Being exceptionally fond of statistics, and even more so, aware of their limitations, I’m often delighted with when those limitations are discovered.
In this article by the Journal of the American Medical Association on Risk Prediction Models for Hospital Readmission the authors looked at a large group of studies and tested for their predictive statistical significance with interesting results.
This remains yet another indication of our limitations in treating “populations” rather than patients. Despite how exquisitely educated we may be or think we are, the incredible complexity of human biology and spirit requires that humility be practiced, as well as science.
