerns as could be met, were
met by the parents giving their consent, a matter over which the researchers
clearly took great care. But duty-based concerns about intentionally creating
a risky environment were not treated as overriding. Rather, the goal-based
need for sound science overrode the duty-based need to protect the babies
from deliberate exposure to harm. If a proposal for non-therapeutic research
is found to be acceptable it will be because the duty-based concerns have been
subsumed by the goal-based scientiWc justiWcation. If it is found not to be
acceptable it will be because the duty-based concerns are so strong that they
outweigh the goal-based need to conduct the research altogether. In this case,
in my view, the duty-based concern should have been paramount.
However, the researchers themselves, the research ethics committee which
approved the research, and the editor who published the results, clearly
thought long and hard about the issues. The journal published the paper with
the research Wndings, which itself included a discussion of the ethical issues
(Parkins et al., 1998a), a commentary by Savulescu (1998), a commentary by
the chairman of the research ethics committee which passed the proposal
(Hughes, 1998), and the authors reply to the two commentaries (Parkins et
al., 1998b). Those concerned gave the matter considerable thought, and that
must be respected.
My own Wnal word is that the case supports the view that duty-based
concerns should remain pre-eminent against goal-based ones. If the re-
searchers had started with their duty of care to the infants as the moral basis
for the research, they would have been more likely to think of using an
observational design for the investigations.
Duty to care versus patient autonomy: non-therapeutic healthy
volunteer research indicates the need to protect subjects from harm
even if they consent
Healthy volunteer research is distinguishable by its use of healthy people,
usually young men, as guinea pigs for research purposes. Most often it takes
place within the pharmaceutical industry when new chemical entities are
tried for the Wrst time in humans (Phase I studies). These volunteers will be
paid for their services. They will typically spend some time in Phase I research
units, which, governed by comprehensive guidelines (ABPI, 1991), have
proper facilities for care and resuscitation if necessary. Medical and nursing
care are present and the experiments are conducted according to strict
regulations (ICHGCP, 1997). Giving a new chemical entity to a human being
carries numerous foreseeable and unforeseeable risks. Research will have
been conducted in animals, but animals are not perfect models of humans,
and much will still not be known about the toxicity of the novel entity, nor
howthe human body will digest and excrete it. The purpose of Phase I studies
is to Wnd these things out. Phase I studies do not seek to discover the
therapeutic eYcacy of the new drug. This waits until after the drug is known
to be safe for humans to take.
Very toxic drugs such as chemotherapeutic agents for cancer treatment are
not given to healthy volunteers because the toxicity levels are known to be
high. Such drugs have to be tested in patients straightaway. That presents its
own set of problems, since although the moral framework is clear, that the
hoped-for beneWt should outweigh the predicted risk, both risk and beneWt
are hypothetical, and so the actual decision is very diYcult indeed. It is made
more diYcult by the fact that patients will often have tried every other
treatment option and may accept, in their desperation, anything which holds
out only very tiny chances of helping, even when the risks are very high
indeed (McBride, 1994).
Most new chemical entities, however, start by being studied in healthy
volunteers and these trials are transparently non-therapeutic. Because the
volunteers are healthy adults it is possible to have a very clear understanding
with them that there is no therapeutic intention in the trial and that there will
be some risk. Information sheets can be long and detailed, because there need
be no concern that these participants will be feeling too ill to go through them
carefully and think about what they are being told. Volunteers are also
expected to be honest about their own medical and social history, so that the
researchers can reject any for whom participation in their trial would be
unwise or even medically dangerous. Participants, as was said, are paid.
Healthy volunteer study of eproxindine hydrochloride, 1985
In 1985 shock waves went through the clinical research community and
research ethics committees when the death of a healthy volunteer was
reported (Darragh et al., 1985). The medical examination he had been given
had not picked up, and he had not spoken of, a depot injection of Xupen-
thixol on the day before his death. The study was of eproxindine hydrochlo-
ride, a new anti-arrhythmic agent. The study was designed to investigate
tolerance of humans to single, rising, intravenous doses of the drug given by
infusion over Wve minutes. One hundred and ninety-two doses had already
been given over a period of seven months to healthy young men using the
same routes, and no adverse reactions had been seen. This young man,
however, had a sudden cardiorespiratory arrest and died. Afterwards it
became known fromnon-medical sources that he had attended a psychiatric
outpatient clinic where he had received an injection, which was, as subse-
quent blood tests showed, of Xupenthixol. Although the precise mechanism
for the interaction between the study drug and Xupenthixol is not known,
both are basic drugs and are highly protein bound.
The young man, together with all the other participants, had been told that
all medicines, illicit social drugs, and alcohol were prohibited throughout the
study. At no time did the volunteer who died admit to receiving any
medication, nor did he disclose information relating to his recent medical
history. He had Wrst attended the clinic as a healthy volunteer six years
previously, and had participated in about two studies per year since then. He
had been independently examined by eight diVerent physicians, none of
whom had detected the condition and treatment which would have excluded
him from being a participant in the trials at the clinic. As the authors of the
report point out, the case highlights the diYculties of screening out volun-
teers. Without full disclosure of the medical history, most comprehensive
medical examinations and tests will be inadequate.
Kidology
A story told in the British Medical Journal in 1998 rather nicely illustrates the
need to guard against over-eager and mendacious healthy volunteers, who
assume that, being in the hands of doctors and research ethics committees,
they will not come to any harm.
The train journey, the pile of papers, and the BMJ. The man opposite catches my eye.
I see youre a doctor.
Self aiming at closure: Yes, Ive a lot of work to get through.
Im on my way back from London. I was in Harley Street being interviewed as a
volunteer for a drug trial of antidepressants. I had to use some kidology.
Yes?
Im not depressed but they pay better in London than they do in Scotland or
Manchester but I dont think I succeeded. The doctor advised me to use cognitive
therapy. They said they would pay my expenses by cheque but eventually agreed to
give me cash.
Oh?
Yes, the trials are advertised, the best pay about 100 a day to volunteers. For a 20
day trial thats 2000. The worst trial was when I had to be woken up every hour to do
mental tests, but usually its like being on a health farm.
What about making sure you dont come to harm?
Oh, they have a committee of vicars and lawyers to decide its all right, and its nice
to see your regular friends.
My train drew into the station; I was no longer irritated at the interruption. (Boyd,
1998)
Audit of volunteer screening procedures
An audit of the screening procedures for healthy volunteers reported in 1992
showed the importance of obtaining reports on the volunteers fromtheir GPs
(Watson and Wyld, 1992). Information so received revealed important facts
not disclosed during the medical history or examination of the healthy
volunteers themselves. The audit was of the rejection rate and stage at which
rejection took place in 831 subjects all volunteering for the same study. The
screening procedure at this unit was in three stages: Wrst, the volunteers Wlled
out a questionnaire with personal, social and medical details; at this stage 117
of the 831 were rejected. Second, a medical history was taken, and a physical
examination performed with clinical chemistry and haematology investiga-
tions and serology for Hepatitis B and HIV. A urine sample was tested for
protein, glucose, blood and bilirubin, and also for drugs of abuse. At this
stage 45 volunteers were rejected. Stage three consisted of obtaining informa-
tion from the GP of each of the volunteers. The questionnaire the GP was
asked to complete asked about disorders of the central nervous system,
cardiovascular disorders, disorders of the respiratory system, disorders and
diseases of the alimentary system and the genito-urinary system, psychiatric
disorders, alcohol and/or drug abuse, and any other disorders such as
diabetes mellitus. The GP was asked about any prescribed drugs and evidence
of adverse drug reactions, and about the involvement of his patient in any
previous research studies within the last six months. The GP was asked if
there was any reason he felt the volunteer should not participate in the
research. Finally, he was asked how long his records went back on the
volunteer in question. At this stage 44 more volunteers were rejected.
On the basis of information from GPs, four volunteers were rejected for
depressive illness, 18 for alcohol dependency, four for drug dependency, and
six for neurological disorders (including one case of epilepsy and four of
recent head injuries, two of whom had been prescribed prophylactic anti-
convulsant drugs). The understated conclusion of the authors is: Application
to the GP for medical information on potential volunteers is an important
step in the screening of volunteer subjects for clinical research.
Does payment to research subjects make a moral difference?
In healthy volunteer studies, the research participants are in a much better
position than patients for the purposes of obtaining consent. They are
healthy and competent, and there is no need to rush the consent procedure.
We could think that the duty-based approach becomes redundant in these
trials, for if the science is good (goal-based) and the subjects are fully
consenting, that is to say, are perfectly well aware of any risk (right-based),
there are no other moral concerns. However, the mendacity about social or
medical habits indicated in the audit of volunteer screening procedures was
almost certainly due to the lure of money, which aVected the exercising of
their judgement, thus raising a question mark over the validity of the
right-based criterion being met. It can be assumed that the reason individuals
choose to take part in healthy volunteer studies is that they would like the
remuneration oVered. But in many other circumstances individuals are paid
for doing risky jobs, such as city couriers, stuntmen and women, and
security guards. Is there a diVerence between those situations and that of the
healthy volunteer who knowingly takes a risk for money?
The additional factors in the medical research context that may make a
diVerence are the presence of healthcare professionals and the reference to a
research ethics committee. The potential healthy volunteer may or may not
be aware of the latter (the depressive manque was) but he will certainly know
about the former because they are conducting the trial. He may assume that
the researcher will not do anything that will harm him, even if he is told that
she will. To this extent the research must honour the duty-based moral
requirement not to be too risky. It would be wrong to propose risky research
and let the prospect of payment lure the healthy volunteer to accept those
risks when, as I say, he will have a belief that a doctor would not deliberately
expose him to harm. But only the volunteer himself can be held responsible
for lying about social and medical habits that will expose him to risks if he
continues with them and also enters the trial. The lure of money may
encourage mendacity but for that no-one can or should be held accountable
but the volunteer himself.
Women of child-bearing potential as healthy volunteers
Until recently, women of child-bearing potential, that is, healthy young
women, were routinely excluded from Phase I studies because of the risk to
the fetus if they were by some chance pregnant. Hence, healthy volunteer
studies were almost always on young men. This reXected a duty-based
concern. It meant that data from Phase I studies were always based upon
studying male participants. The diVerence, of course, is that whereas a man is
taking a risk just with his own body, the woman will be taking a risk both with
her own body and with the fetus. However, it seems inconsistent not to rely
on a woman knowing that she is pregnant and disclosing that information for
the fetus safety, when researchers have to rely on a mans honesty about his
drinking and other social habits.
In this chapter some examples have been considered where the duty-based
considerations of acting in the patients best interests weigh suYciently in the
balance as to force either goal-based or right-based assumptions to be
reconsidered. The primary goal-based assumption brought into question is
that it is of overriding importance that the science of a research project is as
pure as possible. The primary right-based assumption is that if people have
been fully informed of what they are to consent to, are competent, and have
given their voluntary agreement, then they have a right to go forward to be a
research participant.
In considering the issue of scientiWc rigour the use of placebo controls in
therapeutic research was discussed. The arguments on both sides were
investigated, seeing how a goal-based thinker, such as Ciociola, tries to show
that the scientiWcally rigorous trial is also in accordance with a duty-based
ethic, because patients would come to no real harm and could be excluded if
there was any possibility that they would. We saw how a duty-based thinker,
such as Freedman, tries to show that trials which honour the duty of care to
patients can also be scientiWcally rigorous. Examples of trials in which the use
of placebo was contentious were looked at: folic acid in pregnancy; trials of
peptic ulcer disease; and ondansetron trials. It was noted in conclusion that
whilst goal-based and duty-based approaches might reach out to each other,
they do not, in fact, meet in the middle. It was suggested that for this reason it
was important to put the duty-based duty of care in primary place, but at the
same time remain open to the need to conduct scientiWcally rigorous trials,
and then let each case stand on its own merits.
A more worrying example of a trial in which the goal-based considerations
overrode duty-based ones was then described and analysed. This was the trial
of babies, exposed to the same atmospheric conditions of an inter-continen-
tal aeroplane. Although the immediate risk was not great, and indeed
mimicked atmospheric conditions into which babies are introduced every
time their parents Xy with them, the conditions were thought to contribute to
sudden infant death syndrome, so the longer-termrisk was considerable, and
unknown. The researchers created an ideal experimental situation, which
produced a scientiWcally satisfactory result, when they should have used the
less rigorous observational model of research which is employed when
investigating other sorts of risky behaviour, such as smoking. Goal-based
considerations had overridden duty-based ones in an inappropriate way. If
the researchers had started from the duty-based foundation of not harming
their research participants, rather than starting with the need for scientiWc
rigour and considering the safety of the babies only secondarily (with how-
ever much care), they would have been more likely to propose a diVerent trial
design, such as an observational one. I am not suggesting that the researchers
did not take the risks to the babies suYciently seriously, because clearly they
thought long and hard about the issue. What I am suggesting is that the
notion of not harming was not the prior consideration. It came after the
research was designed. From a duty-based starting point, other research
design possibilities might have opened up, as has been discussed. Without
such a basis, they were not even mentioned.
Finally, an example of duty-based morality clashing with right-based
morality was considered. The evidence from the audit of screening potential
healthy volunteers through their doctors was that to rely on a persons
self-selectionin this situation will mean that some individuals will go forward
as volunteers who should not. On the presumption that the individuals
volunteer because of the money involved, such remunerationshould never be
for undergoing risk and researchers should live up to the expectation that
they would not deliberately cause harm. However, mendacity about social
habits cannot be the responsibility of anyone but the liar, even if he is being
paid.
The discussion in this chapter has focused on the duty-based issue of the
doctor/researchers duty of care to her research subjects, which demands that
she acts in their best interests. This need comes into conXict with both
goal-based issues of scientiWc validity, and right-based issues of patient
autonomy. In seeking to come to a reasonable conclusion when faced with
this conXict it was proposed that the duty-based perspective remains pre-
eminent, in the sense that that is where each doctor should start. However,
the goal-based moral demands are suYciently strong to warrant serious
attention from every doctor. Therefore, while staying with her duty to care
for each of her patients, each doctor must remain open to the need to be
continually developing new treatments and improving old ones. The doctor
is concerned for the patient she is currently treating, and for all those in the
future who will need care. The danger of the duty-based approach being
arrogant, because of the doctors necessary assumption of knowing what is in
her patients best interests, is tempered by the right-based demand that she
seek to know and respect the views of her potential research participants.
Whilst she might need to question a healthy volunteers eagerness to be her
research participant, she should certainly not question his refusal. It is to
issues of consent as they have arisen in particular cases that we now turn.
