Matched risk-set analysis and sensitivity
Respect sampled sets and assumptions
The analysis must respect risk-set sampling and matching. State any sampling weights or corrections and how repeated appearances enter uncertainty estimation. Conditional likelihood notation alone does not specify every sampling design.
Flow diagram
Steps
Define sampled-cohort contrast
Link the relative rate or hazard contrast to incidence-density sampling. Controls may recur and later become cases; the interpretation does not require a rare-disease approximation.
Choose matched analysis
Specify conditional likelihood or another sampling-compatible estimator.
Respect reuse and dependence
Account for matching, repeated controls and sampling structure as required.
Plan sensitivity analyses
Assess matching choices, exposure lags, missing measurements and exclusions.
Specify precision and reporting
Document uncertainty, sampling limitations and cohort-to-sample accounting.
Matched risk-set analysis and sensitivity
End flow
Tables and supporting information
Matched analysis plan
Justify the estimator using the actual sampling process.
Assumptions and rationale
An unmatched generic case-control analysis may not preserve the risk-set design or target contrast.
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.
Conditional likelihood contribution
Sⱼ is a sampled matched set containing one case c and its controls; x is a scalar exposure. A model combines contributions across sets. Under appropriate incidence-density sampling, exp(β) estimates a relative rate or hazard. This illustrative form assumes sampling compatible with the conditional likelihood. Unequal sampling probabilities, additional covariates, repeated sampling and uncertainty need explicit treatment; the rare-disease approximation is not required for the rate-ratio interpretation.