Understanding Randomized, Controlled Clinical Trials
What are randomized, controlled, double-blind clinical trials?
A clinical trial is a research study conducted in human volunteers. Clinical trials are designed to decisively answer specific questions about vaccines or new therapies, such as whether they are safe and effective. Clinical trials are conducted in phases, starting with small Phase I studies that look primarily at safety, and progressing to large Phase III clinical trials, which are designed to show whether or not a vaccine or other medical technology is effective at either preventing or treating a disease. These trials lead to the licensure of a vaccine or therapy for public use. Other intermediate studies, such as Phase IIb test-of-concept trials, can also be used to give initial indications of efficacy (see VAX September 2005 Primer on Understanding Test-of-Concept Trials). The final stage of evaluation, Phase IV, occurs after a vaccine or therapy is licensed and is being used by large numbers of people, but these studies are not always required or completed.
The best way to determine if a vaccine or therapy is effective is to test it in a randomized, controlled, double-blind clinical trial. This type of trial is often referred to as the gold standard in medical research and provides the strongest evidence for the efficacy of an experimental product. Clinical trials of AIDS vaccine candidates are conducted in this manner to determine whether or not they are effective at protecting people from HIV infection or have some degree of partial efficacy that limits disease progression in individuals who become HIV infected even after receiving the vaccine (see VAX May 2007 Primer on Understanding Partially Effective AIDS Vaccines).
A controlled clinical trial compares the vaccine candidate or therapy being tested to either the best available treatment for that disease or, in the case of a preventive technology like a vaccine, against an inactive substance known as placebo that has no biological effect. AIDS vaccine candidates are tested in placebo-controlled trials—one group of volunteers is given the experimental vaccine candidate, while another group, called the control group, receives placebo. This allows researchers to detect any differences between the two groups regarding safety or efficacy.
For safety, it is valuable to compare any possible side effects in individuals who receive the vaccine candidate with those in volunteers who receive an injection of an inactive substance. The efficacy of an AIDS vaccine candidate in protecting against HIV infection is determined by comparing the number of individuals who become HIV infected—through exposure to the virus in their community—in each group. To say whether or not a vaccine candidate is partially effective, researchers compare the quantity of HIV in the blood, known as the viral load, in individuals from the two groups who become HIV infected through natural exposure to the virus during the trial.
Researchers can conclude whether a vaccine candidate is effective or not by looking at the difference between the vaccine and placebo recipients in either the total number of newly HIV-infected individuals or in their viral loads. If there is no difference, researchers can conclude that the vaccine candidate is ineffective. This was determined recently in the Phase IIb test-of-concept trial, known as the STEP trial, of Merck's AIDS vaccine candidate (seeSpotlight).
Whether a volunteer in a clinical trial receives the vaccine candidate or placebo is determined completely randomly by a computer program. However the randomization process involves more than simply dividing volunteers into two groups. For the results between the vaccine and placebo recipients to be truly comparable, the composition of these groups must be similar. For example, if the vaccine group involves only women, and the placebo group involves only men who have sex with men, the results between the two groups aren't comparable because it is impossible to rule out whether or not the route of HIV transmission may have affected the efficacy of the vaccine candidate.
Several factors must be considered during the randomization of volunteers, including sex, age, race, and geographic location. In AIDS vaccine trials, volunteers are also randomized based on behavioral factors that put them at increased risk of HIV infection, such as number of sexual partners. If the distribution of different factors is equivalent between the vaccine and placebo groups, a trial is randomized properly.
However there are always some factors that researchers can't account for during the randomization process. These are called confounding factors because they are not distributed evenly between the two groups and therefore can bias the results. Statistical analyses of completed clinical trials can sometimes help explain the effects of such confounding factors.
Another factor in the design of clinical trials that adds credibility to the results is double-blinding, which requires that neither the volunteers nor the researchers know who is receiving the vaccine candidate or placebo. Double-blind trials give more accurate results because individuals do not alter their behavior based on whether or not they are receiving the vaccine candidate. But some trials, such as those that offer a surgical intervention like circumcision, can obviously not be blinded and are referred to as open trials.
Several precautions are taken to keep trials blinded. Volunteers in a vaccine trial are assigned code numbers and staff members at a clinical trial site are only given a syringe labeled with that individual's code number. The pharmacist at the site, who prepares the syringes containing either the vaccine or placebo, only has access to the volunteer's code number and does not see any of the trial volunteers. Also, the placebo formulation is given in the same quantity as the vaccine and is made to look identical.
Researchers and volunteers usually do not find out who received vaccine or placebo until all volunteers finish their study visits and the trial is considered complete. Sometimes, such as in the STEP and Phambili trials, researchers decided to unblind volunteers before the trial is technically complete (see Removing the blindfold).