In 1948, British doctors published the results of the first clinical trial of the effectiveness of

streptomycin in tuberculosis. One group of patients was treated with streptomycin, the other -
according to the then standard pharmacotherapy regimens. The distribution of patients into
groups was performed according to the table of random numbers. The principle of
randomization - "randomly selected groups" - has become the gold standard of medicine. The
most acceptable and reliable is a randomized study with the principle of double-blind control.
During a randomized study of the effectiveness of a drug in a particular disease, groups of
patients (at least two) are distributed randomly. This achieves the practical identity of groups
of participants in quantitative and qualitative indicators. Analyze and evaluate the effectiveness
of a particular type of medical intervention. Non-randomized studies suggest that patients be
divided into groups at random, where random distribution is not possible for technical or ethical
reasons.
Cohort studies involve the formation of two or more groups (cohorts) of patients, of which
only one evaluates the relevant medical or therapeutic intervention, although the clinical
outcome is recorded in all groups. Observations can last for years (for example, the effect of
smoking on the development of lung cancer).
Transverse (or one-time) studies are conducted by questioning, examining, collecting
answers to a specific question among doctors and patients. Examination and collection of
information about the patient (or group of patients) is carried out once. This makes it possible
to establish the picture of the disease in one patient (or group of patients), to clarify the
symptoms, to determine the individual manifestations and severity of the disease. The end
result is a description of the disease in an individual patient, and in a set of options - is a study
of the relationship of some symptoms with the variant of the disease.
Case-control studies are performed in situations where the expected clinical effect is very rare
or slow. Form a group of people from individual cases of the disease or clinical effect. Next, a
control group is selected from persons without such a disease or condition, but similar in
important prognostic characteristics - age, sex, comorbidities. The number of patients who have
experienced certain adverse and undesirable effects is calculated in all groups. The results are
correlated taking into account known and measurable prognostic factors.
A case description or series of cases is a brief report of successful treatment or manifestations
of threatening complications of pharmacotherapy, which is essential for prompt medical
information. The value of the method lies in obtaining prompt notification of the complexity
of treatment, the occurrence of side effects, etc., because waiting for years for relevant more
reliable information is often impractical.
Recommendations for the management of patients should be systematized according to the
principle of reliability of the effectiveness and appropriateness of use.

Levels of evidence and degree of recommendations

In practice, health professionals can use many potential sources of information about medical
interventions:
• materials of researches carried out by medical experts or experts from other
areas;research
• materials and other information from pharmaceutical and other companies;
• research reviews and clinical guidelines;
• opinions of experienced specialists (experts);
• opinions of colleagues;
• their personal experience;
 • Patient testimonials based on personal experience.

For a doctor, the most valuable are studies published in scientific medical journals. This is due
to the fact that articles in journals are strictly selected and edited, which reduces the likelihood
of receiving poor quality information, or incomprehensible, uninformative message.
In turn, scientific reports in journals do not always contain the results of original research. It
can also be comments, discussions. Along with medical and biological studies, as well as
studies performed on animals are published in journals.
The most convincing information is research, characterized by the fact that it is a systematic
process conducted according to a clearly defined protocol, which seeks to exclude or explicitly
indicate the researcher's own passions and allow to obtain results that are relevant to patients /
clients and practitioners, working in this field.

Hierarchy of provability of information sources

systematic reviews

randomized clinical trials

cohort studies

case-control research

case study

description of individual cases

editorial articles

animal studies

in vitro research

Research data have different levels of evidence. Using the "pyramid of evidence", The doctor
should always prefer the results of the most evidence-based research. In terms of the
effectiveness of therapy and prevention, the most evidence-based studies are randomized
clinical trials (RCT). In the case where there are many RCT, systematic reviews allow to take
into account the differences between them and to get a generalized assessment based on the
whole set of RCT. Therefore, it is considered that the conclusions of systematic reviews are
more conclusive than the results of individual RCT.

Gradations (classes) and levels of evidence were developed at Oxford:

Class I - the presence of consensus and / or evidence of the effectiveness, appropriateness and
effectiveness of the procedure.
Class II - conflicting evidence and lack of consensus on the effectiveness and appropriateness
of the procedure
IIA - "scales" of evidence / consensus tend to the effectiveness and appropriateness of the
procedure;
IIB - "scales" of evidence / consensus tend to inefficiency and inexpediency of the procedure;
Class III - the presence of consensus and / or evidence of ineffectiveness and inexpediency of
the procedure, and in some cases - even its harmfulness.

