Tom Hollon, PhD
by Miriam Komaromy, MD
updated October 24, 2010
Before a new drug, surgical procedure, or therapy becomes available to
the public, it must go through a rigorous testing process and be evaluated
by the US Food and Drug Administration (FDA). This testing process consists
of a series of clinical trials that are designed to test the safety and
usefulness of the new drug compared to the current standard treatment.
clinical trials that make headlines are usually what are called phase
III trials. These are large-scale tests with hundreds or thousands of
patients. They are the culmination of earlier phase I and phase II trials
that include many fewer people and still earlier preclinical experiments
with animals. They are also the final tests in humans before the FDA is
asked to authorize sale of new medicines.
trials are designed to test whether a drug is safe for humans, and whether
the drug is effective in treating human diseases or conditions. Although
the drug has generally gone through extensive animal testing before the
trial begins, animal trials cannot always predict how new medicines will
affect humans. Even the most painstaking tests with animals give only approximate
indications of how people will respond to drugs. At some point, after thorough
study in animals (and when the FDA is convinced human experimentation will
probably be safe), tests with humans become necessary.
volunteers are required because people are highly variable in how
they respond to drugs
Not only is it necessary to test new drugs in humans, but they need to be tested in a large number of humans in order for the results of the trial to be clear. The reason so many volunteers are required is that people are highly variable in how they respond to drugs. It is not unusual, for example, for a drug be somewhat effective in only 30 percent of those who take it. For medical researchers to prove such a slight benefit requires testing the drug in as many as several thousand patients.
This extensive testing is part of what drives up the cost of new drugs the average development time is over 10 years and costs from 500 to 700 million dollars.
amount of testing has gone on by the time a drug is ready for phase III.
In general, a new drug or treatment goes through preclinical testing in
animals, then small phase I and phase II trials in humans before being ready
for large-scale testing.
volunteers are required because people are highly variable in how
they respond to drugs
a drug is discovered that seems to have medical potential, a drug company
will test it exhaustively in animals, looking for signs it may be poisonous,
cause cancer, or cause birth defects. Animal studies will also be used
to estimate the initial drug doses to be tested in humans.
animal experiments are finished, the company asks the FDA for permission
to begin clinical trials. The FDA only grants approval once they are satisfied
that the animal experiments are sound and that clinical trials are likely
to be safe.
test of the drug in humans is known as phase I. It is designed to find out
if the drug is safe rather than whether the drug is effective. Phase I is
also used to learn what drug doses to use in later trials, how the drug
is broken down in the body and excreted, and study short-term side effects.
I trials are designed to find out if a drug is safe for humans
phase I trial rarely has more than 100 participants. Often healthy people
are enrolled in phase I trials rather than patients on the assumption
that if the drug has unexpected side effects, healthy people have the
best chance of escaping permanent harm. But on other occasions, as with
a drug treating a serious disease like cancer, phase I subjects may be
patients who have failed standard treatments.
drug passes the safety tests of phase I, it advances to a phase II trial
with up to 200 participants. The goal of a phase II trial is to learn more
about safety and side effects, sharpen estimates of proper doses, and get
an early appraisal of whether the drug is going to work. This trial is often
the first time a drug is tested in actual patients.
II trials are often the first time a drug is tested in actual patients
III trial consists of hundreds or thousands of people. Commonly, phase III
will be conducted at several medical centers to see if people treated in
different locales have similar experiences. The central question of a phase
III trial is whether the drug works. Phase III will also give doctors an
extensive look at the drug's side effects. There are many different ways
of conducting a phase III trial.
central question of a phase III trial is whether the drug works
Phase III trials are randomized, double-blind trials. Randomized means people
are assigned at random either to receive the new drug, the standard treatment
for that disease, or a nonfunctional substitute (such as a sugar pill).
This last group is often called the control group, or the placebo group.
Because phase III must answer definitively whether the drug works, it's
important to compare people who receive it with others who do not.
is a way to help ensure that different groups in the trial have similar
characteristics. This makes it easier to compare outcomes between
good example of how people are randomly assigned to study groups is the
phase III trial for Herceptin, which treats a form of breast cancer. In
this trial women either received the standard breast cancer treatment
with Herceptin or the standard treatment without Herceptin. In this case,
there was no control group who received just a sugar pill because to do
so would be to not treat women with a potentially fatal illness. In this
trial, the women who received Herceptin with the standard treatment became
the test group; the others, the controls.
were assigned at random to the two groups. This means that the average
age of the two groups was roughly similar as was the severity of their
cancer so the results can be compared. If the two groups had differed
considerably in age or health, researchers would not be able to tell if
the drug itself was effective, or if the women in that group were just
younger or healthier.
trial was also considered double-blind, which means that neither the women
nor their doctors knew who was receiving Herceptin. Of course, it is natural
for doctors to want to know what treatments their patients are getting,
and in single-blind trials they do. There, only the trial participants are
unaware of which group they are in. But double-blind trials are considered
better because they prevent doctors from acting on preconceived notions
they may have about whether or not the drug works.
means that neither the doctor nor the trial participant knows whether
the participant is receiving the experimental treatment
example, a doctor in the Herceptin trial might have been tempted to offer
extra treatment to participants that weren't getting Herceptin. But the
physician's attempt to compensate for the fact that the participant wasn't
receiving the study drug would have collided with the requirement to keep
groups comparable in everything except who received the new drug. Unintentionally,
the doctor might have interfered with the trial, casting doubt on its
conclusions. This is why it is considered good practice for phase III
trials to be double-blind.
every trial is blind. In unblinded trials, often described as open trials,
both doctors and participants know what treatments are being given. Trials
of surgical procedures and comparisons of medical devices are often by
nature open. One of the problems with an open drug trial is that many
participants may not want to take placebos, because they presume the drug
will be better. Open trials, like single-blind trials, are considered
to be more prone to error than double-blind procedures.
patients are being treated with a combination of drugs, as is current
practice for HIV infection, a new drug may be evaluated by testing it
in combination with other drugs rather than by itself. A factorial design
trial may be used for this purpose. A simple factorial design would have
one group testing therapy A, another testing therapy B, a third group
testing A and B combined, and a control group testing neither A nor B.
Factorial designs are considered an efficient way to test medicines in
combination, but their results are not always easy to interpret.
a crossover trial, each participant gets both treatments being tested.
Some participants are assigned at random to receive drug A, and later,
drug B. Others receive B, then A. To produce valid results, the effect
of the first drug must end before the second drug is taken, and vice-versa.
This requirement can be hard to satisfy, and is one reason crossover trials
are not often used.
drug trials test drugs designed to treat diseases affecting fewer than
200,000 Americans. Some are rare genetic diseases that occur when missing
or defective enzymes prevent essential biochemical reactions from happening.
Because affected individuals are so few, an orphan drug may be tested
only on a small number of participants, who generally are so sick that
if the drug works, their improved health is readily apparent.
Lee, Chi-Jen (1993) Development and Evaluation of Drugsfrom Laboratory
through Licensure to Market. Boca Raton: CRC Press.
L. M., et al. (1996) Fundamentals Of Clinical Trials. St. Louis:
M.E. (1998) The Impact of the Orpan Drug Act. Modern Drug Discovery,
D and Feiden, K. L. (1995, March). Faster Evaluation of Vital Drugs. Scientific
American. p. 48-54.
R. (1998) Her-2The Making of Herceptin, a Revolutionary Treatment
for Breast Cancer. New York: Random House.