The Future of the Good Laboratory

Practice Regulations
SANDY WEINBERG

Tikvah Therapeutics, Inc., Atlanta,

Georgia, U.S.A.

BACKGROUND AND METHODOLOGY

There are many time-tested methods for predicting the future: you can examine tea leaves; gaze
into a crystal ball; throw I Ching coins; or seek out Delphic oracles. Since I am not a strong believer in
magic, I prefer to examine the intestines of goats.

No, reading goat intestines is not magic, though it might be clothed in a magical aura. But for a
primitive tribe of shepherds, the random selection of representative animals and examination of their
internal organs for signs of disease, parasites, of genetic mutation is a scientically valid and effective
method of predicting the likely future health of the herd, and the economic success of the tribe. It is
prediction of the future though extrapolation of the early trends discernible in the present. In addition,
while it is not without limitationsunanticipated events can disrupt normal developmentsextending
trends represents a sound method of looking forward.

So let us examine the metaphoric goat intestines of the good laboratory practice (GLP)
regulations, noting that all prediction carries a risk of uncertainty resulting from unexpected twists, but
that generally trends seemto progress along generally established pathways.

CONTINUED TREND TOWARD AUTOMATION

There isn't much risk, for example, in predicting a continuing trend toward laboratory
automation. In Part Because Of falling prices, in part because of increasing complex needs, and in part
because of growing regulatory acceptance, laboratories have been increasingly automated over the past
20 years. Over that time, I have conducted more than 300 laboratory audits. The last time I saw an
industrial GLP lab without any automated equipment, information systems, or data collection system
was in 2001, and it was about to be replaced by a fully automated laboratory information management
systems (LIMS). Today it is hard to imagine a laboratory that does not utilize some computer-controlled
equipment, and it is difcult to nd any lab manager who is not at least considering a further
automation step.

With the nal release of 21 CFR Part 11 and its risk assessment interpretation many laboratories
have converted to electronic standard operating systems (SOPs), allowing lab workers to have instant
access to SOPs while easily controlling revisions and modied versions. At the same time, LIMS that can
collect and evaluate data have grown in sophistication and reliability even as they have fallen in price:
most GLP labs have or are considering LIMS. Increasingly, those LIMS add another level of automation
sophistication, actively seeking rather than passively accepting data. In that active mode, the LIMS
signicantly contribute to the design of experimental methods and protocols.

Robotic devices capable of directing pipettes to inject microarray trays, or moving test tubes into
position, of passing samples from automated station to station, and even of optically scanning bacterial
colonies are increasingly in use, and are increasingly interconnected. In some limited function laboratory
applicationswater quality testing, for examplerobotic lab-in-a-box devices that can process a
sample through a variety of standardized tests are available and in increasingly common use.

This trend toward increased in automation number of collateral effects. In laboratories in which
protocols are standardized automated systems can increase accuracy, decrease personnel costs, and
enhance quality control (QC). Computerized systems are generally immune from the small slippages of
attention and care that result in minor but sometimes cascading human errors through the boredom of
repetitive tasks. On the other hand, computers are prone to much more spectacular errorsfor
example, to one off recording of test results from a long line of samples. But while these errors are
dramatic in scope, with effective QC efforts they are generally detectable and hence correctable. The
insidious minor corruptions produced by humans are much more likely to continue undetected.

On the other hand, the creative limitations of automated systems make innovation,
improvement, and inspired insights much less likely. Computers never seem to experience aha
moments, and reliance on the rote procedures of automated systems reduces the likelihood of
accidental discovery or insightful improvement. In short, automation in a late stage quality assurance
laboratory is probably a signicant advantage, while the same use of sophisticated tools in an early stage
research and development laboratory probably has mixed value. In both cases, though, the automation
of laboratories under carefully controlled and documented conditions improves the regulatory
environment: operating compliance GLP lab is easier when the critical functions are taken out of the
hands of employees and delegated to non innovative; always complain; never bored, tired or hungry;
innitely repeating automatons.