113
8
Case studies of right-based issues
Introduction
This chapter will look at the practical application of right-based morality in
research on humans, which is expressed as the need to obtain consent and to
respect conWdentiality. These are the two primary moral demands of right-
based morality, as I have deWned it. By consent is meant the adequately
informed, voluntary agreement of a competent person to participate in
research. By conWdentiality is meant the non-disclosure of private informa-
tion which has been given in a medical context, where no consent to
disclosure has been given or can be assumed.
When consent and conWdentiality were discussed at a theoretical level it
was argued that they were essential moral requirements. In the context of
particular cases, goal-based demands for rigorous science and duty-based
demands to act in a patients best interests also weigh in the balance. I
propose to consider the right-based requirement for consent in the face of
goal-, duty- and, paradoxically, right-based obstacles to its being successfully
obtained. Empirical studies of the consent procedure for these purposes will
be concentrated on. The extent to which the use of consent forms can be
regarded as helping or hindering the consent process will also be considered.
Under conWdentiality one argument against respecting it, which arises from
the goal-based need to conduct records-based research, will be considered.
Consent
Consent to participation in research is demanded by the right-based prin-
ciple, that no person should be treated merely as a means to an end, but
always also as an end in herself. The idea that anyone is merely expendable is
abhorrent; it has given rise to such atrocities as those committed by the Nazis
in concentration camps, or by the injustice of apartheid in South Africa. That
anyone should be expendable for the good of others or in the service of a
greater good (as in being a research participant) does not make it any more
right. The Nazis and the government of South Africa thought they were
acting for a greater good. It does not follow, however, that no-one should be
used in the service of others. The signiWcant additional factor that makes the
114 Case studies of right-based issues
use morally acceptable is that each person agrees that she is so used. It is one
thing to be forced to serve others and quite another willingly to do so. The
latter is to be highly recommended. In medical research this means that
consent should always be sought.
Nevertheless, there are numerous ways in which the consent process can be
undermined in the context of research. Some of these attacks are eVective. To
illustrate this, right-based, duty-based and goal-based constraints on the
consent process will be looked at, starting with right-based concerns because
they could be said to be among the most diYcult to answer. Duty-based and
goal-based concerns, whilst compelling, could be overruled by using right-
based morality as a trump card. Thus, the response would be given that no
matter, for example, that the giving of information to a person in order to
seek her consent will cause her distress (a duty-based concern), it is neverthe-
less more important that she is informed and makes up her own mind.
Alternatively, the response could be, if the obtaining of consent means that
important research will be severely curtailed because of low recruitment (a
goal-based concern), that again, it is still more important that people are not
used merely as a means to an end. Right-based morality is the basis of the
demand that consent be sought, so if it presents diYculties in obtaining
consent, these cannot be answered so easily.
Right-based difculties with consent: the empirical evidence
There are numerous opinions about what the consent procedure needs to
consist of if it is to succeed. The relentless disagreement between diVerent
research ethics committees and between research ethics committees and
researchers on this point alone is suYcient witness to this fact. Rather than
rehearse the disagreements, some empirical studies which highlight some of
the diYculties will be looked at. There are clearly more studies of this area
than there is room to discuss here, but a good spread of studies will be
referred to, hopefully giving a fair picture.
Patients felt they were well-informed
First, there is a positive result from a study conducted some time ago
(Goodman et al., 1984). This research surveyed patients from two trials
which were studying the ventilatory eVects of postoperative analgesia. There
were 18 patients in each of the trials. Fourteen out of the 18 patients in the
Wrst study returned questionnaires; 18 out of 18 returned questionnaires in
the second study. In the Wrst study the questionnaire was sent to all 18
patients who had completed the postoperative phase of the study. In the
second study the questionnaire was given to patients after the operation but
before they were discharged from hospital.
The results were promising for the success of the consent procedure: of the
115 Consent
32 respondents, 30 thought the explanation prior to their giving consent was
satisfactory. Eleven had felt obliged to participate in the trial, but not from
any feeling of coercion, but rather from a sense of duty to others. Of the eight
who found discomfort from the operation, only one felt that it was worse
than the prior explanation had given him to believe.
Kings review of the literature
Other studies, however, have not been so positive in their Wndings. Jennifer
King (1986) undertook a review of the empirical research on the consent
procedure; her Wndings are mixed, to say the least. One observation from the
review was that patients needs and wishes for information are generally
greater than many doctors realize or cater for, although, as King points out,
the degree of information wanted will inevitably vary with the individual
patient and the condition she suVers from. She quotes one researchers
conclusion: To tell or not to tell is not the issue, but rather what information
is appropriate for a particular patient (Hoy, 1985).
Three studies in the review showed a failure of patients properly to
understand what they had been told (Garnham, 1975; Leeb et al., 1976;
Cassileth et al., 1980). However, the method in these studies relied on
patients recall of what they had been told, in some cases after a few months
had elapsed. This is not really a measure of the patients understanding at the
time: there are plenty of situations in which we take in information, under-
stand it and act upon it, and then forget about it soon afterwards. Four other
studies showed more success in helping patients understanding when certain
strategies and communication techniques were adopted to improve informa-
tion giving (Ley, 1979; Woodward, 1979; Holtzman et al., 1983; Simes et al.,
1986).
Other studies reviewed by King show duty- and goal-based problems,
which will be turned to under those headings. Obviously, a number of studies
have been conducted since 1986, and some of these will be looked at now.
Volunteers tend to be less well-educated
In 1990, a sad but unsurprising Wnding was reported in Australia: that parents
who volunteer their children for research are less well-educated, and hold
fewer administrative or professional jobs than those who do not. Their
motivation is to help others, but it also comes from a desire to gain more
information and to help their own children (Harth and Thong, 1990).
Patients do not understand information sheets
Another study in 1990 was even more depressing, giving reasons to believe
that patients do not understand what they are told, no matter how hard
clinicians try to make written information comprehensible (Sutherland et al.,
1990). An information sheet and consent form for a hypothetical trial was
given to patients. The inclusion criteria were: all patients had been admitted
to one hospital requiring postsurgical treatment for cancer; they all knewthey
were being treated for cancer; they all understood written and spoken
English; they all gave consent to participate. The study excluded patients who
had previously consented or refused to enter a clinical trial, or who had
undergone chemotherapy. Patients were asked to underline the information
on the sheet which was pertinent to their decision to take part in the
hypothetical study. Seventy-four per cent did not indicate that both risks and
beneWts were pertinent. Twenty out of 50 refused to take part in the study; for
70% of those, only information about risks was pertinent to their decision.
Thirty out of 50 agreed; for 33% of those, the only information pertinent to
themwas about risks, while 10% noted neither beneWts nor risks. This part of
the results indicates only that diVerent patients are moved by or want to
know about diVerent information, a Wnding already established. However,
the researchers also gave patients a multiple choice test of the meaning of four
of the statements in the information sheet. Between 26% and 54% of the
answers, depending on the statement, were wrong. One of the statements was
a particular type of cancer responds to radiation treatment in 10% of cases.
Forty-six per cent of respondents thought, correctly, that that meant: on
average, for 10 out of every 100 patients, the tumour will decrease in size after
radiation treatment. But 22% thought it meant: the radiation treatment is
about 10% eVective in an individual patient. Another 22% thought it meant:
10% of patients are cured. Ten per cent thought it meant: there is a 10%
chance of survival. Another of the statements read: a process called random-
ization is used to select your treatment in this clinical trial. Sixty-eight per
cent, correctly, thought that meant: each individual patient has exactly the
same chance of receiving the drug, or not receiving the drug, as any other
participating patient. But 14% thought it meant: the process will select the
best treatment for me, and 18% thought it meant: the doctor decides which
treatment is the right one for me. The lack of understanding of randomiz-
ation is supported by Wndings in later studies which I shall come to shortly.
Benets remembered more than risks
A report in the British Medical Journal of a study in Chicago indicated that
cancer patients enrolled in a Phase I trial do not take in much of the
information given to them prior to agreeing to take part. Only one third
correctly thought the study was of dose escalation; the other two thirds did
not know or thought it was therapeutic. The British Medical Journal, citing
the researcher who conducted the survey, observed:
Clearly, despite attempts at explanations, patients motivation to get some therapeutic
beneWt is so overwhelming that it obliterates the memories of much of what they have
been told about the study. (McBride, 1994)
117 Consent
This worrying Wnding is partly explained by the research of Slevin and others
(1990) which found that people who actually had cancer are much more
likely to opt for radical treatment with minimal chance of beneWt than people
who do not have cancer, including medical and nursing professionals. In
1994, Chee Saw and others found that patients were more likely to remember
the beneWts they had been told about than the risks (Chee Saw et al., 1994).
The information was about the transurethral resection of the prostate which
patients were about to undergo: the risk of retrograde ejaculation was stressed
when the information was given, and yet 18% of patients did not remember
that this was one of the risks, when they were asked. This selective remember-
ing of only the better news was found in another study (McDaniel Hutson
and Blaha, 1991).
Randomization not really understood: the ECMO study
Qualitative methods were used (Snowdon et al., 1997) to explore in detail
parental reactions to randomallocation of treatment in an important nation-
al neonatal trial: the UK collaborative trial of extra-corporeal membrane
oxygenation (ECMO). The trial involved mature newborn babies with acute,
but potentially reversible, respiratory failure, comparing conventional man-
agement, that is, maximal ventilatory support, with a novel treatment,
ECMO, which involved oxygenation of the blood via an external circuit. The
ECMO treatment meant the baby would be attached to the ECMO circuit by
an operation to insert cannulae into blood vessels in the neck. The treatment
also often involved a journey by air ambulance since there were only Wve
centres in the UK with the equipment. ECMO had been widely adopted
elsewhere in the world, though not subject to properly controlled trials. In
Britain, by dint of careful collaboration, the trial of ECMO involved paediat-
ric units right across the country, and ECMO was only available in the trial.
The research by Snowdon and others explored parents understanding of
randomization. Three observations should be made about the study before
looking at those Wndings which have implications for right-based problems
with consent. First, the parents concerned were in an extremely stressful
situation with a newborn baby who was critically ill and needed treatment
quickly. Second, although parents were given letters and information leaXets
about the trial which had been prepared by the trial steering committee, some
centres were required by local researchethics committees to amend or replace
the trial documentation, and in some instances to use locally idiosyncratic
consent forms. Additional information would have been given to parents at
the discretion of the participating centres staV, since they were asked to
expand on the written information. Hence, the basic information given to
parents was not uniform. Third, there was widespread belief amongst parents
that ECMO was the superior treatment.
For 12 of the 21 pairs of parents interviewed, at least one of the parents
understood the nature of randomization. Of those who did not, some sawthe
treatment decision as therapeutic:
Andrea based her understanding on information she was given about her sons
eligibility for the trial. When the possibility of the trial was raised with Andrea and
Robert, Daniel did not quite fulWl the entry criteria, but would do so if he deteriorated
further. She found this diYcult to explain and eventually concluded that, although
Daniel was dying, he still needed to decline to a stage where conventional therapy
would be of no assistance.
When asked about their perceptions of how randomization was actually
carried out, most parents were aware that a computer was involved; some
thought it was to make a more sophisticated treatment decision, others to
generate a simple binary decision. One parent, interestingly, described the
process of minimization (a reWned version of randomization in which the
computer will ensure the groups are comparable according to pre-set data,
which was used in this trial) but without describing the random element. The
woman said:
It would be randomly selection fromlike a computer . . . We assumed that what would
happen is . . . that the information . . . would have been fed through the computer and
that I assume that if they had em already had information on a baby girl, round eight
pound to nine pound weight, and they had already had on ECMO trial, then they
would do the other . . . I thought that perhaps the reason we didnt go onto the ECMO
trial was because they already had somebody, or previously had somebody, same
weight, same sex, similar problem.
The authors report that by contrast to this account, other parents said they
had received very simple information, such as that the childs name only went
into the computer:
How they always put it was the button would be pressed and if his name came up he
would be gone [i.e. transferred for ECMO].
The authors observe that the comment reXected a belief that several babies
were competing for ECMO places; an understanding of the random decision
but a distortion of the reasons for making the decision randomly. Another
couple were given conXicting information: that only the childs name would
be fed into the computer; and that the childs details would be fed in.
When asked about the reasons for randomization, only four parents
suggested methodological reasons, for example the need in research for
controls. This was typical of how parents viewed the conventional treatment:
You canna really do a trial if you put them all on to it . . . Youve got to leave some to
compare it against.
Most parents described ethical reasons for randomization in a more-or-less
muddled way. A number saw it as a way of not asking the doctors to play God
119 Consent
in allocating a limited resource. Robert, whose son had gone into the ECMO
arm, said:
I just felt there could have been ten other babies with exactly the same problems as
him and now there is nine sets of parents who are now being told that their babys not
being accepted onto the trial. And I did feel a bout of guilt for that but I could have
gone out and . . . danced on water . . . when I got told that hed been accepted.
Most parents were uncomfortable with randomization and only one said it
was important to test properly the new treatment; only one other suggested
it was necessary to randomize in order to assess ECMO for associated
hazards. A number regarded the process as unfair or tough and heartless:
I suppose trials have to be a bit heartless, but youd think that when the baby looks like
theyre dying, youd think theyd just say . . . Oh hell you know lets try the ECMO,
see if it saves this baby, but with that sort of a trial they cant do that can they? They
have to say, Well look, this baby looks like its dying but Im sorry its getting
conventional treatment and thats that.
Many accounts of parents convey a sense that randomization was a trivial
way to make a choice. A doctor should be expected to be more careful and
considered in how he decided.
Robert felt that ECMO was a proven treatment and so should not be
restricted. When he was asked what he was told about how the treatment
would be chosen:
Yes, by randomisation and that that annoyed me. It didnt annoy me at the time
because we got it [ECMO] but since then the very thing thats stuck in my mind all the
time is who gives them the right to play God with babies lives? And why the hell have
we got it on trial when its been in the States and its got an 89% success rate or
whatever . . . ? Why is the National Health playing around with this? You know they
wouldnt play around if America suddenly came up with a cure for cancer . . . Why are
they playing around with babies lives? Er, you know who gives them the right to sit
here with say 10 babies and think well this this one here you know will will suit the
trial, you know. Why not all 10 of them? Why isnt it available everywhere so
everybody has a fair chance?
This study gives an insight into the meaning of randomization to lay people.
Randomization is not generally very well understood, and people do not
appreciate that the doctors may genuinely not know which is the better
treatment, or indeed that ECMO was not proven to be eVective. The wide-
spread belief, clearly shown in these responses, that ECMO was the superior
treatment, would have made a big impact on the ability or willingness of the
parents to comprehend randomization. Whether or not there was strong
equipoise in the doctors minds, there was clearly no such balance of uncer-
tainty in the minds of the parents. This is a good demonstration of the
problems with communication that the consent procedure has to overcome if
it is to be more than a semblance of the real thing.
