A decision-analytical perspective on incorporating multiple outcomes in the production of clinical prediction models: defining a taxonomy of risk estimands

• DOI: 10.1186/s12916-025-03978-3
• Publication Date: 2025-03-06T15:23:19Z
• Authors: Glen P. Martin; Alexander Pate; Siân Bladon; Matthew Sperrin; Richard D. Riley

Clinical prediction models (CPMs) estimate an individual's risk of current or future outcome events, using information available about the individual at time prediction. While most CPMs are developed to predict a single event, many clinical decisions require considering risks multiple events. For example, decision-making for anticoagulation therapy involves assessing both blood clot and bleeding, while around interventions multimorbidity prevention requires understanding developing long-term conditions. However, determining when how incorporate outcomes into remains challenging. This article aims raise awareness present examples where such is essential help inform decision-making. A range analytical methods develop multiple-outcome CPMs, but there frequent malapropisms heterogeneity in terminology across this literature, making it difficult identify/compare possible methods. Selecting appropriate method should depend on intended estimand-the type predicted that we wish CPM estimate-but often not defined reported. Using decision-analytical perspective, taxonomy estimands frame different contexts requiring CPMs. We outline four levels estimands: (i) single-outcome risk, (ii) competing-outcome (iii) composite-outcome (iv) combinations. demonstrate utility-theory lens can define estimand given scenario, based model's use. Clearly defining reporting all model studies. framework aids selecting task development.