Why randomized controlled trials fail but needn't: 2. Failure to employ physiological statistics, or the only formula a clinician-trialist is ever likely to need (or understand!)

D L Sackett
2001 CMAJ - Canadian Medical Association Journal  
Because statistics has too often been presented as a bag of specialized computational tools, with morbid emphasis on calculation, it is no wonder that survivors of such courses regard their statistical tools as instruments of torture [rather] than as diagnostic aids in the art and science of data analysis. -George W. Cobb 1 From the underside It was the seventh time I had taken a course in basic biostatistics, and I vowed that this time I was actually going to understand it. The year was 1974,
more » ... nd we were on sabbatical in London. The Beatles had started to beat on each other, Ali beat Frazier, the good guys beat Nixon, Lord Lucan beat his nanny, and I could still beat the odds of getting run over while bicycling between NW3 and St. Thomas's Hospital. I had the time, I had the "coal-face" experience (by then I'd been a PI on several randomized controlled trials [RCTs]), and I already knew the English, Greek and Latin bits (I'd aced my 6 previous courses). Moreover, I'd had the good fortune to have worked with biostatisticians who were not only brilliant methodologists but also outstanding teachers. Alas, history repeated itself. At the end of my studies I was as incapable of applying this course to my current phase III RCTs as I had been incapable of applying my medical school biostatistics course to my patients 18 years earlier. Chastened, I returned home to find that the new crop of would-be trialists who also had successfully completed graduate courses in biostatistics were as confused as I when they tried to integrate how sick their prospective study patients might be, how well their outcomes might be ascertained, and how powerful their interventions might be with how many patients they might need to enrol and how certain they could be about any of their conclusions. Even today, 26 years later, the young clinical-practice researchers who come to our Trout Workshops up here in the woods still find it difficult or impossible to see the practical forest among the statistical trees.
pmid:11706914 pmcid:PMC81587 fatcat:frtiil3rsnhgdibh5md4kdjfsy