In turn, the degree of proof of the effectiveness and feasibility of the procedure is divided into
three levels of reliability:
Level A - data obtained from at least two randomized trials;
Level B - data obtained in one randomized clinical trial and / or meta-analysis, or in several
non-randomized trials;
Level C is a consensus of experts' beliefs based on research results and clinical practice.

During the clinical study establish:

• improvement of schemes and terms of drug dosing;
• interaction with food or other drugs;
• the influence of certain factors of the drug on survival, etc.

In a clinical study, the goal of treatment of patients is determined by surrogate endpoints -

the parameters of the disease, which provide a direct or long-term result of the factor.
There are three types of endpoints:
• primary endpoints are the leading indicators that indicate a possible prolongation of the
patient's life (reduction of overall mortality, mortality from the disease);
• secondary endpoints are characterized by improving the quality of life of the patient or
by reducing the incidence of non-lethal complications, or by alleviating the clinical
signs of the disease;
• Tertiary endpoints are indicators that are not relevant to improving the quality of life or
prolonging it, but may indicate the ability to prevent disease by eliminating risk factors.

Clinical trials
Significant progress in elucidating the causes and spread of diseases has led to the development
of new methods for their diagnosis, treatment and prevention, many of which have reduced the
incidence, especially infectious. All this contributed to the strengthening of the empirical
approach in medicine. The peculiarity of the approach is the focus on the direct study of
phenomena. Real observation is used as a method. Only such an approach, according to most
scientists, has largely guaranteed the effectiveness of applied methods of diagnosis, treatment
and prevention of diseases. The predominance of the empirical approach in medicine has led
to the fact that almost by the middle of the XX century. judgments about the effectiveness of
diagnostic methods and methods of treatment of patients were based mainly on personal
experience, the experience of this team and authoritative opinion.
However, in the XV-XVI centuries. some scientists have believed that the potential
effectiveness of treatments and disease prevention, although consistent with empirical
knowledge, must be evaluated experimentally. Experiment (experience) is a general scientific
method of testing causal hypotheses by means of controlled intervention in the natural course
of the studied phenomenon. The purpose of epidemiological experimental research is to assess
the potential and actual effectiveness and safety of prophylactic and medicinal products,
methods and schemes of treatment, diagnosis and prevention of diseases.

A clinical trial (CT) is a controlled experimental study in which subjects receive prophylactic,
diagnostic, or therapeutic agents to evaluate their effectiveness and safety.
The general rules for involving human CT follow from the Nuremberg Code and the more
detailed Helsinki Declaration of the World Medical Association. Later, in order to streamline
preclinical and clinical studies, the WHO in 1974 developed "Guidelines for the evaluation of
drugs for human use." Subsequently, on the basis of this document in the United States were
developed national rules for CT, issued in 1977 under the name "Rules of Good Clinical
Practice" (Good clinical practice, GCP). Then similar rules were adopted by other countries
(EU, Japan, Canada, Australia). In order to harmonize them, international conferences were
held (the first in 1991, www.ich.org), in which drug manufacturers played a significant role,
so GCP does not fully meet the requirements of the Declaration of Helsinki. The language of
these documents also differs. While researchers talk about medical intervention trials as
research options that focus on evaluating the effectiveness and safety of the intervention, the
ICH documents use the more general term "research." Since 1998, the WHO has been
implementing the project "Implementation of International Standards in the Practice of Clinical
Trials in Central and Eastern Europe". Thus, GCP rules have been developed to ensure that CT
results are reliable and accurate and protect the rights, inviolability and confidentiality of
subjects. They cover the entire chain of clinical trials - setting, conducting, performing,
controlling, inspecting, registering, analyzing and reporting on CT.

The concept of three E in clinical trials should meet: Efficacy; Effectiveness; Efficiency
(Benefit for the individual patient / population / society as a whole).
The process of clinical trials of new drugs includes four interrelated phases.

Classification of experimental epidemiological studies.

Randomized clinical trial: evaluation of the potential efficacy and safety of immunobiological
drugs and medicines.
Randomized field study: evaluation of the potential efficacy and safety of immunobiological
drugs.
Continuous field study: assessment of the real efficacy and safety of immunobiological drugs
and medicines.

