PREDPERF: Stata module module for calculating simple indices describing the accuracy and precision of one or two continuous-value predictors, based on a series of predictions matched with corresponding ‘true’ observations

In pharmacometrics, as in many other disciplines involving modeling, we regularly need to validate a model built to predict the values of a continuous marker by comparing its predictions with a series of “true” observations. We also often need to compare two models on the basis of their respective performances. A seminal paper by Lewis Sheiner and Stuart Beal, still widely cited today, laid the conceptual foundations for evaluating the predictive performance of a model for a continuous outcome variable (Sheiner LB, Beal SL. Some suggestions for measuring predictive performance. J Pharmacokinet Biopharm. 1981;9:503-12. doi: 10.1007/BF01060893). The accuracy of predictions is reflected in their mean bias relative to observations, while the precision of predictions is reflected in their root mean square residual error with respect to the observations. Sheiner and Beal showed how to estimate simple confidence intervals around these two indices, and how to test simple hypotheses. This unpretentious program implements these calculations in STATA, while providing an illustrative graph of the calculated indices over a scatterplot of observations versus predictions. It offers the option of calculating absolute bias and imprecision, or their relative counterparts (i.e. expressed as percentage ratios), or their logarithmic counterparts (i.e. based on the logarithm of the values compared).