E-records could test drug effectiveness

PHILADELPHIA, Feb. 10 (UPI) -- U.S. researchers predict national electronic medical records could test drug efficacy on a par with randomized trials.

However, researchers caution the need for at least 10 years of national data in the United States would be needed before electronic medical record data bases could be used to measure treatment efficacy.

"An appropriately configured electronic medical records data base could offer an invaluable tool, but we need to get to work now on how to configure it properly," study leader Dr. Richard Tannen of the University of Pennsylvania School of Medicine in Philadelphia says in a statement.

In the study, published in the British Medical Journal, Tannen's team examined outcomes measured by six previously performed randomized trials and the incidence of stroke, heart attack or death in a group of 8 million patients -- Britain's general practice research database.

After standard bio-statistical methods adjusted for differences in the database's treated and untreated groups there was no differences between database outcomes and randomized clinical trials in nine out of 17 outcomes. New biostatistical methods used to control for differences prior to the time the study began eliminated the remaining differences

The researchers say their findings show an electronic databases combined with new biostatistical methods could yield valid results.

"That's the real message of our paper -- this can work," Tannen says.