The drug undergoes serious tests before it appears on pharmacy shelves. The following practice
is accepted in the world: at first these tests are carried out within the framework of the
preclinical phase (preclinical phase), which implies the development of the drug in research
centers and laboratories. Usually, organizations that develop new drugs are called research and
development organizations. In large pharmaceutical companies there are departments of
research and development (Research and Development Departments). However, many small
companies are developing 3-4 new drugs or even just one drug. Often the financing of such
companies is provided by the issuance of shares, which mobilize funds for research. Upon
completion of the preclinical phase, such Research and Development companies can sell their
formula to large pharmaceutical companies or start conducting CT themselves. As a rule, they
have neither the experience nor the ability to conduct CT, then they begin cooperation with
contract research organizations (Contract Research Organizations).
Inclusion and exclusion criteria. Criteria for inclusion of patients (subjects). Needed to describe
the population (general population) to which the patients included in the study correspond.
Exclusion criteria. Needed to create a homogeneous sample, ie less variability of variables in
the initial state and in estimating the magnitude of the effect of the intervention. Persons with
severe comorbidities, life-threatening conditions or interfering with experimental conditions
(for example, with dementia) are excluded from the number of CT participants. Thus, the
statistical sensitivity of the experiment increases.
Consent of participants. Ideally, all patients who meet the inclusion criteria should participate
in the trial. In practice, not all patients agree. Some may prefer one of the proven treatments
and do not want to give it to the tried and tested method. Others do not want to be studied or
choose another treatment. Such patients are not included in the study. It is necessary that the
response rate, ie the share of people who responded to the request to participate in the study,
was quite high, at least 80%. Patients will follow the recommendations depending on the
acceptability of the study. The results of treatment of such patients are higher and do not depend
on treatment. Subjects, actively choose a method of treatment, are treated more diligently, more
correctly perform the appointment. This property of people is called accuracy or diligence, but
more often - compliance.

Planning the number of participants. The number of patients included in the experiment
(sample size) should be justified, based on:
• the expected level of efficiency;
• research structures;
• pre-established threshold of statistical significance of effect detection;
• prevalence of the disease.

When planning a study, expect the number of patients to be sufficient to detect the expected
effect. Calculations are quite complex, they are performed using statistical programs.
Randomization - random distribution of patients into groups. Its purpose - minimal differences
between groups, by all indications are random, not intentional. The methodology of statistical
analysis of data proceeds from the principle of random staffing of groups: differences of groups
by definition are casual.
Randomization is carried out in different ways: the use of tables of random numbers, computer
programs. Sometimes randomization is replaced by pseudorandomization (division into groups
by the first letter of the name, date of birth, medical card number, day of the week of admission
to the clinic, etc.). Its application can affect the correctness of the sample and, accordingly, the
evaluation of results. The most terrible adverse consequence of pseudorandomization - will be
known to belong to each patient to a particular group (main or control). Thus, the main
condition of randomization will not be fulfilled - hiding its results; its most important function
- ensuring the blind nature of the study - will not be realized. In studies where no measures
were taken to conceal the results of randomization or concealment was insufficient, the
assessment of the effectiveness of the intervention was overestimated by approximately 25%.
To ensure concealment, reliable technical measures are taken (for example, after the
registration of the patient who agreed to the test, information about him is entered in the
database of the organizer of the study).
Placebo. When evaluating the effectiveness of a new drug, the question arises about its
effectiveness, ie the ability to reduce the likelihood of adverse outcomes compared to the
absence of intervention. In the control group, the lack of intervention may be psychologically
unacceptable for patients, leading to non-compliance with the test regimen. Patients left
without treatment switch to self-medication. That is why patients in the control group are given
a substance (procedures performed) that do not differ from active intervention. A placebo is
usually a dosage form that lacks the active ingredient, such as a tablet form that is identical in
color and shape to the active ingredient but contains only an indifferent substance, kaolin or
starch, and for injectable forms, an isotonic sodium chloride solution. The use of placebo is not
always possible, and sometimes unethical, for example, when it is unacceptable to deprive
patients of effective treatment. Then the control group is prescribed standard treatment and
placebo, and the main group - standard treatment and the study drug. The effectiveness of a
new drug is easier to show in comparison with placebo, in comparison with an existing drug it
is necessary to prove a greater or the same effect of a new drug.
It is believed that the use of placebo has a positive effect, the "placebo effect". The beneficial
effect of placebo is related to its psychological effect on the patient. Placebo has little effect
only on results that reflect the patient's subjective condition (sleep quality, pain intensity).
Clinically important results are not affected by placebo (life expectancy, remission duration,
functional defect, etc.).
Difficulties in prescribing the drug. Regardless of the nature of the intervention (therapeutic,
diagnostic, prophylactic), it should be clearly described and standardized.
At appointment of some interventions selection of a dose does not cause difficulties: parenteral
administration of drug according to the scheme provides receipt in an organism of a certain
amount of active substance. The use of oral forms of drugs already leads to difficulties in
dosing. Depending on compliance, patients may not take the daily dose, and in the case of
severe side effects - and reduce it altogether. There are interventions that are difficult to dose.
These include surgery, chiropractic, acupuncture.
During CT, treatment previously prescribed to the patient is usually stopped. The period after
cessation of previous treatment and before the beginning of CT is set so that the concentration
of the active substance decreases. If patients in the main group are taking additional drugs (co-
intervention), there may be a shift in results toward higher efficacy. If patients in the control
group use the same drugs as in the main group (contamination, contamination), the result may
be shifted towards the ineffectiveness of the drug.
The trial takes measures to prevent contamination and co-intervention and to increase patient
and staff compliance with the protocol's proposed actions. One way is to conduct the
introductory stage to the test. At this stage, patients who do not follow the regimen are detected,
for example, by determining in the urine of substances introduced into the drug as a label. Then
only executive patients are included in the trial. Co-intervention and contamination are almost
inevitable, they must be taken into account when analyzing the data.

