Confounding adjustment strategy
Causal question → covariates → adjustment method → diagnostics → analysis strategy
Flow diagram
Steps
Causal question and target effect
Define exposure strategies and target population/estimand.
Causal covariates and temporal ordering
Choose confounders using substantive knowledge; distinguish mediators and colliders.
Adjustment strategy
Specify outcome modeling, stratification, matching, weighting or another justified approach.
Overlap and balance diagnostics
Assess support, balance, extreme weights and model adequacy.
Revise strategy or restrict target
Prespecify responses to inadequate overlap and record any changed target population.
Estimate effect and assess residual bias
Record uncertainty estimation and sensitivity to unmeasured or mismeasured confounding.
Tables and supporting information
Covariate rationale
Do not automatically include post-exposure variables.
Adjustment diagnostics
Planned criteria and completed results must remain distinct.
Identification assumptions
Record exchangeability, consistency, positivity and model assumptions appropriate to the chosen method.
Causal rationale and evidence
Attach clinical and methodological support; the editor does not discover confounders.