Lessons from case studies

In this final chapter, we review some practical issues of development, validation, and updating of prediction models, based on the empirical experience from case studies as described in this book, and modelling experience in other medical prediction problems. We consider the essential elements to successful modelling: sufficient sample size; emphasis on validation; using, not ignoring, subject matter knowledge. Tentative recommendations are made, recognizing that specific circumstances may ask for specific approaches. We end this chapter with a description of the case studies used throughout this book, where data sets are available through the book's Web site.