There are further, related questions to ponder. Does it matter that the
parents may well have agreed to enrol their babies into the ECMO trial not
because of the uncertainty of which of the two treatments was better, but
because it was the only way their baby had any chance at all of receiving
ECMO, since it was not oVered outside the trial? If ECMO had been available
without having to enrol in the trial, would the vast majority of parents have
refused to enter their babies for the trial and opted for ECMO instead? If so,
the ECMO trial would never have been able to happen and the superiority of
ECMO would not have been established with scientiWc rigour. This is a
goal-based imperative. By only being oVered ECMO in the trial, however,
parents were not really making a voluntary choice about enrolling their
babies, they were being coerced on the basis of their beliefs about ECMO.
Does the fact that they may have been wrong about ECMOs superiority
make the coercion morally acceptable? That is to say, the parents may have
felt an obligation to enter their babies into the trial because that was the only
way they could be sure they were doing the best they could for their child, and
therefore their consent was not really valid, because it felt to them that they
had no choice. But because the basis for their sense of obligation was
objectively wrong (that ECMO was indubitably the superior treatment), the
parents actually made the right decision, albeit for the wrong reason, because
no one was worse oV on the basis of it (the babies best interests were served,
which is the duty-based imperative) and the future surgical care of babies in
their condition was helped (the goal-based imperative). But what then of the
right-based imperative that we should be of service to others only by our free
will?
It seems to me that this conundrum is a fact of life in many circumstances
other than medical research on humans. My reasons for agreeing to certain
ways of being treated are always going to be mixed and based upon a less than
perfect understanding of the facts. This is not a trivial observation, because in
ethics, motive matters a great deal. We do not think morality is well served if
we try to justify lack of understanding by asserting that, actually, our decision
is objectively right. But motive can be separated from understanding. That is
to say, the desire of the parents to do the best for their babies, regardless of
how they came to understand how to do that, can be applauded. Indeed, had
ECMO been available outside the trial, given the state of their beliefs about it,
it would have been questionable if they had put their babies into the trial. By
extension, if the research project was one where consent was not given
vicariously but on a persons own behalf, the motive to help others could be
applauded if it played a part in the decision of the individual to enter the trial.
121 Consent
Duty-based difculties with consent
The theoretical diYculty that duty-based morality poses for consent is that it
causes harm. In the above example of the ECMO study, the duty-based
moralist might argue that none of these diYculties would have arisen if the
parents had simply not been told that their babies were in a trial, or if they
had been pre-randomized, with only those who had been randomized to
ECMO being asked if they would like their child to be in the study. This
circumvents the diYculty; it does not solve it, and is anathema to the
right-based moralist whose response to the misconceptions of the parents is
to try and improve communication with them, certainly not to conclude that
consent should therefore not be sought. But there are compelling reasons for
the duty-based approach to be taken seriously, even if, in the end, it is
rejected.
Empirical evidence of whether information increases anxiety
One area that has been the subject of study is whether giving information
increases anxiety, and what is the psychological/physiological eVect of giving
information to patients for the purposes of seeking their consent. Kings 1986
review looked at some studies of this. Her conclusion was that it depends
upon the patient, the condition, and the way in which information is given.
There is certainly some evidence that information about side eVects of
treatment can lead to those side eVects being experienced (e.g. Cairns et al.,
1985). There is also evidence, however, that giving information can improve
patients recovery (Wallace, 1984, 1986). Kerrigan and others (1993) ob-
served that detailed information did not increase patient anxiety in men
undergoing elective inguinal hernia repair. However, in Chee Saw et al.s
study (1994) referred to above, 54% of respondents did not want detailed
explanations, trusting in the doctor to give them the right treatment.
Patients consent should not automatically be sought in therapeutic
research
An article by Tobias and Souhami (1993), which was, in duty-based fashion,
entitled Informed consent can be needlessly cruel, cites arguments and
evidence that often slip into goal-based reasoning (so the article might have
been called: Informed consent prevents good clinical research), which will
be examined below. They do, however, make some duty-based points about
the eVect of the consent process on the patient. Their most signiWcant point is
that the double standard that currently exists between therapeutic research
and treatment should go, and that the decision whether or not to inform a
patient about what is happening to her should be based, in both situations,
on her psychological and physiological state. The underlying premise is that
the clinical trial is as therapeutic for the patient as treatment in the ordinary
way, since the trial is comparing new and standard treatments, and patients
have a 50% chance of receiving the better treatment. The doctor does not
know which it is, and therefore it is in accordance with his duty of care to his
patient to put her in the trial. If in his view the patient would be unduly
distressed by information, then, as he would if he were oVering treatment in
the ordinary way, he should withhold information, but still enrol her in the
trial, because that is doing what he believes is best for her. Tobias and
Souhami think that patients, when they are told what is wrong with them,
particularly when it is a condition as frightening as cancer, want reassurance
and a clear prognosis, not to be told that their doctor does not know what is
best. They point out that in the consultation and discussion of the trial the
doctor may describe the advantages of the experimental treatment, only to
have to back peddle if the patient is then randomized to the conventional
therapy arm. Although the right-based moralist would argue that it was the
doctors fault for enthusing about the novel treatment when he was supposed
to be in equipoise, we can see from the ECMOtrial Wndings and the Chee Saw
study that it is not so easy to convey a balanced picture. Tobias and Souhami
pick up the point that is demonstrated in the ECMO Wndings, that patients
are disturbed by the knowledge that a computer, not a doctor, will be making
the decision about their treatment.
Problems with this argument
Right-based moralists would argue that people need to be informed, ought to
be informed, and if that is at the price of their peace of mind then that is a
price that must be paid. There is a diVerence between the ordinary thera-
peutic situation and that of the clinical trial, from the point of view of the
patient, because in the ordinary situation (as the patient believes) the doctor
knows what to do for her. In the context of research the doctor does not
know, and the patient is helping him to Wnd out for the sake of future
patients. Were she to Wnd out later that this is what had happened without
her knowledge or consent, she would not be best pleased. Nor would she be
inclined to trust her doctor in the future. From the patients perspective, she
has been deceived, and she has also been made to be of service to others
without her consent. Tobias, Souhami, Chalmers (personal communication,
2000), Baum (1986, 1993) and others will argue that in the ordinary clinical
situation the patient is also deceived because the doctor does not recognize
that there are other treatment options. So there is deceit in both circumstan-
ces, but at least in the context of a trial some good is being achieved for future
patients. The right-based perspective looks from the patients eyes, however,
and from here, the trial and the treatment situations are diVerent because in
the former the doctor would knowingly deceive the patient if he did not seek
her consent, whereas in the latter the doctor is also in the dark and no
123 Consent
deliberate deceit has been practised. The issue is, again, one of motive rather
than objective fact, or what works out for the best in the end.
Lord Scarmans prudent patient test
Lord Scarman, who was one of the judges involved in the Sidaway v Bethlem
Royal Hospital Governors (Sidaway, 1985), commented later:
Even before Sidaway I was troubled and I remain nowvery puzzled as to the ethical
and legal implications of the so-called randomised clinical trial, where, ex hypothesi,
the doctor does not know, and is using, to put it baldly, his patient for the purposes
of very worthwhile experimentation. . . As the lawstands at the moment, I would have
thought that any doctor who allows his patient to go into a randomised clinical trial
without telling him runs a very real risk if things go wrong and the patient suVers
injury or damage . . . One can see the value of this form of experimentation . . . but I
am bound to say that I think there is a danger for the doctor in going ahead with
subjecting, if that is the correct word, his patient to a randomised clinical trial without
warning him. (Scarman, 1986)
Goal-based difculties with consent
The goal-based argument against consent is that seeking consent from
patients reduces recruitment rates into trials to the extent that the trials
themselves are in jeopardy. It is this conclusion at which Tobias and Souhami
arrive when they state:
The issue of informed consent has thus become a major barrier to the successful
conduct of randomised clinical trials in cancer. The many practical diYculties have
led to low levels of recruitment, especially where there is a substantial diVerence
between the treatment policies being compared (Tobias and Souhami, 1993)
Empirical evidence that consent reduces recruitment
There is some empirical evidence to support their concerns. The develop-
ment from duty-based to goal-based arguments in their article is rational
given that one of the reasons that recruitment levels are low, apart from
patients not wanting to join trials once they have been informed, is the
reluctance of their doctors to talk to themabout randomization (Taylor et al.,
1984). This is, presumably, out of concern for their patients well being,
although, as one correspondent pointed out, it may be that the doctor is
simply transferring his anxiety about challenging the eYcacy of a new
product to his patient (Walsworth-Bell, 1993).
The ISIS Trial in the United Kingdom and the United States
Tobias and Souhami cite the experience of the international study of infarct
survival (ISIS) investigators. The second ISIS trial, ISIS II, studied strep-
tokinase and aspirin in acute myocardial infarction. The patients who were
eligible for this trial were, therefore, extremely ill and in no Wt state to give
consent to participate. They were, in other words, incompetent at the time
when consent would need to be sought. Nevertheless, what Collins described
as humanly inappropriate written informed consent conditions were de-
manded of the investigators in the United States. These led to a very poor
recruitment rate compared with the UK, where consent was permitted to be
sought in whatever way was best for the patient (Collins et al., 1992). This was
usually retrospective or sometimes, I regret to say, not taken at all. In the
lengthy correspondence which followed the publication of Tobias and
Souhamis article many opinions were expressed on the matter, but some
were able to give actual examples of the goal-based problems consent had
caused. These were not only related to poor recruitment rate but also to the
validity of the research results:
We recently had to relax the time between the diagnosis of breast cancer and
recruitment to a study of adjuvant treatment. Because of observations that bisphos-
phonates decreased morbidity due to bone metastases in patients with breast cancer
we wished to determine whether these agents might delay the appearance of skeletal
disease when given to women at the time that breast cancer was diagnosed. We
intended to test the intervention at the earliest possible time in the natural course of
the disease, whereas the trauma to patients resulting from the diagnosis, radiotherapy,
chemotherapy, and informed consent have permitted us to randomise our patients
only up to six months after diagnosis (Kanis and Bergmann, 1993).
The breast conservation trial and the Presidents Commission
In Kings 1986 reviewthe United Kingdombreast conservation trial is cited as
an example of a trial which had to be stopped due to poor recruitment
following information given to patients (British Medical Journal, 1983).
However, the Presidents Commission study which aimed to identify reasons
for refusal of treatment found that the outstanding reason for refusal was bad
communication (Presidents Commission, 1983).
Patients will not happily agree to being randomized
More pertinently for our discussion, the Simes et al. (1986) study already
referred to found that the better informed the patients were, the less likely
they were to agree to randomization. This seems to be because the tendency
to want to establish a preference for one or other treatment is stronger than
the acceptance of genuine uncertainty.
Research which cannot take place if consent is sought
Finally, there are those research projects which could not happen at all if
consent were to be sought, not because diYculties in obtaining consent
scupper recruitment, but because the outcome of the research depends upon
125 Consent
the subjects not knowing they are in a trial. One such example (McArdle et
al., 1996), was a study of diVerent levels of psychological support for women
following breast cancer surgery. Varying levels of support from diverse
sources (routine care; routine care and breast care nurse; routine care and
support organization; routine care, nurse and support organization) were
compared with each other to see which was the most beneWcial. The research
showed that morbidity was consistently lower in the group which received
care from the breast care nurse only.
If the women concerned had known that they were participating in a
research project, their responses to the support would have been aVected and
would have rendered the study scientiWcally invalid. The goal-based require-
ments demanded that the right-based requirements were waived. My own
comment on the research at the time was this:
Does this [not obtaining consent in order to conduct the research] make the trial
unethical? The answer to this question depends on the balance of apparently compet-
ing moral claims. If you regard the research question as paramount you might argue
that the need to do the research outweighed the moral requirement to consult the
subjects of the research. If you regard the subjects welfare as paramount then you
would be happy in this case since the subjects would suVer no harm. But if your
overriding concern is that the wishes of people being used as a means to someone
elses ends should Wrst be consulted then you would not consider this research to be
ethical. By extension, of course, any research which needed to deceive the research
subjects in order to obtain a result would not be regarded as ethical. (Foster, 1996)
Interjection: should research be published if it was conducted
unethically?
Both the study of infants in high-altitude tents (see Chapter Seven), and the
study of post-operative support for breast cancer patients, involved waiving
important moral principles, the former of a duty-based nature, the latter of a
goal-based nature. A question often pondered by journal editors is whether
research which has been conducted unethically should be published. The
question is consonant with that of whether we should beneWt from any of the
research conducted so unethically in the Nazi concentration camps. It is
more diYcult to answer than the question of whether unethical research
should deliberately be conducted so that others may beneWt. That question, at
least in its baldest theoretical form given here, should clearly be answered in
the negative. But what about research which has already happened?
Some would argue that it is better that some good come out of unethical
research than none, rather like the argument that fetal tissue should be used
therapeutically, to ensure that some good comes out of an abortion. The
counter argument to this is that such a stance will encourage bad practice in
the future, giving house room to the idea that an action may be wrong in its
content but it will be made right by its consequences. In any case, goes this
side of the argument, if an action is wrong, then it is wrong, and no future
outcome accidentally enjoyed by virtue of it makes it right. The debate is the
classic one between consequentialist and deontological thinkers, in our terms
the goal-based and the duty/right-based thinkers. Intuitively and idealisti-
cally I would like to believe that I would not knowingly beneWt fromanothers
harm. A diVerent circumstance from medical research might elicit a diVerent
response from this, however. Suppose a Nazi tortures a Jewish freedom
Wghter who confesses that a bomb is placed outside some building. The Nazi
tells me this information, and I know that my child is inside that building.
Would I be wrong to save my child because the information was obtained
unethically?
In the case of research, there is a diVerence between the status of research
that has been conducted unethically according to goal-based standards, and
that which has been unethical according to duty- or right-based standards. In
the case of bad science (goal-based faults), the results cannot be extended to
shed light on a new situation. In the case of deontologically bad (duty- or
right-based faults) but methodologically good research the results may be
relevant to a future situation. Suppose the research on the eVects of the
atmospheric conditions of an aeroplane on babies had shown that there is
indeed a causal relationship between exposure to 15% oxygen and sudden
infant death syndrome? However unethically the research had been conduc-
ted, that knowledge ought to be publicized, for the sake of future babies.
