The Cochrane Handbook for Systematic Reviews of Prognosis Research and Prediction Models is the very first of its kind. Understanding, summarizing and predicting outcomes in populations and individuals is essential to modern healthcare planning and decision making, especially with the advent of personalized or stratified care. As such, there are thousands of new prognosis studies published each year in the health literature, many of which now use approaches attributed to artificial intelligence including machine learning. This has led to a rapid rise in prognosis research findings, including tens of thousands of prediction models (predictive algorithms) purporting to predict and inform health outcomes in individuals and populations. However, the quality of primary prognosis studies is known to be highly variable, and there are often conflicting reports about the usefulness of particular prognosis findings or specific prediction models.
Systematic reviews of prognosis research and prediction models are needed more than ever before. This Cochrane Handbook provides a formal, transparent and rigorous framework to identify, appraise and summarize existing prognosis research evidence to inform healthcare and health policy. It describes the concepts of prognosis and (primary) prognosis research in healthcare, and details the entire process of preparing, conducting, analyzing, reporting and maintaining systematic reviews of the three main types of prognosis research: overall prognosis, prognostic factors and prediction models. Indeed, this handbook essentially covers three handbooks in one.
This Cochrane Handbook has been produced by the Cochrane Prognosis Methods Group under the senior leadership of Carl Moons and Richard Riley. It provides a foundational framework, with recommendations and a step-by-step guide for anyone involved in planning, conducting or appraising systematic reviews of prognosis research and prediction models. As such, it serves as a pivotal reference for new and experienced reviewers alike.
Part I: Prognosis and Prognosis Research
Part II: Methods for Reviews of Prognosis Studies
- Data extraction from prognosis studies
- Extracting and deriving statistical estimates from prognosis studies
- Risk of bias and applicability assessment
- Meta-analysis I: key principles and a general inverse-variance approach
- Meta-analysis II: subgroup analysis, meta-regression and small-study effects
- Meta-analysis III: one-stage models and advanced approaches
- Summary of findings tables and grading certainty of evidence
- Reporting the prognosis review
Part III: Special Topics
Important notes
- A book version (first edition) of this Handbook, to be published by Wiley in 2026, will comprise chapters 1-17.
- Chapters are freely available for personal use only via a Cochrane Account (don't have an account? Set one up here). These chapters are not for general distribution, and all content remains the copyright of Cochrane.
- These drafts will undergo further work. Further drafts may be posted here before finalization.
- If you would like feedback on any draft chapters, please use this form. Your comments will be considered in finalizing the drafts for the printed copy of the handbook.
- If you use the chapters in your work, please cite the material as set out at the beginning of each chapter, including the exact date of the version.
Editorial team
- Karel GM Moons (Editor)
- Richard D Riley (Editor)
- Anneke Damen (Associate Editor)
- Lotty Hooft (Associate Editor)
- Alfonso Iorio (Associate Editor)
- Nicole Skoetz (Associate Editor)
- Katrina Williams (Associate Editor)