Because of nancial pressures to increase efciency; because of increasing capabilities and

TREND TOWARD INCREASED FOOD AND DRUG ADMINISTRATION INVOLVEMENT

The United States Food and Drug Administration (FDA) is always squeezed by completing
priorities. On the policy level, the FDA has to carefully balance public access to drugs (the approval of
new potentially life saving products) with the safety of drugs (restricting approval until safety is assured)
in an environment in which the media and the public can not seem to understand that every product
produces some side effects that are likely to negatively affect someone; in which every problem seems to
be someones fault; and in which everyone seems to know (or quickly meet) an attorney.

On a nancial level, the FDA is squeezed by a Congress that is always trying to control budgets
while increasing the scope of the agencys responsibility. It seems that the legislators have no problem
voting for a tight budget one day and publicly criticizing the agency for inadequate scrutiny the next.

On a scientic level, the agency is squeezed by geometric increases in biochemical information,

Currently the FDA is responding to two pressure vectors. First, ageneral realization thatmany
Americans nd the cost of drug products to be a limiting factor in implementing promising therapies has
let to the inclusion of product cost as a factor in the access to new products. These cost pressures have
realigned the safety versus access balance as excessive regulation; maximizing safety has emerged as a
signicant factor limiting public access. In the past, greater care resulted in higher levels of safety and
assured if somewhat delayed access. Now, with cost concerns added to the scenario, excessive safety
care not only delays access but may actually (and certainly perceptually) prohibits access for the working
poor. The end result is the increasing use of risk assessment to control the depth and extent of
regulatory concerns, and a new pressure to minimize regulatory expenses.

The second pressure, in some opposition to therst, calls for in creased regulatory involvement.
Publicity and lawsuits related to problems with vaccine shortage attributable in some public media to
too infrequent FDA inspections; criticism of the FDA investigation / inspection process regarding a
possible contaminating fungus in a contact lens solution plant; post market problems with some pain
blockers, some weight reduction therapies, and some heart therapies have led to demands for increased
FDA scrutiny both before and after approval. Of course, fullling these demands requires additional FDA
resources, carved from an already stretched budget, and adds to the end cost of products just as
demands for cost controls are peaking.

The role of laboratories, particularly for quality assurance (QA), places the GLP rmly in the
spotlight as the FDA struggles to deal with these competing demands for increased regulatory scrutiny
coupled with decreased regulatory (nancial) burden. Short term, the FDA is considering increasing the
fees it charges pharmaceutical companies-an answer that will meet immediate needs, but which will
eventually result (as price and cost elasticity decreases) in product cost increases.

In the longer term, expect the FDA to search for new technological solutions that can provide a
way out of the dilemma.
Test
facilities
operation of
animal care]
species storage, 75
reagents and solutions
discarding of, 7172
labeling of, 7172
records, GLP compliance
and, 228229
standard operating
procedures, 6571
archiving of, 71
availability of, 69
balance of detail, 67
deviations from, 69
electronic format, 7071
formats of, 6567
IACUC review, 6869
topics to include, 6768
updating of, 69
functional control and,
199200
integrity, 8283
out of range, 200
system
GLP definition of, 35
surveillance inspections
and, 249
user tests, 199200
validation tests, 199200
within range, 200
Test and control
articles, 7986
article/carrier mixtures, 8386
expiration dates, 8586
characterization of, 7982
handling, 8283
testing integrity, 8283
labeling requirements, 81
reserve sample sizes, 82
stability of, 8081
Test and control handling,
facilities and, 5859
TOP. See technical operating
procedures.
Training
documentation, GALP
requirements and,
139148
GLP inspection and, 172
Transmission, data loss during,
203
21 CFR Part 11
auditing checklist, 115128
computer systems, regulation
of, 112114
review of, 114128
independent audits,
114128
User tests, computer system
validation and,
199200
VAI. See voluntary action
indicated.
Validation
computer systems, 193222
EPA GALP, 194195
cost controls, 210215
evidence, 197205
data integrity, 202204
design control, 197199
functional control, 199200
managerial control,
201202
operation control, 200201
system reliability, 204205
report, 207208
tests, 199200
Voluntary action indicated (VAI),
Establishment
Inspection Report,
253254
Water, contaminant analysis of,
7678
Within range (normal range)
testing, 200
Written records, equipment and,
6465