Even when results from early clinical trials are overwhelmingly positive, medications once approved often prove less effective than first hoped. Indeed, there’s a joke among doctors about newly-approved drugs that goes like this: ”Prescribe it now…while it still works.”

Why do drugs lose their lustre when they begin to be widely used? For many reasons. Initial enthusiasm on the part of investigators may bias the results to make them seem more favorable, even in controlled studies. Drug makers, who fund the clinical trials that are conducted to win approval, may intentionally make the results appear rosier by fiddling with statistics or leaving out less-encouraging results.

Now a new study offers another surprising reason. Volunteers chosen for clinical trials may not represent the patients who ultimately end up taking the drugs.

Researchers at the University of Texas Southwestern Medical Center looked at participants in an investigation called the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, which set out to measure the effectiveness of several treatment alternatives for depression. Only 22 percent of the 2,855 volunteers in that trial, the scientists found, would have met the criteria for inclusion in a typical antidepressant medication study.

Why not? Again, there are plenty of reasons. Patients who are obese or have diabetes are typically excluded from drug studies, for example, even though a large proportion of depressed patients (indeed, of all patients) fit into those categories. Patients who have tried other antidepressants without success are also often excluded from trials of new psychotropic drugs. Patients who report having suicidal thoughts or who have other psychiatric illnesses are also barred. As Dr. Madhukar Trivedi, a professor of psychiatry at UT Southwestern, who led the study, points out: “These are the patients impacted by depression the most — highest suicide potential, highest unemployment rates, highest social impairment, and they are likely to produce poorer outcomes. That population doesn’t get studied systematically in traditional pharmaceutical industry studies.”

Dr. Trivedi’s startling conclusion: “We are basing our judgment of clinical care in the United States on samples of patients that are totally different from the patient population actually treated in primary care and mental health facilities.”

The same problem is likely to afflict studies of many kinds of drugs, from insomnia medications to blood pressure pills.

In order to get clear results from studies, researchers do need to set criteria for eligibility. But if those criteria eliminate many of the very people most likely to use the drug, what good are the results? We think the scientific committees that approve research protocols need to do a better job of assessing eligibility criteria. The goal: to make sure that the people enlisted for clinical trials represent a good cross-section of the people most likely to use a drug once it’s approved.

© 2009 PDQhealth

Tags: drug testing, eligibility criteria