It is, I still contend, important to remain deeply intuitively uncomfortable
with this situation. The usefulness of the results does not exonerate the
unethical conduct of the research. Allowing good consequences to justify bad
actions encourages bad actions. Therefore, the messy policy of journal edi-
tors, to publish research that is important scientiWcally but that has failed to
meet duty- and right-based criteria, but to do so hedged around by commen-
taries, has, on balance, to be welcomed. The fact that it is diYcult and
uncomfortable, and satisWes no one completely, is a witness to our continued
moral sensitivity and watchfulness. That is to be celebrated, not least because
it ensures that attempts will continue to be made not to start research that is
unethical. Better for the journal editors not to be faced with manuscripts of
unethical research. Needless to say, there will always be debates, keeping
researchers and research ethics committees busy, about the degree to which
goal-, duty- or right-based concerns have to be compromised or outweigh
each other before a research proposal becomes downright unethical.
Written consent
Written consent the patients signature on a piece of paper indicating her
agreement to take part in a research project may seem to represent one of
the details of the consent procedure which has been Wnally established as a
127 Consent
necessary detail and is therefore not open to question. The reason I want to
discuss it here is because there is evidence that patients regard the practice of
signing a consent form as a means of protecting the doctor, which is not what
its purpose was supposed to be. I would argue that the signature on the piece
of paper symbolizes the false worship of autonomy. It signiWes a handing over
of the moral responsibility for what follows to the patient and away from the
doctor. The price of patient autonomy in this symbolism is the loss of the
sense of being looked after by the doctor.
Empirical evidence that patients think their signature protects the
doctor
The empirical evidence shows that patients think the signature on the
consent form makes it a legal document absolving them of the right to make
any claim against the doctor who treats them. Chee Saw et al. (1994) found
this view in 62% of the patients in their trial. Dawes and Davison (1994), in a
study of the success of the consent procedure in surgery for ear, nose and
throat, stated: More than two thirds [of the study subjects] thought signing a
consent form primarily signiWed agreement to undergo treatment and that it
was a legal document; 54% thought there was an important medico-legal
aspect. When asked which of the following statements best reXected their
attitude when signing the consent form, 14% thought it protected the doctor
against being sued, while 90%thought they had to sign it. A study by Penman
et al. (1984) found that 63% of patients receiving investigational chemother-
apy said that the consent form played no part in their decision to accept
treatment, saying that it was their trust in their doctors guidance that was
most important. A more recent study supported this Wnding (Mark and
Spiro, 1990).
Consent as a separate, necessary procedure
The consent procedure, which involves oVering adequate information and
eliciting a voluntary response, is not covered by the signature on the piece of
paper. Moreover, the consent procedure itself stands outside the doctors
responsibility for his patient which should be there whatever the patient has
agreed to. These points were well put by a correspondent discussing the issue
of consent to blood transfusion:
Inform[ing] patients adequately . . . is exactly the factor that ensures that consent is
both appropriate and valid, regardless of whether a formal consent formis signed by a
patient in the presence of a healthcare professional . . . The signature of a patient on a
consent form does not mean that he or she assumes responsibility for the decision to
give a transfusion: this remains with the doctor (Richardson and Jones, 1998).
In the Sidaway case the issue was whether valid consent had been obtained
prior to an operation. Lord Scarman said, Mrs. Sidaway signed the usual
consent form, in which she declared the nature and purpose of the operation
had been explained to her. But the case went through all the courts in an
eVort to establish what was really said to Mrs. Sidaway.
What, then, is the purpose of the signature? Is it to prove not to the law
courts, but to those others who are tasked with checking on the researcher,
that he is doing what he should be doing? How will research ethics commit-
tees know that any attention has been paid to consent if there is no signed
consent form somewhere in existence? How will the Good Clinical Practice
auditors (ICHGCP, 1997) know that the procedures for consent have been
properly followed if there are no signed consent forms to look at? Yet it is
within most peoples experience to sign something they have not read,
particularly if it is being held out by a (busy) doctor or a nurse, whom they
believe has their best interests at heart.
The effect of the consent form on the research participant
The evidence points to the suspicion that, for a research participant, signing a
form has the eVect of making her doubt that her doctor is entirely happy
about what he is about to do to her, so much so that she has to sign
something saying she will not sue the doctor if it goes wrong, even though the
form does not indicate as much. If the consent form were just a bureaucratic
example of the increasingly regulated times then it would not matter quite so
much. But it does not help the patient, and it does not help the consent
process.
Concluding remarks
Under the general heading of right-based concerns consent has been identiW-
ed as the most important moral claim. The obligation to obtain consent arises
from the principle that no-one should be used merely as a means to an end
but always also as an end in herself. That, I contended, is a moral principle
which stands alone, not needing support from goal-based reasons of conse-
quence, nor from duty-based reasons of not harming. It is simply the right
that accrues to human beings. Of course, it can only accrue to human beings
who have autonomy to exercise. The implications of this were discussed in
Chapter Four. In this chapter the various ways in which the principle of
consent is threatened were considered. They come in three guises, which
approximate to the three moral approaches which have been the framework
for this book. The right-based threats were found in a large amount of
empirical evidence that the consent procedure does not work very well.
Information is misunderstood, and people do not feel free or well enough to
make voluntary decisions. In particular, there is a block about the meaning
and implications of randomization. The apparent inability of many people to
understand what random allocation is may be symptomatic of a refusal to
129 Consent
believe that their doctor is so uncertain of what is best for them individually
that he will prefer to toss a coin to allocate treatment. The objective, scientiWc
arguments for randomization cannot easily be faulted, and many would
objectively agree that if a valid result is to be reached, randomization is the
best way. It is diYcult for an individual to accept that for herself, however,
because she knows that she is a peculiarly unique individual, diVerent from
every single other individual alive or dead. The paradox was nicely described
by Professor Baum, who told a working party on informed consent that he
had spoken to a patient about a clinical trial for which she was eligible. She
had responded by asserting that she did not want to make the decision for
herself, but wanted Baum to decide what was best for her. He said that he felt
the best thing for her was to be in the trial, where she would have a 50%
chance of receiving the best treatment. She said that she did not want to be in
the trial. Which treatment would you prefer, then? asked Baum. I dont
know, you decide, was her response.
Duty-based threats to the principle of respect for autonomy show them-
selves more in theoretical arguments than in any empirical evidence that
people will be adversely aVected either by too much information, or by being
required to decide for themselves what to do about their treatment. Particu-
larly in the context of therapeutic research, where, it is argued, patients are
receiving as good treatment as if they were being treated in the ordinary way,
it should be left to the discretion of the researcher whether or not to seek
consent. However, the diVerence between treatment and therapeutic research
is the element of randomization, and as we have seen, that is a signiWcant
diVerence in the minds of patients. Unless she is explicitly told to the
contrary, a patient will assume that her doctor has given her the treatment
which her doctor (not a computer) has chosen for her, on the grounds that he
believes it is the best treatment for her as an individual. Hence, not to tell
patients about randomization, however diYcult doing so might be, consti-
tutes deceit.
The goal-based arguments against consent can be compelling, if good
research is seriously under threat because of the consent requirement. Never-
theless, just as it would be wrong to conduct research which broke a goal-
based rule by being badly designed, so would it be wrong to conduct research
which broke a right-based rule by not seeking consent. The importance of
respect for autonomy is too great to be easily overruled by consequentialism.
Finally, the practice of seeking written consent, a practice which has
become standardized during the 1990s, was considered. It was argued that its
value is very limited indeed, and that it may well do more harm than good, if
the good we are seeking is a proper consent procedure. There is irony in the
fact that the signed form neither protects the practitioner, nor does it allay
patient fears. Its continued use seems irrational and counterproductive.
Condentiality
ConWdentiality is a right-based issue because it concerns the protection of
personal information which, if it does not belong to the person, certainly
does not belong to anyone else. Informationwhich has beengiven in a clinical
context should always be assumed to be conWdential.
Should records-based or epidemiological research take place if it
compromises patient autonomy?
Records-based research is typically conducted by epidemiologists, and in-
volves looking only at patient records with no interaction with patients at all.
As said in Chapter Four, the principle governing conWdentiality in the
medical context is that while consent to disclosure may be inferred if the
disclosure is to those who are directly involved in the patients health care, it
cannot be inferred if the recipient of the information is not so involved. This
is the situation of the records-based researcher. The results of the research
may lead to improvements in health care which might eventually beneWt the
patients whose records are being scrutinized, but the intention is not to
beneWt those patients directly. Hence, this is non-therapeutic research being
conducted without patient consent, using information which is personal to
those patients. The numbers involved may be huge, and the researcher
unlikely to know the patients. Indeed, the researcher is unlikely to be the
slightest bit interested in who the patient is whose records he is examining: he
will only be looking for what interests him, and that will not be any unfortu-
nate medical secrets about the patient. These practical observations do not
detract from the fact that for such research to take place, the principle of
conWdentiality has been violated. But, say epidemiologists, this research has
to take place:
Probably every hazard we know of has required access to medical records, often with
the details of the exposure kept in one place and the details of outcome in another,
requiring a formal link between the two. The associations between taking oestrogens
in pregnancy and cancer, ionising radiation and leukaemia, and use of oral contracep-
tives and venous thrombosis were all documented in this way. Recently, the risk of
limb reduction defects associated with chorionic villus sampling in early pregnancy
was identiWed through linking records. Such work can also provide reassuring infor-
mation on the safety of medical procedures for example, conWrmation that am-
niocentesis does not cause clubfoot. The comparison of the results of blood tests from
pregnant women with and without speciWed congenitally malformedfetuses has led to
the development of antenatal screening tests. (Wald et al., 1994)
One solution to the problem of breaching conWdentiality is to anonymize the
records before the researcher sees them. This has serious resource implica-
131 Condentiality
tions (a goal-based problem), although it is possible. But in any case, the
examples of the beneWts of records-based research quoted above all depended
upon the records being named so that they could be linked. Another solution
would be to seek the consent of all the people whose records are involved.
This, again, has such huge resource implications that much, if not all,
records-based research would simply not be able to commence on that basis.
However, the right-based moral claim to privacy is strong, so much so that
versions of the two solutions given above were threatened in European
legislation, although they were in the end waived if the disclosure was for
research purposes. When it looked like they would be made law, Wald and
others (1994) argued strongly that such moves had to be resisted. They gave
the examples listed above as arguments for being allowed to continue with
their epidemiology, which are all goal-based, in that they seek to justify the
breach of conWdentiality by reference to the vast amount of good it has done.
They further argue, using a sort of compromised right-based approach, that
records were frequently shared already for the purposes of care of the patient,
and that:
In practice, the duty of conWdentiality is interpreted as applying not only to the doctor
directly involved in the care of the patient but also to those with whom he or she
judges the information may be shared; there is, in eVect, a professional duty of
collective conWdentiality . . . Once the sense of sharing personal medical information
among doctors subject to a strict duty of collective conWdentiality is accepted, it
becomes irrelevant whether the activity is for research, teaching, or care. The only
relevant issue is to ensure that the collective conWdentiality is secure and that no
breach occurs that could in any way adversely aVect the individuals concerned. (Wald
et al., 1994)
The lack of distinction between the uses to which the records are being put is
disingenuous. Of course it makes a diVerence: consent can be inferred, as I
said, if the reason for disclosure is in the best interests of the patient; such an
inference cannot be made otherwise.
It would seem that the conXict is a moral one: either the right-based
demand for respect to privacy strictly interpreted wins, and such research,
much needed, cannot take place; or the research is permitted to take place for
goal-based reasons and the right-based claim is overridden. In the UK,
guidelines from the Department of Health (1996) provide a fudge which
nevertheless lets the goal-based view take the upper hand: they say that if a
researcher has a contract with the National Health Service, and is therefore
bound by the rules of conWdentiality that all National Health Service person-
nel are bound by, and if the local research ethics committee agrees, then the
researcher may have access to patient records without consent. The fudge
provides a loophole: epidemiologists who wish to gain access to patient
records but who have no National Health Service contract simply obtain a
short-term contract. This does mean, of course, they must abide by rules of
conWdentiality.
Concluding remarks
The solution provided by the UK guidance is useful, since it allows epi-
demiologists to continue their work whilst at the same time emphasizing the
importance of the conWdentiality principle. That is the best that can be hoped
for. The principle of conWdentiality is important. But to rule out all epi-
demiological research automatically by making the principle absolute does
not accord with commonsense (a good test for ethical decision making). It
ignores the context, which is usually a mistake. By hedging the breach around
with conditions, respect is paid to the principle without letting it become a
dictator.
This chapter has considered right-based issues as they arise in research on
humans. The issues are those of consent and conWdentiality. How the prin-
ciple of respect for autonomy may come under threat from right-, duty- and
goal-based perspectives has been considered, as has what threatens conWden-
tiality, and how strong the arguments are for undermining it. Solutions have
not been provided, but rather we have shown how it is possible for there to be
disagreement about the ethics of a research project. By dividing up the
diVerent ethical concerns into three approaches we can see how a project can
fulWl one but not necessarily all of them. Although the decision still has to be
taken, if the approaches come into conXict, as to whether or not a research
project should go ahead, the analysis by reference to the three approaches
helps to clarify what, if anything, we are compromising by virtue of our
decision. It also, importantly, retains the integrity of the diVerent moral
claims. By making explicit the fact that, for example, right-based morality has
been waived in order to allow an epidemiological research project to happen,
the issue has not been fudged but we have rather been clear about what we are
doing. The task has also been given of demonstrating why the right-based
principle should be subsumed in this particular case. Having to do that
means having to take the principle seriously. That can help prevent unethical
research through ignorance of the moral issues at stake.
133
9
A framework for ethical review: researchers,
research ethics committees, and moral
responsibility
Introduction
How well does the three-approaches system of ethics work in the context of
research on humans? Three approaches to making moral decisions have been
identiWed (following Botros): goal-based, duty-based and right-based. Goal-
based morality is consequentialist, duty- and right-based morality are deon-
tological. As they have been deWned here, duty- and right-based morality
consider the content of the action itself; the distinction between them is the
way in which the rightness of the action is judged. Duty-based morality asks
that the action be in accord with moral principles that are believed to be right
regardless of the situation. Right-based morality, by contrast, considers the
wishes and concerns of those aVected by the action, not only in terms of
whom the action will make happy and whom not, but also in terms of their
views on the content of the action.
The three approaches combined
Although libraries of books have been written about versions of these three
approaches, and indeed about other ways of thinking morally about actions,
the three-approaches model works well for the purposes of considering the
ethics of research on humans because it can form a framework for ethical
review of research projects. Rather than try to decide which of the three
approaches is the best one, I have found that all three approaches have their
place in the decision-making process. Each is deWcient in some way, but the
deWciency is made up by the other two. Hence, the failure of goal-based
morality to consider the content of the action which leads to greater happi-
ness would mean that some morally abhorrent acts might be countenanced.
