The disease detectives. Pt. 4, Epidemiology IV; epidemiology and statistics.

Statistics are essential for epidemiology, but they can be interpreted in different ways. Studies of oral contraceptives have indicated that there could be increased risk with third generation pills. Could the experts be influenced by whether they are paid by drugs companies or by patients? How important is the idea of statistical significance? Is it an arbitrary hurdle? How do policy makers decide when it is necessary to act? Richard Lilford explains the statistical theories of Thomas Bayes (d.1761). Prior belief modified by data gives posterior belief. Life is about degrees of uncertainty. Lilford says it is impossible to eliminate subjectivity from interpretation of statistics.