Paul Meier, who was among the most influential biostatisticians of his generation and helped bring mathematical rigor to medical research in the years after World War II, died Aug. 7 at his home in Manhattan.
He had suffered a disabling stroke about 10 years ago and died after a series of recent strokes. He was 87.
(Family Photo) - Paul Meier, a biostatistician who co-invented the Kaplan-Meier curve, died Aug. 7 at his home in Manhattan.
As a biostatistician, Dr. Meier brought probability theory, data analysis and logic to bear on problems in biology and medicine. In the late 1950s, he co-invented the “Kaplan-Meier estimator” as a way of depicting survival and other important outcomes in medical experiments. It has been used in tens of thousands of scientific studies.
In addition to introducing analytical techniques, he was an early and successful proponent of “randomization” in clinical trials. The idea of assigning subjects in a medical experiment to one treatment or another solely on the basis of chance horrified many physicians.
In developing ways to more easily figure out which treatments worked and which ones didn’t, Dr. Meier “must already have helped save tens of thousands of lives,” said Richard Peto, a leading biostatistician at the University of Oxford in England. “I and hundreds of others use his methods every week in our work.”
Dr. Meier’s claim to fame was the two-dimensional, X- and Y-axis graph that can be found in the New England Journal of Medicine, the Lancet and dozens of other medical journals each week.
The Kaplan-Meier curve is a way of illustrating what happens to a group of subjects over the course of an experiment in which everyone starts in the same state: alive, disease-free, unpregnant, et cetera. If the subjects receive different “interventions” that affect health — say, half are getting an active drug and the other half a placebo — the graph depicts the two subgroups’ experiences as two diverging lines.
The problem that Dr. Meier and his collaborator, Edward L. Kaplan, solved was how to calculate an outcome — such as the chance of surviving five years after a cancer treatment — when not everyone participates in the experiment for the same length of time.
That problem is not uncommon, because clinical trials are time-consuming affairs and it often takes months or years to enroll enough volunteers.
The consequence is that a five-year study of survival after a cancer treatment might have only 50 or 60 percent of its subjects observed for the full five years. The rest are in for less time. Kaplan and Dr. Meier derived a series of equations that allows every patient’s experience — death or survival for whatever length of time observed — to contribute to the ultimate calculation of survival.
“It was a very, very important advance,” said Steven N. Goodman, associate dean for clinical research at Stanford University’s medical school and editor of the journal Clinical Trials. “It seems so elementary now.”
Kaplan, then at the University of California Radiation Laboratory, and Dr. Meier, then at the University of Chicago, published their 24-page paper in the Journal of the American Statistical Association in 1958.
Loading...
Comments