Outcomes ("target" signs) - events that will evaluate the effectiveness of treatment or other
interventions. Types of results:
• clinically important results (mortality, life expectancy, frequency of exacerbations,
maintenance);
• intermediate;
• indirect;
• surrogate results.

Quality of life. When assessing the effectiveness of the intervention should not forget to assess
the quality of life. Sometimes a small increase in life expectancy can be achieved by
unacceptable suffering during treatment, such as oncology. To assess the quality of life,
complex scales are used, the final score of which is obtained as a result of summing up various
information (about the intensity of pain, mood, breathing, ability to wash yourself, take care of
yourself).
Termination of the trial.
The duration of CT is planned based on the number of participants, the expected frequency of
cases and the difference between interventions (the size of the effect), the planned statistical
significance of the result. It is incorrect to test until the result is statistically significant, because
sooner or later statistically significant differences can be achieved. That is why the duration of
the CT is set in advance.
In long-term trials, the rules for terminating CT are set due to the need to maintain the safety
of participants and the possible receipt of convincing results in favor of one of the studied
interventions.
Trial with data analysis depending on the prescribed or received treatment. The results of
controlled randomized trials can be analyzed and presented in two ways: either on the basis of
the appointment of a treatment in randomization, or on the basis of treatment actually received
by the patient. The correct presentation of the results depends on the question.
• If the question is which treatment tactics are best for making a clinical decision, then
the analysis based on the treatment prescribed at randomization should be used,
regardless of whether all patients actually received this treatment. This approach is
called intention to treat analysis. Advantages of this approach: the question asked
corresponds to what is usually of interest to the clinician when prescribing treatment,
and the compared patients are really divided into groups randomly. Disadvantage: if
many patients do not receive the proposed treatment, the differences between the
experimental and control groups disappear, the probability of a negative test result
increases. In this case, the lack of differences between groups can be interpreted
differently: either the experimental intervention is actually ineffective, or it was simply
not applied.
• If we are interested in whether experimental treatment is really better than control. In
this case, the analysis based on the received treatment is more suitable for the answer,
ie an assessment of the effect of the treatment that each patient actually received and
regardless of what treatment was prescribed to him at randomization. The mechanism
of the studied effect is clarified. The disadvantage of this approach is that if most
patients do not receive the proposed treatment, the trial ceases to be randomized and
becomes a routine cohort study. This means that all differences between groups,
excluding the method of treatment, must be leveled in some way (by imposing
restrictions, selecting pairs, dividing into subgroups or standardization) to achieve full
compatibility, as is the case in non-experimental studies.