*To mask research question/hypothesis to participants
|}
==Confounding and Effect modification==
{{#mermaid:
flowchart TD
crude[Crude analysis] --> strata[Stratified analysis]
strata --> node1[[Are stratified estimates similar to each other?]]
node1 -- "YES, similar to each other" --> MH[Mantel-Haenszel adjusted analysis]
MH --> node2[[Is MH adjusted estimate similar to crude estimate?]]
node2 -- "YES, similar to crude" --> No([There is no Confounding nor Effect modification]) --> R1["Report crude (MH adjusted) estimate"]
node2 -- "NO, different from crude" --> Co([There is Confounding]) --> R2[Report MH adjusted estimate]
node1 -- "NO, different from each other" ----> EM([There is Effect modification]) --> R3[Report stratified estimates]
R2 -.-> node3[[Are all possible confounders measured<br>and all possible stratification conducted?]]
node3 -- "YES, rigorous" --> noR([MH adjusted estimate should be<br>closer to the true population value])
node3 -- "NO, unfeasible" --> yesR([MH adjusted estimate should<br>still include residual confounding])
}}