Wednesday, 17 August 2011

Paul Meier, RIP

Chances are pretty good you've never heard of Paul Meier.  But if you've taken an FDA-approved medication, or purchased life insurance, your life has been affected in a profound way by his work.

This weekend, read the announcement of the passing Dr Meier, who was one of the true giants of mathematical statistics.  In my opinion, his work with Edward L Kaplan (put the two names together, and perhaps if you've trained in stats or clinical research, you may be closer to recognition) has had more impact on clinical research than any other statistician, and perhaps more than any clinician of any other scientific discipline.

In 1958, Drs Kaplan and Meier submitted to the Journal of the American Statistical Association a manuscript titled "Nonparametric estimation from incomplete observations," in which a method for estimating life expectancy and mortality from data including missing or censored observations was outlined.

It's safe to say that there is almost no pharamacological research programme today that does not involve Kaplan-Meier curves.  From cancer survival to outcomes of heart surgery, this single paper is pretty close to fundamental in clinical trial design.

That same year, Dr Meier also authored a paper called "Clinical evaluation of new drugs," which was published in the journal Annual Review of Medicine, in which he argued that randomised controlled trials (RCTs) were essential in establishing the safety and efficacy of new medicines.

It's hard to fathom nowadays that there actually was an argument for assigning to random placebo and treatment arms patients to assess the effectiveness of a potential treatment, but it was quite controversial at the time.

Modern medical research is based upon the RCT, and countless lives have been saved/improved as a result.


No comments: