Researchers at the Stanford University School of Medicine have developed a computer algorithm which has enabled them to identify "true" drug side effects in millions of reports made to the FDA by patients and physicians. Published on March 14 in Science Translational Medicine, the research also identifies previously unknown interactions between pairs of drugs. The first author, Stanford graduate student Nicholas Tatonetti, developed a method that would allow a case control study within that data by matching groups of people with the most similarities possible, save for one drug variable, which allowed them to tell if the drug was the cause of an adverse event report. Tatonetti and senior author, Russ Altman, Stanford professor, used the technique to analyze reports in the FDA's Adverse Event Reporting System to discover unidentified side effects and interactions, and then tested their prediction by analyzing electronic health records of patients at Stanford Hospital & Clinics. They confirmed 47 new drug interactions identified in the study of AERS. Two databases were built to make the work publicly available; OFFSIDES contains 1,332 drugs each with an average of 329 new adverse events, while TWOSIDES includes 59,220 pairs of drugs, with a total of 1,301 adverse events that cannot be clearly caused by either drug alone. It is believed that the databases will help physicians better tailor prescriptions to their patients, and will further drug discover by finding drugs that cause similar side effects, an indication that they may affect the same biological pathways and have more than one possible clinical indication.