Index Last Issue Contact DQ News Latest Issue
'Evidence-Based Medicine' - Data Quality Implications
A December 16th article in the "Science Times" Section of The New York Times examines the implications of "evidence-based medicine" on the American healthcare system. According to the Times, evidence-based medicine trains doctors to search medical journals and databases for tests and treatments that have helped large numbers of patients. Then, doctors apply the data to their own patients, under the statistical assurance that what holds true for groups is likely to be valid for individuals.
Evidence-based medicine involves trying whenever possible to base medical decisions on sound research data. The strategy is being widely hailed in the U.S., Britain, and Canada as a long-sought method to allow the results of the best medical research to hospitals and doctors' offices where medical decisions are made.
According to the Times, doctors usually usually rely on a combination of habit and casual intuition, using treatments they are familiar with or heard good things about. Doctors who practice using evidence-based medicine search the medical literature for specific research studies or randomized controlled trials. About a million randomized controlled trials have been completed and reported in the past 30 years. Using the Internet, a doctor can find relevant data in a few minutes. Proponents of evidence-based medicine believe that the methodology will eventually cleanse medical care of obsolete and dangerous practices.
Critics of evidence-based medicine point out that randomized controlled trials often exclude groups of patients who are elderly, sick, or don't meet a study's criteria. They believe that outcomes analysis (which looks at the medical histories of large groups of patients) is a better predictor of a medical treatment's value. Computer searches of large patient databases can indicate what test, treatment, or doctor led to the best outcome in a different situation.
Unfortunately, there is nothing in the Times article about the data and information quality of the databases being searched. Nor is there any mention of data quality problems with randomized controlled studies and patient medical records. The article was written by Times science reporter Abigail Zuger, and appears on page C1.