PREDICTIVE DIAGNOSTICS FOR LOGISTIC-MODELS

Novel methodology is implemented to assess the predictive power of cov ariate information associated with sequential binary events. Logistic models are first fitted on the basis of a subset of the observations a nd then evaluated sequentially on the rest. The probabilistic forecast s are compared to the outcomes via a scoring function, but as most val idation samples are small, the usual reference distribution for the te st statistics is inadequate. However, bootstrap-based distributions ca n easily be constructed. The first example pertains to the evaluation of screening tests for major depression. It illustrates that goodness- of-fit and predictive assessments lead to the selection of very differ ent models. The second example deals with the prediction of a major ev ent in the natural history of HIV-induced disease. It shows that this type of analysis can reveal features missed by other approaches.