New algorithm detects adverse reactions caused by drug combinations

A new algorithm developed by researchers to search through the U.S. Food and Drug Administration’s (FDA) enormous and complex record of drug interaction data reveals what can happen when patients take prescription drugs in combination.

The FDA’s Adverse Event Reporting System (AERS) data, an ever-growing database of hundreds of thousands of drug reactions reported directly to the agency from health care professionals and consumers, provides valuable information about prescription drugs and their possible side effects. The new algorithm, however, unlocks more of the database’s potential by finding previously unknown relationships between different drugs and their consequences for patients.

Russ Altman, a Stanford University bioengineer and lead author of the study, called the algorithm “a step in the direction of a complete catalogue of drug-drug interactions.”

As the science journal Nature explained, drugs are tested for safety in pre-market clinical trials before they’re approved for use in the general population, but the size of a trial needed to discover all possible drug-drug interactions would make it an impractical, if not impossible, task. Drug trials normally do have strict criteria for participants, with their condition and other drugs they take noted, but “Once a drug hits the market … things can get messy as unknown side-effect pop up,” Nature explains. “And that’s where [the] Algorithm comes in.”

“Even if you show a drug is safe in a clinical trial, that doesn’t mean it’s going to be safe in the real world,” Paul Watkins, director of the Hamner–University of North Carolina Institute for Drug Safety Sciences, told Nature. “This approach is addressing a better way to rapidly assess a drug’s safety in the real world once it is approved.”

The algorithm, which automatically corrects for several sources of bias, such as age, gender, and disease, helped Altman and his team compile a database of more than 1,300 prescription drugs and potential side effects that were not listed on the label for those drugs. The results were astounding. Instead of an average of 69 side effects for each drug (as listed on their labels), the algorithm detected an average of 329 adverse side effects.

What the FDA will do with all this additional data remains unknown. Mr. Altman told Nature that he plans to present his findings to the agency, which could use the algorithm with its existing drug-surveillance programs to remove bias. But where regulators go from there is unknown.

“We’ve just released a database with 10,000 or more adverse events,” he told Nature. “I do not expect the FDA to uncritically take these results and add them to every drug label.”