Duty-based morality Wlls in the gap, because it asks that actions in general
adhere to certain moral principles, among which would include not deliber-
ately harming anyone when acting. On its own, duty-based morality can be
blinkered, in that the moral principles do need to be applied intelligently in
134 A framework for ethical review
given contexts. Goal-based thinking provides the context: it identiWes a goal
and describes how it needs to achieve that goal. Duty-based morality, though,
should not be put oV by strident goal-based claims: however justiWed the
research might seemin terms of its outcome, there are some research projects
which are too risky to research subjects to be acceptable. The experiment with
babies in 15% oxygen tents discussed in Chapter Seven was an example of
that. At certain points, the duty-based moralist will argue that whatever the
consequences, there are some things that should not be countenanced. For of
course, unlike duty- and right-based morality, goal-based morality not only
prohibits certain actions (which lead to a decrease in happiness), it also
positively requires certain other actions. If an action will increase happiness,
then it is morally wrong by goal-based standards not to do it. This argument
can be used by researchers to justify what they do. The controlled trials which
validate new treatments not only do not violate moral (goal-based) prin-
ciples, they are also positively demanded by moral (goal-based) principles,
because the consequences of a treatment not being properly validated are far
worse than the consequences of its being validated (e.g. Wing, 1975). This is
an argument that has to be taken seriously. The context of moral decision-
making matters and the goals of a research project set that context. Ignoring
the goal-based imperative by making duty- and right-based principles abso-
lute creates a tyranny within which no one Xourishes, just as making science
and its pursuit an absolute creates a tyranny.
Typically, the duty-based concern will be with the harm to which the
research participants will be exposed. Any harm from research should be
justiWable and acceptable. In therapeutic research, the harm should be no
greater than that which patient would expect if they were patients only and
not research participants as well. The beneWts they might expect fromtherapy
should also be the same, as far as anyone can tell, in the research context as
compared with the practice context. This requires strong equipoise, or a
genuine balance of uncertainty about the eYcacy of the comparative arms of
the study. In non-therapeutic research, the duty-based moralist requires
harm to be minimized for the research subjects who can expect no thera-
peutic beneWt from participation.
Duty-based morality, as well as courting the danger of preventing clini-
cians from doing anything new and experimental, is also lacking in its
consideration of the autonomy, or reasonableness, of the people most af-
fected by the action itself. In the context of research, these are the research
participants. The trust relationship between doctor and patient means that
patients will assume that a doctor will not ask them to do things which are
harmful. Therefore, doctors should ensure that in non-therapeutic research
the risk is minimal. Nevertheless, doctors should also be wary of being overly
paternalistic in their assessments of what procedures their patients would
accept for the sake of others. Most human beings are capable of and want to
135 A framework to assist ethical review
think about whether what they are about to do is right or not. This is true of
the researcher in a research project, but it is also crucially true of the research
participants. The duty-based moralist may think she knows what is best for
her research participants; but they may take a diVerent view. Hence, the
potential research participants need to be asked whether they are willing to
take part in research, and whether they think that whatever they will have to
face as research participants constitutes an acceptable risk or not. If a trial has
satisWed goal-based requirements by aiming at a laudable goal, it has the
potential to do good. Blocking it because of a fear on behalf of the potential
research participants is not always a good idea. Those who will be exposed to
the risk should be asked what they think.
Moreover, even if the duty-based moralist is entirely satisWed that the
research is not exposing subjects to undue harm, for instance in the well-
designed therapeutic research project, it is still wrong not to ask the subjects if
they mind participating. Again, I have discussed suggestions, which are a
mixture of goal- and duty-based, that if the research is aiming at a good goal,
the researcher is in strong equipoise and the research participants will really
be treated just as well, if not better, in a trial than otherwise, then they do not
need to be consulted about whether they should take part or not. This
viewpoint is an aVront to persons who will thereby be used merely instru-
mentally. In any case, there may be all sorts of reasons particular to individ-
uals why participating in a research project is unacceptable, however much
good their doctors may think it will do themand future patients like them. So
the moral necessity of seeking consent from subjects to participate in re-
search, demanded by right-based morality, must be respected. The tendency
of duty-based morality to speak on behalf of others is thus modiWed by right-
based morality.
A framework to assist ethical review
The three-approaches framework can be translated into a series of questions,
the consideration of which helps lead to a sound and reasonable analysis of
the ethics of a proposed research project. They are oVered in order of
appearance, not in order of importance.
Goal-based questions
What question is the research project addressing?
Is the research aiming at a goal which is good and desirable?
How will the research achieve that goal so that the results are reliable?
How will the results of the research be disseminated?
Duty-based questions
Are the procedures, which research participants will have to undergo for
the trial to reach a successful conclusion, unacceptably risky?
If therapeutic research, is there equipoise?
If non-therapeutic research, are the risks greater than minimal?
Right-based questions
Will consent will be sought from potential research subjects?
What procedures to obtain consent will be followed?
How will conWdentiality be respected?
Resolving conicts between the three approaches
Goal-based questions to set the context
When considering the ethics of any given research project we should begin by
asking what question the research project is addressing. This sets the context
for the moral analysis which follows. The next question is goal-based because
it asks whether the question being investigated is important. Next comes the
question of the way in which the question is going to be answered. This is,
essentially, another context-setting rather than moral question, because all it
does is to identify the appropriate research method for the type of question
being asked, a morally neutral step. However, the way a research project is
designed is crucial to its moral acceptability, since if it is not designed in such
a way that the question will be answered, then it is a waste of all sorts of
resources. Worse still, the answer may be wrong but believed, and it will then
be the cause of wrong or bad practice being introduced into medical care. The
fourth goal-based question is about the dissemination and assimilation of the
results of the research. Again, the question is morally neutral but the answer is
not. Research which never sees the light of day, particularly if it has been
properly designed and has asked a question whose answer is worth knowing,
is also a waste of all sorts of resources.
Goal-based and duty-based moral imperatives in conict
Once these four questions are dealt with the research needs to be considered
froma duty-based perspective. Is the research design, which has been decided
is the right one for this research question, going to expose the research
participants who have to adhere to it to undue risk of harm? Here is where the
137 Resolving conicts between the three approaches
Wrst signs of conXict can begin to show themselves. If the research is non-
therapeutic, the duty-based thinker may have profound problems with any
risk to the research participants, no matter how important the goals of the
research may be. If there is a non-therapeutic research project with a very
desirable goal but which exposes participants to harm, should it take place or
not? This is the conXict between goal-based and duty-based morality. Goal-
based morality says the research should happen; duty-based morality says
not. If the goal-based moralist wins the argument, what harmmight befall the
research participants? If the duty-based moralist wins, what developments in
medicine might thereby be curtailed?
If the research is therapeutic, equipoise is demanded by the duty-based
thinker. Often this will be present in the design of the research with its
diVerent arms giving diVerent treatments, all comparable. However, the
goal-based argument for the need for placebo controls tips the balance
against duty-based acceptability, for in placebo-controlled trials, as we saw,
unless the standard treatment for the condition under research is in fact no
treatment, strong equipoise is not present. Another dilemma emerges at the
point of interaction between goal- and duty-based morality. The scientiWc
merit of the placebo-controlled trial is very signiWcantly greater than active-
controlled trials in the eyes of many, so much so that the US Food and Drug
Administration requires a placebo control unless the condition being treated
is life-threatening or serious. The duty-based thinker balks at the fact that
patients are not being treated equitably in a trial, that they have a chance of
receiving no treatment, when had they been patients in the ordinary way they
would have received some. The goal-based thinkers may seek to compromise
themselves to the duty-based thinkers by restricting placebo-controlled trials
to conditions which are not serious, as the FDA does, hence arguing that
exposure to placebo would not cause outright harm. But, as one strident
duty- (and right-) based thinker has put it, as far as the patient is concerned,
she is ill enough to have sought treatment. Thus, in the eyes of at least one
person, arguably the most interested and therefore the person whose opinion
counts for most, placebo, when standard treatments exist, is not acceptable
(Freedman et al., 1997b). Despite the FDAs requirements, the issue is not
settled, and such a policy may produce a really unethical trial by duty-based
standards because it potentially allows scientiWc rigour to override protection
of individuals from harm.
Goal-based and right-based
The next, right-based, set of questions relates to the need to treat all the
research participants as ends in themselves. Hence their consent must be
sought and their conWdentiality respected. In the case of any research project
involving competent participants, whose private data will be protected from
disclosure save with their agreement, these right-based requirements do not,
on the face of it, conXict with goal- and duty-based requirements. However,
some research projects scientiWc validity depends upon research participants
not knowing that they are subjects in a trial, such as the trial comparing
diVerent follow-up care for women having treatment for breast cancer
discussed in Chapter Eight. In cases such as these, either the womens consent
is not sought, or the research is useless. Either goal-based or right-based
moral demands win out; they cannot both be fulWlled.
Another way in which goal- and right-based demands conXict is research
which can only happen if patient conWdentiality is breached. This is the
situation with medical records research discussed in Chapter Eight. As we saw
in that case, UK regulations (Department of Health, 1996) have come to a
decision to allow the goal-based requirements to override the right-based
ones, whilst taking into account the minimal level of insult and so not
ignoring the right-based approach.
Duty-based and right-based
Right- and duty-based morality come into conXict when the duty-based
concern for the welfare and well-being of the potential research participant
will, in the eyes of the moral agent or researcher, be threatened by seeking
consent to participate in a trial, and yet, as far as her (duty-based) eye can see,
it would be most beneWcial for her patient to go into his trial. The conXict
between the duty-based concern for the patients welfare and the right-based
concern for his autonomy is given an added complication, and edge, by the
goal-based problem that seeking consent reduces recruitment rates. In some
cases it reduces them so much that trials have to be abandoned, which to the
predominantly goal-based thinker is a tragedy of quite extreme dimensions,
particularly as, to her mind, the patients who refused entry would have been
treated as well if not better in the trial than outside it. Does autonomy have a
price? Some would argue that it does not, as it is that which makes us human,
and demands that we treat others as ends and not means. It is our recognition
of a persons autonomy which makes us treat him or her as of equal worth to
ourselves. It was the failure to recognize the inmates of the concentration
camps as humans with autonomy, namely to dehumanize them and thus
make them expendable, which made it possible for the Nazi doctors to do
what they did. But even with such compelling arguments as these, the goal-
and duty-based arguments should not be automatically overridden. For some
people make foolish decisions. Although they do have that right, if their
foolish decisions aVect many other people, perhaps there is an argument for
ignoring them, or preventing them. Some would say that clinical trials of
important new treatments, put in jeopardy because patients do not know
what is best for them, are examples of this. Others would disagree.
139 When the three approaches fail
Goal-based and right-based again
A Wnal conXict might be seen between right-based and goal-based demands
when the subjects are incompetent and hence their consent cannot be sought.
If a research project involving incompetent participants takes place, the
goal-based approach has won. In a way it is a false hypothesis, however, since
in the case of incompetent participants right-based morality has nothing to
say, for consent cannot be sought, and it is confusing to argue as if it can. In
these cases, the duty-based demands for protection become doubly import-
ant, and the right-based concerns for respecting patient conWdentiality re-
main. I might argue that because the research participants cannot speak for
themselves, duty-based concerns must always override goal-based ones.
Whilst this seems intuitively right, I would still caution attempts to make
hard and fast rules.
When the three approaches fail
The foregoing discussion has assumed that it is possible to answer the moral
demands of each of the three approaches, even if it is not always possible to
answer all three of them in any one research project. Cases discussed in
Chapters Six to Eight have shown that it is questionable whether even this can
be so. For although the goal-based thinker can identify goals of research, it is
more often than not hard to decide just how important they are. Although
she can deWne the methods needed to answer the research question, rarely
will a research method answer it perfectly. The duty-based thinker can seek to
protect research participants from foreseeable harm, but no one can protect
them from unforeseeable harm. The right-based thinker can demand that
consent be sought, but no one can guarantee that consent is actually ob-
tained, however happy the patient is to sign the consent form.
Do these observations lead to the conclusion that medical research on
humans is probably always immoral? I think if I answered that question in the
aYrmative I would have missed the point of the moral endeavour. For after
all, we are not gods to determine our own and others fate. We can try to do
good, but we cannot be sure that we will succeed. What, in the end, the three
approaches oVer is a formulation of our motives for acting, not a recipe for
their moral success. The three approaches ask that we seek to magnify the
good (goal-based); that we are harmless (duty-based); and that we respect
other peoples autonomy (right-based).
This book has been addressed primarily to the moral agent, that is, the
researcher, and not to the research ethics committee, although it is every bit
as relevant to the latter as it is to the former. I hope it is relevant to everyone
who is involved in medical research on humans, at whatever level and in
whatever way. It is just that when it comes to making moral decisions about
whether to act or not it is up to the agent of the action to make those
decisions for him- or herself. We should not ask others to make those
decisions for us. We can seek advice, and we can take and obey orders, but in
the end we should only do so because we believe it is right. It may seem that
we are driven to make certain decisions. Moral thinking is a way of waking up
to the fact that how we act is up to us, and, therefore, since actions have
consequences, we should do as much good as we possibly can. It is a way, in
short, of becoming free and responsible, a choice which is available to every
rational being. This book is addressed to researchers because it is up to them
to take on the moral responsibility for the research on humans which they
conduct. The three approaches are meant to help that process work intelli-
gently. If there is an emotional commitment in researchers to the intentions
the three approaches identify, they will be acting as morally as they can and
are to be supported for that, whatever the success or otherwise of the
enterprise.
Research ethics committees
Research Ethics Committees came into being partly because researchers were
not, as far as those who were commenting on the situation could tell,
behaving as well as they should, and partly because a shift in moral emphasis
from goal- and duty-based thinking to right-based thinking, with which
doctors had seemingly not kept up, meant that some sort of external pressure
needed to be put on researchers to adopt a more right-based approach. More
kindly, but more tellingly, the idea was to provide a means by which
researchers could consult the views of those who were not so closely con-
cerned with their research goals as they were, and so were in a better position
to judge the ethics of what they were doing more fairly. This, although well
meant and probably needed, held out the possibility of moving the moral
responsibility for a research project one step away from the researcher herself
and on to the shoulders of the research ethics committee which, in the end,
has become not an advisory body but one which has positively to approve the
ethics of any research on humans before it takes place. Whilst the work of
research ethics committees has been really good, and their function very
necessary, their creation and development into gatekeepers for research on
humans has a sad inevitability about it. It is inevitable because, as will be
shown, it is hard to see how things could have developed otherwise. It is sad
because the keenness of moral thinking that I have tried to demonstrate is so
essential in the researcher has had the edge taken oV it by virtue of her having
to consult and follow the views of others. I am not decrying such actions,
but once rules are written most people tend not to think about why they take
141 Research ethics committees
the formthey do, and whether they are really good rules or not. The rules that
now surround research on humans are commendable but they do detract
from the researchers sense of personal moral responsibility. In what follows
the history of the establishment of research ethics committees in the UK is
recounted to show what I mean. In the US, where the equivalent Institutional
Review Boards were established by law from the start, the general point still
applies, that ethics is more than merely obeying a pre-written code.
