What are the pills in your medicine cabinet, and how do you know they're best for you? When drug companies seek approval to market new medicines, they must show the U.S. Food and Drug Administration the results of all the tests they've run on volunteer patients - at first on only a few, then on dozens, and finally on hundreds or sometimes thousands. After winning approval, the companies typically sponsor reports of those tests in medical journal publications, which many doctors often rely on to determine whether to prescribe new drugs for their patients. Now a skeptical team of medical investigators at UCSF has accused the major drug companies of bias by distorting the results of their trials in those publications, making it hard for doctors to judge for themselves the pros and cons of prescribing the new drugs. As a result, the researchers say, patients may sometimes be taking medicines they don't need - or with unwanted side effects - that their doctors have prescribed on the basis of inadequate information. The UCSF team, led by Lisa A. Bero of the medical center's Institute for Health Policy Studies, probed the details of 164 drug trials involving as many as 1,500 patients over a two-year period and then examined reports on those trials that were published in medical journals, as well as those that remained unpublished. We found really important information from the official trial reports that were either not published at all or that stressed mostly the positive results of trials in the published versions," said Kristin Rising, a physician at the institute who did the major investigation.