Study details misconduct in drug research.

Study details misconduct in drug research

An orthopedic surgeon under investigation for scientific misconduct by the Food and Drug Administration broke into his own office, settling fire to a file room and throwing medical records into a whirlpool bath. The doctor's shenanigans did him no good: FDA barred him from running clinical drug trials after an audit revealed purported study patients who had never participated in the trial.

Although FDA penalized the surgeon in this case, some drug investigators who flagrantly violated scientific standards escape punishment, according to a report in the May 5 JOURNAL OF THE AMERICAN MEDICAL ASSOCAITION. FDA must take stronger action to weed out incompetent or dishonest researchers, say Martin F. Shapiro of the University of California, Los Angeles, and Robert P. Charrow, formerly with the Department of Health and Human Services in Washington, D.C.

FDA regularly sends investigators to examine laboratory records and other data kept by researchers hired by drug firms to study experimental drugs. Shapiro and Charrow analyzed data collected from 1,955 such audits conducted by FDA from June 1977 to April 1988. Auditors found "serious" deficiencies -- such as failure to obtain informed patient consent -- in 12 percent of audits prior to 1985 but in only 7 percent since that date, the researchers report. "It's nice to see that the rate of serious deficiency has fallen," Shapiro says. "But I think most people would agree that 7 percent is too much."

The overall decline suggests FDA's audit program may deter some drug researchers from conducting sloppy or fraudulent research, Shapiro says. But high rates of deficiencies remain in certain areas, he adds. FDA found serious study-protocol violations--in which physicians failed to follow a detailed scientific method -- in 25 percent of audits performed before October 1985 and in 27 percent of audits performed since then. Researchers who dsiregard protocol can skew or invalidate study results, Charrow notes.

Even when investigators were found guilty of scientific misconduct, some escaped censure, the authors found. Shapiro and Charrow examined 395 "for cause" audits, which are more thorough investigations triggered when a routine audit reveals irregularities. FDA disqualified or placed restrictions on researchers in 16 percent of these cases. However, the vast majority--84 percent--did not result in any disciplinary action. In a small number of cases, FDA allowed researchers who had more than once deliberately violated regulations to continue their studies after they promised they wouldn't repeat past mistakes.

Researchers guilty of misconduct fall into several categories, Charrow says. Some deliberately fabricate data to advance their careers. Others cut corners in order to boost their research output and their earning capacity. Still others make mistakes through incompetence or inexperience.

FDA must get tough with researchers who repeatedly flaunt scientific standards, the authors argue. They propose suspending researchers immediately after an audit reveals substantial misconduct. Under the current system, such scientists may continue their studies pending a hearing. To weed out incompetent researchers, Shapiro and Charrow suggest that FDA give would-be investigators an examination, certifying those who pass to conduct clinical drug trials.

The authors stop short to recommendin g that the National Institutes of Health adopt a similar audit program -- an idea that has received a great deal of attention in the waske of several highly publicized cases of alleged fraud involving NIH-supported research (see box). Charrow points out that basic biomedical scientists must pass a rigorous peer review before getting NIH grant money, a process that helps eliminate shoddy researchers from the start. In contrast, investigators evaluating drugs for FDA approval contract directly with pharmaceutical companies. FDA can veto a firm's choice but does not put sicentists through a peer review, Charrow says.