The Nuremberg Code
The 1946 Nuremberg Code (Annas and Grodin, 1992) was drawn up follow-
ing the trials at Nuremberg of, among others, the Nazi doctors who had
conducted research on concentration camp inmates with such chilling
cruelty. The Code consists of ten principles to govern research on humans,
the Wrst of which is a categorical requirement that consent should be obtained
before anyone becomes a research subject. This Code was, as far as the
literature can tell us, largely ignored by researchers, perhaps because they
thought that the Code was only for badly behaved doctors like those tried at
Nuremberg. Moreover, the Code dealt only with non-therapeutic research.
During the 1950s, then, research was conducted but not under any sort of
ethical scrutiny apart from that of the doctors themselves. This was in
keeping with attitudes generally, for the doctors would regard themselves as
professionals bound by strict ethical principles, and their patients would
simply do as they were told.
Pappworths guinea pigs
During the 1960s, however, attitudes began to modify. One instigator of
change from within the medical profession was Dr Maurice Pappworth. He
was investigating and writing throughout the 1960s of abuses of medical
privilege in the interests of research. In 1969 his book Human Guinea Pigs was
published. It begins with the following accusation:
For several years a few doctors in this country and in America have been trying to
bring to the attention of their fellows a disturbing aspect of what have become
common practices in medical research. These practices concern experiments made
chieXy on hospital patients, and the aspect of them which is disturbing is the ethical
one. In their zeal to extend the frontiers of medical knowledge, many clinicians appear
temporarily to have lost sight of the fact that the subjects of their experiments are in all
cases individuals with common rights and in most cases sick people hoping to be
cured. As a result it has become a common occurrence for the investigator to take risks
with patients of which those patients are not fully aware, or not aware at all, and to
which they would not consent if there were aware; to subject them to mental and
physical distress which is in no way necessitated by, and has no connexion with, the
treatment of the disease from which they are suVering; and in some cases deliberately
to retard the recovery from that disease so that investigation of a particular condition
can be extended. (Pappworth, 1969)
His book describes the procedures of hundreds of experiments which bear
out this observation. His sources of information were not diYcult to Wnd:
they were chieXy articles published in the medical press. All he did, duty-
based fashion, was to describe the methods, adopted in the research in order
to gain the answer, from the point of view of the procedures which the
patients would actually have undergone.
Sir Austin Bradford-Hill and Mrs Hodgson
In the lay public, an organization called the Patients Association was begin-
ning to make concerned noises. These were Wrst prompted by some of the
points raised in a lecture given by Sir Austin Bradford-Hill, a renowned
medical statistician. He was speaking about the dangers of trying to codify
ethics and make rules that apply for all time to research on humans. Some of
his observations were prophetic, and these will be returned to later. Here,
however, I reproduce the point that really irked Mrs Hodgson, chairman of
the Patients Association:
Personally, and speaking as a patient, I have no doubt whatever that there are
circumstances in which the patients consent to taking part in a controlled trial should
be sought. I have equally no doubt that there are circumstances in which it need not
and even should not be sought . . . Surely it is often quite impossible to tell
ill-educated and sick persons the pros and cons of a new and unknown treatment
versus the orthodox and known? And in fact, of course one does not know. . . Can you
describe that situation to a patient so that he does not lose conWdence in you the
essence of the doctor/patient relationship and in such a way that he fully under-
stands and can therefore give an understanding consent to his inclusion in the trial? In
my opinion nothing less is of value. Just to ask the patient does he mind if you try
some new tablets on himdoes nothing, I suggest, to meet the problem. . . If the patient
cannot really grasp the whole situation, or without upsetting his faith in your
judgement cannot be made to grasp it, then in my opinion the ethical decision still lies
with the doctor, whether or not it is proper to exhibit, or withhold a treatment.
Three months later, an editorial in the British Medical Journal responded:
By his remarks about not telling patients of the whys and wherefores of an experiment,
Sir Austin has attracted the attention of Mrs. Helen Hodgson, chairman of the
Patients Association, who in the British Medical Journal of May 18 sharply challenged
him. Mrs. Hodgson, according to an article in the Times (June 17) headed Now a
Voice for the Patient, is moving forward with some determination. A questionary is
being sent out to members of the Patients Association which says this: Clinical trials
are obviously going on, and on a big scale. Patients are not told if they are receiving
new or orthodox treatment. I maintain that they should be told. I think it begs the
question to say that it is diYcult to explain these things to ignorant and sick people.
We agree with Mrs. Hodgson. (British Medical Journal, 1963)
The Medical Research Council, the World Medical Association and the
Royal College of Physicians
The medical profession, through its professional bodies, had not been entire-
ly idle during this time. The Medical Research Council issued a statement in
its report for the year 196263, entitled Responsibility in investigations in
institutions (Medical Research Council, 1964). This gave clear ethical prin-
ciples that should govern research involving human subjects. The World
Medical Association drew up the Declaration of Helsinki in 1964, also giving
principles for research on humans, against the production of which Sir
Austin Bradford-Hill was arguing in his 1963 lecture. Most eVective of all,
however, was the report issued by the Royal College of Physicians in 1967 to
boards of governors, medical school councils and hospital management
committees or equivalent bodies in non-medical institutions, and subse-
quently endorsed by the Ministry of Health (1968). This Report quoted the
Surgeon-General of the United States who, in 1966, had called upon the
institutions which were state funded to take full responsibility for the re-
search carried out in them. The College of Physicians Report said that
medical institutions had a responsibility for ensuring that all clinical investi-
gations carried out within their hospitals were ethical and conducted with the
optimum technical skill and precautions for safety. The Report then states:
This responsibility would be discharged if in medical institutions where clinical
investigation is carried out, it were ensured that all projects were approved by a group
of doctors [my italics] including those experienced in clinical investigation. This group
should satisfy itself of the ethics of all proposed investigations. (Royal College of
Physicians, 1967)
The Ministry of Health
The endorsing Health Notice from the Ministry of Health issued the follow-
ing year was sent to regional and area health authorities and to boards of
governors, asking them to set up these advisory groups. Prototype research
ethics committees were born.
The growth of research ethics committees power
Until 1991 not every District Health Authority had a research ethics commit-
tee, and until 1991 clear and detailed guidance was not forthcoming from the
government as to how these committees should function, who should sit on
them, and what the ethical criteria were against which to judge the research
proposals being put before them (Foster, 1997b). But the authority of
research ethics committees had grown. They began as advisory bodies,
oVering an independent view of the ethics of research proposals. For their
function to be recognized and used by sometimes reluctant doctors, they had
to have some authority, which meant that it needed to become mandatory to
consult them. During the period from 1968 to 1991 no new legislation was
proposed, as it had been in the United States from the beginning, but custom
and practice required that if research was to take place on human subjects
within the NHS, then research ethics committees had to be involved. What if
a research ethics committee gave a view with which a researcher disagreed?
For the research ethics committees function to have any meaning, the
committee had to have the Wnal word. Advice changed to approval, and
research ethics committees authority strengthened.
The importance of the issues research ethics committees were considering,
albeit not well deWned during this period, was recognized by the editors of
medical journals and those who funded or commissioned research. Over time
it became a rule that a good journal would not publish a researchers Wndings
if her project had not Wrst been approved by a research ethics committee.
Those who awarded funds required ethical approval prior to granting sup-
port. The pharmaceutical industry required its researchers to obtain ethical
approval; and in 1991 the (more) formal establishment of Local Research
Ethics Committees (Department of Health, 1991) led to a move by the
industry to seek approval from every local committee rather than a central
one, as had hitherto been the practice for multi-centre trials. A trial was
deemed to be ethical if it had been approved by a local research ethics
committee; not otherwise. The diVering interpretations of what constituted
ethical research that had developed within committees around the country
meant that local committees often came to diVerent decisions about research
which they had to approve. When the research was multi-centre this led to
major logistical diYculties as the same trial would be judged diVerently
around the country.
Multi-centre research ethics committees
Rather than ask the researcher to judge which were the right opinions, which
was obviously inappropriate by this stage, the solution was to create a new set
of committees, called Multi-centre Research Ethics Committees, which were
to approve the ethics of multi-centre research, with local committees retain-
ing the right to veto research in their patches for local reasons (Department of
Health, 1997). This structure is reXected in proposed European legislation on
the conduct of clinical trials.
Who guards the guardians?
One of the problems research ethics committees face is that although they can
approve or disapprove the research proposal that the researcher puts before
them, they cannot know that what they are shown is what actually takes place
without actually being there. Some committees are setting themselves up as
auditors to ensure that researchers do what the committees agreed they could
do (Berry, 1997; Smith et al., 1997). Concerns of a similar nature are now
being expressed about what happens in research ethics committees meetings,
where members may be making all sorts of mistaken judgements due to lack
of understanding of the research they have to review (Foster and Holley,
1998). Without actually watching the committees in session we could not
know. Some researchers regard the fact that a number of diVerent commit-
tees, reviewing the same research project, come to diVerent conclusions
about it, to be proof that some committees are not working properly (Harries
et al., 1994; Hotopf et al., 1995; Penn and Steer, 1995; Ahmed and Nicholson,
1996). The pharmaceutical industry, required by regulatory authorities to
prove that it has conducted its research in an exemplary manner, needs to be
sure that the research ethics committees which approve its research are
functioning properly. In the United States, auditors from the FDA patrol
committees and this may start to happen in the UK before too long. At some
point a question will be raised as to whether the auditors are doing their work
properly.
The point of telling the story of research ethics committees is not to
undermine their sterling work, which has been undertaken in often extremely
diYcult circumstances for thirty years. I cannot praise the endeavour highly
enough, because it has always been in support of the good. It is rather to
emphasize that no amount of regulation, auditing or public accountability
can detract from the fact that people are morally responsible for their own
actions. The seen to be ethical part of the ideal be ethical and be seen to be
ethical, which is often invoked in the context of research on humans, is
laudable, and the attempts that research ethics committees have made to
make it meaningful are even more laudable. However, appearances can be
deceptive, whereas everyone knows what she/he is up to individually. No one
can hide from him- or herself.
Bradford-Hills prophecy
Bradford-Hill had some unfortunate views about consent, which were at least
partly responsible for the development of the regulatory system now used to
govern medical research on humans. It is ironic that, in seeking to show the
dangers of regulating research, he helped to bring about the very regulation
he was keen to avoid. For the main message of his lecture was that once
written regulations appear, no matter how much their status is said to be
advisory rather than binding, they become the yardstick against which
actions are measured, and ultimately the laws reference point, should there
need to be one. In the end, he said, the Declaration of Helsinki, and other
guidelines which follow, will not help a researcher conduct his research
ethically, they will be the rules by which he is bound. The mark of advice is
that it is followed freely, and can be ignored. The Declaration of Helsinki,
should it be drawn up, would not long remain with that uncertain status.
And, of course, Bradford-Hill was right. By virtue of the regulatory system we
have for research on humans, the Wrst step of which was the Declaration of
Helsinki, the reason researchers conduct the research in the way they do is
because they run the risk of working outside the law if they do not. That, at
any rate, is the threat that calls researchers to order and gives research ethics
committees the power they wield.
Conclusion
This legalism is a far cry from the project of this book, which is to provide
researchers with the tools they need to be able to discern an ethical approach
to their research. I would urge you to conduct a proper moral analysis using
the framework for ethical review that I have oVered, rather than to refer to
rules which cannot anticipate every new situation. If you do use the frame-
work it will be possible to demonstrate that your decision was made on a
sound and reasonable basis, taking all the necessary factors into account.
That, in the end, is what matters most.
References
ABPI (Association of the British Pharmaceutical Industry) (1991). Guidelines for safe
laboratory practice. In Foster, C. (ed.) Manual for Research Ethics Committees, 5th
edn, 1997. London: Kings College.
Ahmed, A.H. and Nicholson, K.G. (1996). Delays and diversity in the practice of local
research ethics committees. Journal of Medical Ethics 22: 2636.
Annas, G.J. and Grodin, M.A. (1992). The Nazi Doctors and the Nuremberg Code. New
York, Oxford: Oxford University Press.
Aquinas, T. (c. 1264) Summa Contra Gentiles. transl. Vernon J. Bourke (1975).
London: University of Notre Dame Press; New York: Doubleday.
Aspinall, R.L. and Goodman, N. (1995). Denial of eVective treatment and poor
quality of clinical information in placebo controlled trials of ondansetron for
postoperative nausea and vomiting: a review of published trials. British Medical
Journal, 311: 8446.
Baum, M. (1986). Do we need informed consent? The Lancet, October 18, 2(8512):
91112.
Baum, M. (1993). New approach for recruitment into randomised controlled trials.
The Lancet, 341: 81213.
Bentham, J. (1789). An Introduction to the Principles of Morals and Legislation. In
Warnock, M. (ed.) (1962). Utilitarianism, pp. 3377. Glasgow: William Collins
Sons and Co.
Berry, J. (1997). Local research ethics committees can audit ethical standards in
research. Journal of Medical Ethics, 23: 37981.
Bhopal, R.S. and Tonks, A. (1994) The role of letters in reviewing research. British
Medical Journal, 308: 15823.
Black, D. (1998). The limitations of evidence. Journal of the Royal College of Physicians,
32(1): 236.
Botros. S. (1992). Ethics in medical research: uncovering the conXicting approaches.
In Foster, C. (ed.) Manual for Research Ethics Committees, 5th edn, 1997, vol. I,
pp. iv, 110. London: Kings College.
Boyd, R. (1998). A view from the man in the seat opposite. British Medical Journal,
317: 410.
Bradford-Hill, A. (1963). Marc Daniels Lecture: Medical ethics and controlled trials.
British Medical Journal, April 20: 10439.
Brewin, T.B. (1993). Logic and magic in mainstream and fringe medicine. Journal of
the Royal Society of Medicine, 86: 7213.
147
British Medical Journal. (1963). Ethics of human experimentation (editorial). July 6,
52489.
British Medical Journal. (1983). Informed consent: ethical, legal and medical implica-
tions for doctors and patients who participate in randomised clinical trials (editor-
ial). 286: 111721.
British Paediatric Association. (1992). Guidelines for research on children. In Foster,
C. (ed.) Manual for Research Ethics Committees. 5th edn, 1997, vol. II (27), pp. 124.
London: Kings College.
