Why do so many trials of psychiatric drugs have negative outcomes? In fact, it is estimated that about one-third to one-half of trials of antidepressants fail to show a significant difference between drug and placebo, said Robert Temple, M.D., director of the Office of Medical Policy and the Office of Drug Evaluation I at the FDA’s Center for Drug Evaluation and Research.

One probable cause of these negative outcomes is that because there is often a lot of initial, rapid improvement in both the placebo and drug groups, which is usually sustained, the placebo group often improves so much that it is difficult for those taking psychiatric drugs to show even greater improvement.

For example, "typically, those receiving placebo might improve from a Hamilton Depression rating of 25 to 15 and thus be only slightly depressed. Consequently, there is not much room for the drug group to show additional improvement, so an effect from the drug may be hard to demonstrate.

"It would be like testing an antidepressant in a population that wasn’t depressed," Temple concluded.

When asked if this initial rapid improvement just occurred with trials for depression, Temple replied, "Our experience over several years shows that it occurs in a wide range of trials in psychiatric conditions including those for obsessive-compulsive disorder, generalized anxiety disorder, psychosis, as well as depression." However, because there are more trials for antidepressants, there are more data on these drugs, said Temple.

Although no one knows why this improvement occurs, Walter Brown, M.D., a clinical professor of psychiatry at Tufts and Brown universities, hypothesized that because it happens equally in the drug and placebo groups of antidepressant trials, it may not be from pharmacological activity. Similarly, Temple said, "I think it’s a characteristic of the studies, not the drug."

So what could be the cause? Brown, as well as Temple and David Hellerstein, M.D., offered their hypotheses on this subject.

Hellerstein has a very interesting hypothesis: "In talking informally, many drug researchers admit that they use a lot of incentives to recruit subjects; however, these are usually not mentioned in their journal articles. The incentives may include free transportation, payment, free lunches, lots of TLC, and even in some cases psychotherapy, whereas other trials might offer brief, impersonal encounters with researchers and no perks."

Hellerstein is the director of the Mood Disorders Research Unit and chief of outpatient psychiatry at New York Beth Israel Medical Center. He is also an associate professor at Albert Einstein College of Medicine.

Regarding psychotherapy, Hellerstein suggested that research assistants who spend a lot of time with subjects are in essence providing supportive therapy. And supportive therapy is an effective form of therapy, according to studies by Hellerstein and his colleagues.

As a result of receiving the perks and TLC, the subjects feel better—just as someone feels better from visiting a spa. And thus both those on drug and placebo undergo the same rate of initial improvement. Theproblem is, as Temple said, the placebo group improves so much it is very difficult to show an even greater improvement in those taking active medication, said Hellerstein, an investigator in many clinical drug trials, mostly on chronic depression.

Similarly Louis Lasagna, M.D., Sc.D., said, "If patients are chosen who are capable of responding to nondrug ‘perks’ to a significant degree, that would clearly leave very little room for demonstrating activity of a drug that really is active." Lasagna is chair of the board and adjunct scholar at the Tufts Center for the Study of Drug Development.

Hellerstein noted a related problem: Researchers cannot tell whether the subjects taking medication feel better from the perks and TLC (rather than the medication), the medication, or a combination of both the perks and TLC and the medication, concluded Hellerstein.

In reviewing these concerns, he suggested that the placebo aspects of both the drug and placebo groups should be monitored and standardized. He also thought that researchers’ papers should include detailed information on exactly what constitutes their placebo treatment.

Another psychiatric research expert, Darrel Regier, M.D., director of the American Psychiatric Institute for Research and Education, questioned whether this initial rapid improvement is occurring. "If it is happening, the pharmaceutical companies have a definite incentive to identify and eliminate it."

When asked about Hellerstein’s proposal to standardize placebo conditions, Regier replied, "I don’t think we have enough details on this problem to actually start promulgating standards. [Before considering standards], Hellerstein’s hypothesis should undergo empirical testing on the relative effect of different recruiting incentive strategies on diminishing the apparent efficacy of therapeutic interventions."

While Temple thought this initial improvement was indeed occurring, he observed that "we don’t know if this initial improvement is just spontaneous improvement or the result of perks and TLC, an effect of minimal psychotherapy/supportive therapy, or a consequence of the patient selection process, because there hasn’t been much research on this topic."

It would be helpful if someone designed studies to determine which, if any, of these factors are responsible, Temple suggested. Once this is known, "investigators could try to get rid of these factors," said Temple, who oversees the approval of drugs in three classes including neuropharmacologicals.

Temple has his own hypothesis on why the improvement occurs: "To get patients enrolled, there is an incentive for researchers and patients to score symptoms high. Once enrolled, this incentive no longer exists, and patients seem to improve dramatically," he said.

Finally, Brown offered his opinion on the placebo dilemma.

Although Brown agreed that the improvement was a probable cause of failed trials, he also pointed out that because this improvement occurs in all antidepressant trials, it cannot be the only reason trials fail.

When asked about Hellerstein’s hypothesis on "perks and TLC" and Temple’s on "scoring symptoms high," Brown replied that they could potentially contribute to this initial improvement and should be studied.

Furthermore, Brown pointed out another factor that might contribute to this improvement: Subjects in trials receive all the beneficial ingredients of being in a treatment situation including a thorough evaluation and personal attention.

"Several of these ingredients have been studied in a systematic way, and we know they have a benefit," Brown noted. He added, however, that there has not been any research to determine whether any of these specific elements are related to this initial improvement.