Background: The introduction of evidence based medicine resulted in a clear shift from implicit to explicit reasoning in medicine, including the appreciation of multivariable diagnostic and prognostic prediction models. This is reflected by a sharp increase in published clinical prediction models. Systematic reviews try to assess and summarize the evidence. There is yet no tool to appraise clinical prediction studies.

Objective: To provide a comprehensive list of items that are relevant for systematically reviewing and critically appraising publications on clinical prediction models.

Methods: To indentify the relevant items for critically appraising clinical prediction research, we studied existing reporting guidelines (including in other medical research areas), various quality assessment tools, systematic reviews of prediction research, and methodologic key publications. Subsequently, experts were consulted for additional items.

Results: Items that are most important for systematically reviewing publications on clinical prediction research include study design, subject selection methods, assessment, definition, and coding of outcomes and candidate predictors, statistical power, statistical techniques used, the reporting and handling of missing values, predictor selection approaches, assessment of predictive performance and validation of the final model, and model presentation.

Conclusions: This overview of items can assist systematic reviewers in the appraisal of clinical prediction studies.