Britten, N. (1995). Qualitative interviews in medical research. British Medical Journal,
311: 2513.
Cairns, J.A., Gent, M., Singer, J. (1985). Aspirin, sulWnpyrazone, or both in unstable
angina. New England Journal of Medicine, 313: 136975.
Cassileth, B.R., Zupkis, R.V., Sutton-Smith, K. and March, V. (1980). Informed
consent: why are its goals imperfectly realised? New England Journal of Medicine,
302: 896900.
Chalmers, I. (1994). Public involvement in research to assess the eVects of healthcare.
Paper presented to the Harveran Society, 12 January, 1994.
Chalmers, I. (1995). What do I want fromhealth research and researchers when I am a
patient? British Medical Journal, 316: 131518.
Chee Saw, K., Wood, A.M., Murphy, K., Parry, J.R.W. and Hartfall, W.G. (1994).
Informed consent: an evaluation of patients understanding and opinion (with
respect to the operation of transurethral resection of prostate). Journal of the Royal
Society of Medicine, 87: 1434.
Ciociola, A.A., Webb, D.D. and McSorley, D.J. (1996). The continued use of placebo-
controlled trials in the study of peptic ulcer disease: a sponsor perspective. Drug
Information Journal, 30: 4339.
Collier, J. (1995). Confusion over the use of placebos in clinical trials. British Medical
Journal, 311: 8212.
Collins, R., Doll, R. and Peto, R. (1992). Ethics of clinical trials. In Williams, C.J. (ed.)
Introducing New Treatments for Cancer: Practical, Ethical and Legal Problems.
Chichester: Wiley and Sons.
Darragh, A., Kenny, M., Lambe, R. and Brick, I. (1985). Sudden death of a volunteer.
The Lancet, January 12th, pp. 934.
Dawes, P.J.D. and Davison, P. (1994). Informed consent: what do patients want to
know? Monash Bioethics Review, 13(4): 206.
Department of Health. (1991). Local Research Ethics Committees. HSG(91)5. London:
HMSO.
Department of Health. (1996). The Protection and Use of Patient Information.
HSG(96)18. London: HMSO.
Department of Health. (1997). The Ethical Review of Multi-centre Research.
HSG(97)23. London: HMSO.
Dodds, E.C., Goldenberg, L., Lawson, W. and Robinson, R. (1938). Estrogenic activity
of certain synthetic compounds. Nature, 141: 2478.
Dworkin, R. (1977). Taking Rights Seriously. London: Gerald Duckworth and Co. Ltd.
Easterbrook, P.J., Berlin, J.A., Gopalan R. and Matthews, D.R. (1991). Publication
bias in clinical research. Lancet, 337: 86772.
148 References
Food, Drug and Cosmetic Act of 1938, 52 Stat. 1040, 21 U.S.C. paragraphs 301 et seq.
at paragraph 505 (d) (as amended 1962).
Foster, C. 1996. Commentary: the ethics of clinical research without patients consent.
British Medical Journal, 312: 817.
Foster, C. (ed.) (1997). Manual for Research Ethics Committees, 5th edn. Kings
College, London.
Foster, C. (1998). Research Ethics Committees. In Chadwick, R. (ed.) Encyclopaedia of
Applied Ethics, vol. 3, pp. 84552. San Diego: Academic Press, Inc.
Foster, C. and Holley, S. (1998). Ethical review of multi-centre research: a survey of
multi-centre researchers in the South Thames region. Journal of the Royal College of
Physicians of London, 32(3): 2425
Freedman, B., Weijer, C. and Cranley Grass, K. (1997a). Placebo orthodoxy in clinical
research I: empirical and methodological myths. Monash Bioethics Review, 16(4):
1226.
Freedman, B., Weijer, C. and Cranley Grass, K. (1997b). Placebo orthodoxy in clinical
research II: ethical, legal and regulatory myths. Monash Bioethics Review, 17(1):
1022.
Garnham, J.C. (1975). Some observations on informed consent in therapeutic re-
search. Journal of Medical Ethics, 1: 13845.
Gillick v West Norfolk and Wisbech Area Health Authority [1986] AC 112, 169, 186,
18889, 195, 201.
Goodman, N., Cooper, G.M., Malins, A.R. and Prys-Roberts, C. (1984). The validity
of informed consent in a clinical study. Anaesthesia, 39: 91116.
Grubb, A. (1992). The law relating to consent. In Foster, C. (ed.) Manual for Research
Ethics Committees, 5th edn, 1997, vol. I, pp. ii, 1721. London: Kings College.
Hare, R.M. (1986). Utilitarianism. In Macquarrie, J. and Childress, J. (eds.) A New
Dictionary of Christian Ethics, pp. 6403. London: S.C.M. Press, Ltd.
Harper, W. (1998). The role of futility judgements in improperly limiting the scope of
clinical research. Journal of Medical Ethics, 24(5): 30813.
Harries, U.J., Fentem, P.H., Tuxworth, W. and Hoinville, W. (1994). Local research
ethics committees: widely diVering responses to a national survey protocol. Journal
of the Royal College of Physicians of London, 28: 1504.
Hart, H.L.A. (1973). Benthamon legal rights. In Simpson, A.W.B. (ed.) Oxford Essays
on Jurisprudence, second series, pp. 171201. Oxford: Oxford University Press.
Harth, S.C. and Thong, Y.H. (1990). Sociodemographic and motivational characteris-
tics of parents who volunteer their children for clinical research: a controlled study.
Hatch, E.E., Palmer, J.R., Titus-ErnstoV, L., et al. (1998). Cancer risk in women
exposed to diethylstilbestrol in utero. Journal of the American Medical Association,
280: 6304.
Herxheimer, A. (1993). Publishing the results of sponsored clinical research. British
HFEA/HGAC. (1998). Cloning issues in reproduction, science and medicine. London:
Department of Trade and Industry.
Holtzman, N.A., Faden, R., Churalow, A.J. and Horn, S.D. (1983). EVects of in-
formed parental consent on mothers knowledge of newborn screening.
149 References
Pediatrics, 72(6): 80712.
Hotopf, M., Wessely, S. and Norman, N. (1995). Are ethical committees reliable?
Journal of the Royal Society of Medicine, 88: 315.
Hoy, A.M. (1985). Breaking bad news to patients. British Journal of Hospital Medicine,
August, 969.
Hughes, V. (1998). Commentary: ethical approval of study was warranted. British
Hurwitz, B. (1998). Clinical Guidelines and the Law. Oxford: RadcliVe Medical Press.
International Conference on Harmonisation Good Clinical Practice (ICHGCP)
(1997). International conference on harmonisation of technical requirements for
registration of pharmaceuticals for human use. Guidelines for good clinical practice.
Relevant excerpts can be found in Foster, C. (ed.) Manual for Research Ethics
Committees, 5th edn, vol. II(2), pp. 130. London: Kings College.
Jones, B., Jarvis, P., Lewis, J.A. and Ebbutt, A.F. (1996). Trials to assess equivalence:
the importance of rigorous methods. British Medical Journal, 313: 369.
Jones, J. and Hunter, D. (1995). Consensus methods for medical and health services
research. British Medical Journal, 311: 37680.
Jones, R. (1995). Why do qualitative research? British Medical Journal, 311: 23.
Kanis, J.A. and Bergmann, J.F. (1993). Full consent may bias outcome of trials (letter).
Kant, I. (1785). Fundamental Principles of the Metaphysic of Morals. Translated by
Abbott, T.K. (1988). New York: Prometheus Books.
Kant, I. (1787). Critique of Pure Reason. Translated by Kemp Smith, N. (1978).
London: Macmillan.
Keen, J. and Packwood, T. (1995). Case study evaluation. British Medical Journal, 311:
4446.
Kennedy, I.M. (1988). Treat Me Right. Oxford: Clarendon Paperbacks.
Kerridge, I., Lowe, M. and Henry, D. (1998). Ethics and evidence-based medicine.
Kerrigan, D.D., Thevasagayam, R.S., Woods, T.O. et al. (1993). Whos afraid of
informed consent? British Medical Journal, 306: 298300.
King, J. (1986). Informed Consent. Institute of Medical Ethics Bulletin, Supplement no.
3 (stand alone publication). London: Institute of Medical Ethics.
Kitzinger, J. (1995). Introducing focus groups. British Medical Journal, 311: 299302.
Lancet editorial. (1993). Clinical trials and clinical practice. Lancet. 1993, 342: 8778.
Langman, M.J.S. (1997). Homeopathy trials: reason for good ones but are they
warranted? The Lancet. 1997, 350: 825.
Leeb, D., Bowers, D. and Lynch, J.B. (1976). Observations on the myth of informed
consent. Plastic and Reconstruction Surgery. 1976, 58(3): 2802.
Lehrer, S. (1979). Explorers of the Body. Garden City, New York: Doubleday and Co.
Ltd.
Ley, P. (1979). Memory for medical information. British Journal of Social and Clinical
Psychology. 1979, 18: 24555.
Linde, K., Clausius, N., Ramirez, G., et al. (1997). Are the clinical eVects of homeo-
pathy placebo eVects? A meta-analysis of placebo-controlled trials. Lancet, 350
(9081): 8348.
150 References
MacRae, K. (1996). Clinical research statistics: some deWnitions. In Foster, C. (ed.)
Manual for Research Ethics Committees. 5th edition, 1997, vol. I, pp. ii, 7. London:
Kings College.
Mark, J.S. and Spiro, H. (1990). Informed consent for colonoscopy: a prospective
study. Archives of Internal Medicine. 1990, 150: 77780.
Marx, K. (1843). On the Jewish Question. In Early Writings. Transl. Rodney Living-
stone and Gregor Benton (1973), pp. 21141. London: Harmondsworth. (Can also
be found in any collection of Marxs early writings.)
Mays, N. and Pope, C. (1995). Observational methods in health care settings. British
Medical Journal. 1995, 311: 1824.
McArdle, J.M.C., George, W.D., McArdle, C.S. et al. 1996. Psychological support for
patients undergoing breast cancer surgery: a randomised study. British Medical
Journal. 1996, 312: 81317.
McBride, G. (1994). Phase one trials can exploit terminally ill patients. British Medical
Journal, 308: 67980.
McDaniel Hutson, M. and Blaha, J.D. (1991). Patients recall of pre-operative instruc-
tion for informed consent for an operation. Journal of Bone Joint Surgery (America),
73(a): 1602.
Medical Research Council. (1964). Responsibility in investigations on human sub-
jects. In Report for the Year 19623, pp. 215. London: HMSO (Cmnd 2382).
Ministry of Health. (1968). Supervision of the Ethics of Clinical Trial Investigations.
HM(68)33. London: HMSO.
Noller, K.L. and Fish, C.R. (1974). Diethylstilbestrol usage: its interesting past,
important present, and questionable future. Medical Clinician of North America. 58:
739810.
Nuremberg Trials (1946). Transcript of proceedings. In Kennedy, I. and Grubb, A.
(1995) Medical Law, Text and Materials, pp. 101124. London: Butterworths.
Pappworth, M. (1969). Human Guinea Pigs. London: Routledge and Kegan Paul.
Parkins, K.J., Poets, C.F., OBrien, L.M., Stebbens, V.A. and Southall, D.P. (1998a).
EVect of exposure to 15% oxygen on breathing patterns and oxygen saturation in
infants: interventional study. British Medical Journal, 316: 88794.
Parkins, K.J., Poets, C.F., OBrien, L.M., Stebbens, V.A. and Southall, D.P. (1998b).
Authors reply. British Medical Journal, 316: 8934.
Penman, D.T., Holland, J.C., Bahna, G.F. et al. (1984). Informed consent for inves-
tigational chemotherapy: patients and physicians perceptions. Americal Journal of
Clinical Oncology, 2: 84955.
Penn, Z.J. and Steer, P.J. (1995). Local research ethics committees: hindrance or help?
British Journal of Obstetrics and Gynaecology, 102: 12.
Pieters, T. (1998). Marketing medicines through randomised controlled trials: the
case of interferon. British Medical Journal, 317: 12313.
Platt, R. (1972). Private and Controversial. London: Cassell and Co.
Popper, K.R. (1959) The Logic of ScientiWc Discovery. London: Routledge and Kegan
Paul.
Presidents Commission for the study of ethical problems in medicine and biomedical
and behavioral research. (1983). The Law of Informed Consent in making Health
Care Decisions, vol 3. Washington DC: Government Printing OYce.
151 References
Reilly, D., Taylor, M., Beattie, N.G.M., et al. (1994). Is evidence for homeopathy
reproducible? The Lancet, 344: 16016.
Richardson, N.G.B. and Jones, P.M. (1998). Responsibility for decision to give
transfusion remains with doctor, not patient (letter). British Medical Journal, 316:
77910.
Roberts, I. (1998). An amnesty for unpublished trials. British Medical Journal, 317:
7634.
Rothstein, J. (1995). Attending to transitions: a medical students encounter with
transplantation. Bulletin of Medical Ethics. November: 1319.
Royal College of Physicians. (1967). Supervision of the Ethics of Clinical Research
Investigations in Institutions. London: Royal College of Physicians.
Sackett, D.L. and Wennberg, J.E. (1997). Choosing the best research design for each
question. British Medical Journal, 315: 1636.
Savulescu, J. (1998). Commentary: safety of participants in non-therapeutic research
must be ensured. British Medical Journal, 316: 8912.
Scarmen, Lord Justice. (1986). Consent, communication and responsibility. Journal of
the Royal Society of Medicine, 79: 6978.
Scott, J., Weir, D.G. and Kirke, P.N. (1994). Prevention of neural tube defects with
folic acid a success but . . . Queens Journal of Medicine, 87: 7057.
Sidaway v. Royal Bethlem Hospital Governors. [1985] 1 AC 871.
Simes, R.J., Tattersall, M.H.N., Coates, A.S., et al. (1986). Randomised comparison of
procedures for obtaining informed consent in clinical trials of treatment for cancer.
Slevin, M.L., Stubbs, L., Plant, H.J., et al. (1990). Attitudes to chemotherapy: compari-
ng views of patients with cancer with those of doctors, nurses and general public.
Smith, T., Moore, E.J.H. and Tunstall-Pedoe, H. (1997). Review by a local medical
researchethics committee of the conduct of approved research projects, by examin-
ation of patients case notes, consent forms, and research records, and by interview.
Snowdon, C., Garcia, J. and Elbourne, D. (1997). Making sense of randomisation:
responses of parents of critically ill babies to random allocation of treatment in a
clinical trial. Social Science and Medicine 45(9): 133755.
Stacey, M. (1991). The potential of social science for complementary medicine some
introductory reXections. Complementary Medicine Research, 5: 1836.
