Prediction performance and usefulness
Assess calibration discrimination and utility
Calibration concerns agreement between predicted probabilities and observed frequencies; discrimination concerns ranking. Report both with uncertainty. Decision utility depends on plausible thresholds and consequences, not only a high area under the curve.
Flow diagram
Steps
Define performance targets
Choose measures appropriate to outcome type and prediction horizon.
Assess calibration
Plan calibration plots, intercept/slope or outcome-appropriate calibration measures.
Assess discrimination and overall error
Specify discrimination, scoring rules and censoring-aware methods where needed.
Assess clinical utility
Specify decision thresholds, consequences and net-benefit analysis only where applicable.
Assess subgroups and uncertainty
Prespecify subgroup evaluations, precision and limitations.
Prediction performance and usefulness
End flow
Tables and supporting information
Performance assessment plan
Specify measures before analysis; leave estimates empty.
Assumptions and rationale
Good discrimination does not imply calibration or clinical benefit. Account for censoring when evaluating time-to-event predictions.
Method and protocol evidence
Attach the protocol, methodological sources and versioned software specification supporting these decisions. Record citations and provenance; do not enter participant identifiers.