Several dozen neuropsychiatrists and others, at a recent meeting in Washington, D.C., emphasized the need for a national postmarketing surveillance system for drugs. But how will the country move to such a system?
"Right now the whole country is moving toward an electronic medical record," said Charles O'Brien, M.D., Ph.D., vice chair of psychiatry at the University of Pennsylvania and moderator of the meeting, convened by the American College of Neuropsychopharmacology (ACNP). He is hopeful that if medical records are linked with pharmacy records, side effects that occur perhaps only in every 10,000 patients can be detected relatively early.
The current system, which relies on premarketing testing, is intrinsically limited, he said. Drugs are tested on perhaps a few thousand patients. Such a small sample size will not detect a problem that is relatively rare but possibly severe, said O'Brien. The challenge is to ensure that newly developed EHR systems can be integrated into databases to track medications after they are on the market.
But Herings told the meeting that his country's system had hardly been a panacea. Problems range from changes in government regulations to charges that studies produced from it have favored pharmaceutical or other interests. The quality of the data entered into the system by physicians and pharmacists, said Herings, is supported by Netherlands laws requiring strict recording of dispensed drugs. The system is now being tested in Germany and Poland, but it will be some years before the results of that expansion are known.
Strom agreed that the current system is inadequate, pointing to reports that 51% of drugs that make it to market eventually have label changes because of major safety issues discovered after marketing. And, he asserted, "I am at least as worried about the remaining 49%, the ones that we don't know about because of limitations of available studies."
However, he pointed in particular to claims data such as those from Medicaid and managed care organizations as good sources with which to do research. Although each has its limitations, such as skewed populations or incomplete inpatient or outpatient data, they also have tremendous advantages, including the size of their populations, he argued. Strom also indicated that there are now databases containing actual electronic health records that are "extraordinarily rich," although they are incomplete in terms of hospitalization records.
"Do the databases need improvement? Absolutely," Strom asserted. Work needs to be done on technologic issues and methodology development. But, he said, "I would argue that this is all a question of brains."
A statement issued after the meeting said that, among other things, the group encouraged ACNP to consider establishing fellowships in pharmacoepidemiology to promote the field and to facilitate training. It added that individuals with solid methodological expertise will be essential to the collection and wise use of data.
In an interview with Drug Topics at the meeting, O'Brien called it a hopeful sign that as part of its push for EHRs, the Department of Health & Human Services (HHS) has a major contract to begin harmonizing the various languages used in EHR systems.