Sutherland, H.J., Lockwood, G.A. and Till, J.E. (1990). Are we getting informed
consent from patients with cancer? Journal of the Royal Society of Medicine, 83:
43943.
Taylor, K.M., Margolese, R.G. and Soskolne, C.L. (1984). Physicians reasons for not
entering patients into a randomised clinical trial of surgery for breast cancer. New
England Journal of Medicine, 310: 13637.
Temple, R. (1996). Problems in interpreting active control equivalence trials. Ac-
countability in Research, 4: 26775.
Thorogood, M. (1996). Observational research: some deWnitions. In Foster, C. (ed.)
Manual for Research Ethics Committees, 5th edn, 1997, vol. I, pp. ii, 8. London:
Kings College.
152 References
Tobias, J.S. and Souhami, R.L. (1993). Fully informed consent can be needlessly cruel.
Tramer, M., Reynolds, D.J.M., Moore, R.A. and McQuay, H.J. (1998). When placebo
controlled trials are essential and equivalence trials are inadequate. British Medical
Journal, 317: 87580.
Vandenbroucke, J.P. (1997). Homeopathy trials: going nowhere. The Lancet, 350: 24.
Wald, N., Law, M., Meade, T. et al. (1994). Use of personal medical records for
research purposes. British Medical Journal, 309: 14224.
Waldron, J. (ed.) (1995). Theories of Rights. Oxford: Oxford University Press.
Wallace, L. (1984). Psychological preparation as a method of reducing the stress of
surgery. Journal of Human Stress, Summer: 6277.
Wallace, L. (1986). Informed consent to elective surgery: the therapeutic value?
Social Science and Medicine, 22(1): 2933.
Walsh, P. (1998). Commentary: doctors can never have a moral holiday. British
Walsworth-Bell, J.P. (1993). Doctors should admit uncertainty (letter). British Medi-
cal Journal, 307: 1495.
Watson, N. and Wyld, P.J. (1992). The importance of general practitioner informa-
tion in selection of volunteers for clinical trials. British Journal of Clinical Pharma-
cology, 33: 1979.
Weiss, R.A. (1998). Xenotransplantation. British Medical Journal, 317: 9314.
Williams, B. (1993). Morality. Cambridge: Cambridge University Press.
Wing, J.K. (1975). The ethics of clinical trials. Journal of Medical Ethics, 1: 1745.
Woodward, W.E. (1979). Informed consent of volunteers: a direct measurement of
comprehension and retention of information. Clinical Research, 27: 24852.
World Medical Association. (1996). Declaration of Helsinki. Latest version: South
Africa.
Wyatt, J.C., Paterson-Brown, S., Johanson, R., et al. (1998). Randomised trial of
educational visits to enhance use of systematic reviews in 25 obstetric units. British
Yunus, M. (with Jolis, A.) (1998). Banker to the Poor. London: Aurum Press, Ltd.
153 References
aaaaaa
Index
abortion 17
act utilitarianism 15, 656
advance directives 58
Alpha error 22, 97
alternative medicine 856
homeopathy trials 85, 868
analytic studies 23
Aquinas, Thomas 32
Aspinall, R.L. 100, 101
autografting using cloned embryos 7980
autonomy 911, 48, 50, 138
of reason 545, 56, 689
respect for 556
Baum, M. 34, 129
Bentham, Jeremy 1316, 656
best interests 911, 69
incompetent children or adults 589
non-therapeutic research and 434
see also duty-based morality
Beta error 22, 97
bias 22
publication bias 268
Black, D. 56
blood transfusion, consent to 127
Botros, Sophie 1011
Bradford-Hill, Sir Austin 142, 1456
breast cancer patients, postoperative
support for 125
breast conservation trial 124
Brewin, T.B. 856
Bristol Cancer Help Centre 27, 86
cardio-pulmonary resuscitation (CPR) 89
case study evaluation 24
case-control studies 23
categorical imperative 346, 68, 69
Chalmers, I. 90
Chee Saw, K. 117, 121, 127
children
childhood leukaemia 2, 4, 16
competence determination 58
decisions on behalf of 589
choice theory 503, 71
Ciociola, A.A. 967, 111
claim-rights 4950
cloned embryos, use for autografting
7980
cohort studies 23
Collins, R. 124
competence 5760, 139
adequacy of information 60
decisions on behalf of an incompetent
child or adult 589
determining a childs competence 58
in the ordinary adult 59
complementary medicine 856
homeopathy trials 868
conWdentiality 612, 72, 113, 1302
epidemiological research and 1302,
138
consensus methods 24
consent 9, 5761, 712, 11328, 138
competence and vulnerable groups
5760, 139
adequacy of information 60
competence in the ordinary adult 59
decisions on behalf of an incompetent
child or adult 589
determining a childs competence 58
duty-based diYculties with 1213, 129
information may increase anxiety
121
Scarmans test 123
should consent automatically be
sought 1212
155
consent (cont.)
goal-based diYculties with 1236,
129
consent reduces recruitment into trials
123, 138
patients agreement to randomization
124
research which cannot take place if
consent is sought 1245, 138
non-therapeutic research 44
right-based diYculties with 11420,
1289
beneWts remembered more than risks
11617
education level of volunteers 115
information provision to patients
11415
patients understanding of information
sheets 11516
patients understanding of
randomization 11720
review of literature 115
to blood transfusion 127
voluntary nature of the decision 60
written consent 1268, 129
eVect of consent form on research
subject 128
patients misunderstanding of 127
consequentialism 1011
see also goal-based morality
controlled trials 3, 5, 79
homeopathy trials 868
random allocation 3, 69, 22
see also placebo use; randomized
controlled trials
cross-sectional studies 23
Davison, P. 127
Dawes, P.J.D. 126
Declaration of Helsinki 55, 143, 146
deontological approach 10, 11
see also duty-based morality; right-based
morality
descriptive studies 23
diethylstylboesterol 45
disease registers 62
dissemination of research results 2530,
678
case studies 8890
futile treatment research 8890
pharmaceutical company research 89
Downs syndrome, heartlung
transplantation and 75
duty to care 1011, 37, 47, 93, 94
versus scientiWc goals 1036
duty-based morality 11, 3146, 6871,
1335
application of 3643, 69, 934, 11012
non-therapeutic research 435
therapeutic research 3943
conXicts with goal-based approach
6970, 1367
conXicts with right-based approach
137
natural law ethics 324, 367, 68
see also placebo use; three-approaches
system
Dworkin, Ronald 1011
Easterbrook, Philippa 267
embryos, use for autografting 7980
epidemiological research 23
conWdentiality and 1302
eproxindine hydrochloride study 1078
equipoise 3941, 69, 137
error
Alpha (Type One) 22, 97
Beta (Type Two) 22, 97
evidence-based medicine 25
encouraging clinicians to put knowledge
into practice 256
limitations of 2830
extra-corporeal membrane oxygenation
(ECMO) study 7, 11720, 121
fetal monitoring during labour 90
Fleming, Alexander 823
Florey, Howard Walter 84
focus groups 24
folic acid trials in pregnancy 956
Food and Drug Act (1962 amendments)
989
Food and Drug Administration (FDA),
USA 98100
Freedman, B. 978, 111
futile treatment research results 8890
Gillick test 58
goal-based morality 1330, 37, 658, 102,
1335
application of 1730
156 Index
dissemination of research results
2530, 678
goals of research 1821, 667, 91
research methods 215, 67
conXicts with duty-based approach
6970, 1367
conXicts with right-based approach
712, 934, 1378, 139
foundations of 1318
maximization of health 13
minimization of harm 13
utilitarianism 1317, 6566
problems with 31
see also three-approaches system
goals of research 1821, 667, 91
case studies 7380
immediate goals and general directions
1819
judging the morality of goals 19
problems with payment of researchers to
conduct
research 20
researchers aims at absolute good
201
versus duty to care 1036
Goodman, N. 100, 101
graft versus host disease 734
Hare, R.M. 15, 66
harm, minimization of 13
Harper, William 8890
healthy volunteer research 70, 10610,
11112
audit of volunteer screening procedures
1089
eproxindine hydrochloride study
1078
implications of volunteer payment
10910
women of child-bearing potential 110
heartlung transplantation, Downs
syndrome and 75
Herxheimer, A. 27
Hodgson, Mrs Helen 1423
Hohenfeldian cluster 4850
homeopathy trials 67, 85, 868
Human Fertilization and Embryology Act
(1990) 79
Human Fertilization and Embryology
Authority 79
Human Genetics Advisory Committee 79
hypoxia trial in infants 1046
immunity
legal immunity 50
theories of 81
individualized care 412
infants
sudden infant death syndrome trial
1046, 111
see also children
information 11415
adequacy of 60
increasing anxiety 121
patients understanding of information
sheets 11516
informed consent see consent
interest theory 503
International Declaration of Human
Rights 55
interviews 24
ISIS (International Study of Infarct Survival)
Trial 1234
Kant, Immanuel 32, 36, 48
autonomy of reason 48, 545, 689
Kennedy, Ian M. 401, 47, 556, 71
Kerridge, I. 28, 29
King, Jennifer 115, 121, 124
labour, fetal monitoring during 90
Lehrer, Steven 81
leukaemia, childhood 2, 4, 16
liberty 51, 523, 56
Medical Research Council (MRC) 956,
143
Ministry of Health 143
moral approaches 1011
see also duty-based morality; goal-based
morality; right-based morality
multi-centre research ethics committees
144
natural law ethics 324, 367, 68
nausea, postoperative, ondansetron trials
1002
non-therapeutic research 38, 435, 94,
10310, 137
best interests and 434
consent and 44
157 Index
non-therapeutic research (cont.)
healthy volunteer research 70, 10610
audit of volunteer screening
procedures 1089
eproxindine hydrochloride study
1078
implications of volunteer payment
10910
women of child-bearing potential 110
judgement of acceptability of risk 45
1046
non-therapeutic tests demanded by research
protocol 42
Nuremberg Code 55, 141
Nuremberg Trials 16, 31
observational research 23, 24
ondansetron trials 1002
organ transplantation 667, 7380
autografting using cloned embryos
7980
maintaining the condition of donor
organs 757
waiting lists 745
xenotransplantation 779
p value 22
Pappworth, Maurice 1412
Patients Association 142
payment
of healthy volunteers 10910
of researchers to conduct research 20
penicillin, discovery of 1, 67, 815
peptic ulcer treatment trials 967
pharmaceutical company research 89
Phase I studies 43, 1067
Pieters, Toine 89
placebo use 423, 69, 93103, 111, 137
FDAs arguments for 98100
folic acid trials in pregnancy 956
peptic ulcer treatment trials 967
scientiWc arguments against 978
Popper, K.R. 80
postoperative nausea and vomiting,
postoperative support for breast cancer
patients 125
power, statistical 22
pregnancy, folic acid trials 956
Presidents Commission 124
prevention 1
privileges 489, 55
Prontosil 84
publication bias 268
publication of unethical research 1256
qualitative research 234
random allocation 3, 69, 22, 129
patients understanding of 11720, 124
randomized controlled trials 8, 213, 25,
2830
design of 3943
equipoise 3941, 69
individualized treatment and 412
non-therapeutic tests demanded by
research protocol 42
see also controlled trials; placebo use
Reilly, D. 868
research
dissemination of results 2530, 678
case studies 8890
collecting the results of research 268
encouraging clinicians to put knowledge
into
practice 256
limitations of evidence-based medicine
2830
goals of 1821, 667, 91
case studies 7380
immediate goals and general directions
1819
judging the morality of goals 19
problems with payment of researchers to
conduct
research 20
researchers aims at absolute good
201
versus duty to care 1036
limitations of 56
methods 215, 67
case studies 808
choice of method 245, 67
observational or epidemiological
research 23
qualitative research 234
randomized controlled trials 213
non-therapeutic see non-therapeutic
research
scientiWc validity of 911
158 Index
value of 35
research ethics committees 1401
growth of power 1434
history of establishment 1416
multi-centre research ethics committees
144
problems with 145
researchers
aim at absolute good 201
judgement of acceptability of risk 45
payment to conduct research 20
right-based morality 11, 378, 4764,
712, 133
applications of 5762, 113
conWdentiality 612, 72, 1302
consent procedure 5761
see also consent
conXicts with duty-based approach 138
conXicts with goal-based approach
712, 934, 1378, 139
deWnitions of rights 4853
choice theory 503, 71
Hohenfeldian cluster 4850
interest theory 503
validity of rights 537
autonomy of reason 545, 56, 689
rights for the weaker party 556
Scarmans test 567
see also three-approaches system
rights see right-based morality
Roberts, I. 27
Royal College of Physicians 143
rule utilitarianism 15, 66
sample size 223
Scarmans test 567, 123
self-determination 11, 53
see also autonomy
Sidaway case 56, 123, 1278
Snowdon, C. 117
Souhami, R.L. 1212, 123, 124
Stacey, M. 85
statistical power 22
stem cells 79
sudden infant death syndrome trial 1046,
111
systematic reviews 28, 2930
teleological approach 1011
Temple, Robert 99100
three-approaches system 72, 1335
duty-based questions 136
failure of 13940
goal-based questions 135, 136
resolving conXicts between the three
approaches 1369
duty-based and right-based 138
goal-based and duty-based 1367
goal-based and right-based 1378,
139
right-based questions 136
see also duty-based morality; goal-based
morality; right-based morality
Tobias, J.S. 1212, 123, 124
Tramer, M. 1001
trials see controlled trials; randomized
controlled trials
Type One error 22, 97
Type Two error 22, 97
typhoid vaccine 82
unethical research, publication of 1256
utilitarianism 1317, 31, 656
act utilitarianism 15, 656
application to medical research 1718
rule utilitarianism 15, 66
Vandenbroucke, Jan P. 87
views of research participants 89
volunteers
audit of screening procedures 1089
education level of 115
see also healthy volunteer research
vomiting, postoperative, ondansetron
trials 1002
vulnerable groups
1046
see also children; consent
waiting lists for organ transplantation
745
Wald, N. 1301
Weiss, R.A. 778
Williams, Bernard 1517
World Medical Association 143
Wright, Sir Almroth 813
written consent see consent
xenotransplantation 779
159 Index

The Ethics of Medical Research On Humans

Uploaded by

Copyright:

Available Formats

The Ethics of Medical Research On Humans

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Ethics of Medical Research On Humans

Uploaded by

Copyright:

Available Formats

What are the three main approaches to moral decision-making discussed in the book?

What are some of the ethical issues surrounding experimentation with new cures on dying patients?

The ethics of medical research on humans

To members of research ethics committees past and present

You might also like