Confounder vs Effect modifier

Question 1

Confounding

Answer 1

Confounding is a systematic error where a third variable obscures the true association between exposure and disease

or Confounding occurs when a third variable distorts the true association between exposure and outcome.

The stratified should be similar but different from the crude OR

Question 2

Properties of Confounding

Answer 2

1. Be associated with the exposure
2. Be associated with the outcome
3. Not be in the causal pathway between exposure and disease

Question 3

Avoid confoundings

Answer 3

• Evaluating associations between variable and both exposure and outcome
• Stratification (Compare crude and stratum-specific estimates (e.g., ORs, RRs)
• Comparing crude and adjusted effect estimates
• Using directed acyclic graphs (DAGs) Identify confounding pathways.

Question 4

Control confoundings by phase

Answer 4

Design phase methods:
* Randomization Equalizes confounders across groups.
* Restriction Limit participants to exclude confounders.
* Matchings Pair participants based on confounder levels.

Analysis phase methods:
* Stratification Analyze data within strata of confounders.
* Multivariable regression analysis Adjust for confounders in regression.

Question 5

Effect Measure Modification (EMM)

Answer 5

EMM occurs when the effect of exposure on the outcome varies according to another variable, involving either:
* Synergism (positive interaction)
* Antagonism (negative interaction)

Effect modification describes differences in the exposure-outcome relationship by another variable.

The stratified must be different

• Reflects true biological phenomena, not error.
• Identifies subgroup differences, offering insights into causal mechanisms.

Question 6

Additive vs. Multiplicative Interaction

Answer 6

• Additive interaction: assessed on absolute risk scale (risk difference).
• Multiplicative interaction: assessed on relative risk scale (RR, OR)

Question 7

Avoid EMM

Answer 7

• Stratification
• Statistical tests of interaction

EMM should be reported and understood as it may represent the most significant study findings.

Stratification can help determine whether a variable is a confounder or effect modifier by comparing stratum-specific effect estimates to crude estimates