CoChat

Recommendations for presenting analyses of effect modification and interaction

Open access
Knol, Mirjam J, VanderWeele, Tyler J · 2012 · International Journal of Epidemiology

The DOI matches its Crossref record; 33 of the 35 references it cites were found in a registry or an index; 5 of the 32 sources its claims cite were checked for support.

Abstract

Authors often do not give sufficient information to draw conclusions about the size and statistical significance of interaction on the additive and multiplicative scales. To improve this, we provide four steps, template tables and examples to assess interaction on additive (RERI, AP, S) and multiplicative scales.

Overview

CoChat Summary

TL;DR

The paper recommends a four-step template for presenting effect modification and interaction analyses, distinguishing effect modification (one exposure across strata of another factor) from interaction (joint effects of two exposures). It argues that reporting stratum-specific effects, within-stratum effects, additive and multiplicative interaction measures with CIs/P-values, and confounders gives readers enough information to assess interaction on both scales.

Key findings

  • Proposes four pieces of information: (i) RRs, ORs or RDs for each (A, X) stratum with a single reference category, possibly the stratum with lowest risk of D; (ii) RRs, ORs or RDs for A within strata of X; (iii) interaction measures on additive and multiplicative scales with CIs and P-values; and (iv) the A–D confounders adjusted for. For interaction between A and B, step 2 also presents A within strata of B and B within strata of A, and step 4 lists A–D and B–D confounders.
  • A cited survey of 225 cohort and case–control studies found 61% addressed effect modification or interaction, but only 11% presented individual effects of both exposures and the joint effect of both exposures; most used multiple reference categories. In a random sample of 50 case–control and 25 cohort studies, only one reported a synergy index and none reported RERI.
  • Effect-modification example: antidepressant use and diabetes modified by chronic disease score. RRs were 1.16 (0.95–1.42) for antidepressant use only, 1.55 (1.30–1.84) for chronic disease score ≥1 only, and 2.07 (1.73–2.47) for both; RERI was 0.36 (95% CI −0.003–0.73), P = 0.052; ratio of RRs was 1.15 (0.89–1.49), P = 0.282.
  • Interaction example: dietary vitamin E and XRCC1 codon 399 genotype on prostate cancer. ORs were 1.04 (0.42–2.60) for Arg/Arg genotype only, 1.22 (0.53–2.82) for low vitamin E intake only, and 2.40 (1.02–5.63) for both; RERI was 1.14 (95% CI −0.57 to 2.85), P = 0.19; ratio of ORs was 1.89 (0.56–6.34), P = 0.30.

Methods

  • Methodological recommendations paper with template tables and examples, distinguishing “effect modification” (causal effect of intervening on one exposure across strata of another factor) from “interaction” (causal effect of intervening on two exposures).
  • Uses two illustrative examples: a cohort study by Knol et al. on antidepressant use and diabetes modified by chronic disease score, and a case–control study by van Gils et al. on dietary vitamin E and XRCC1 codon 399 genotype on prostate cancer risk.
  • Analysis focuses on presenting RRs, ORs or RDs with CIs and P-values, calculating RERI and ratio of RRs/ORs for additive and multiplicative interaction, using the delta method for CIs, and recommending RRs over ORs when possible.

Limitations

  • Stated: the paper focuses on RERI as the additive interaction measure; the synergy index is also sometimes considered but should only be used if the doubly unexposed group has the lowest absolute risk. ORs are non-collapsible, so effect modification may vary by covariates, and when the outcome is common ORs may exaggerate interaction measures. In the gene–environment example, population stratification was not controlled, so causal interaction interpretation depends on absence of genetic confounding.
  • Evident from the text: the recommendations are prescriptive and template-based rather than empirically validated in this paper; the illustrative examples show interaction estimates whose CIs include the null, so they demonstrate presentation format rather than confirm interaction.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1093/ije/dyr218

Cites

33 of 33 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

3 Valid · 0 Not valid · 16 Not checked · 11 Own results and interpretations

2 more claims

Create a free account to see all 30 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 12
  • Probably not a substantial claim. 15

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 33 references

Research works printed but not linked: 2

  1. Rothman, 1986, Modern Epidemiology
  2. Rothman, 2002, Measuring interactions, 168

Cited by

See all 1,494

Metadata

Authors
Knol, M. J., & VanderWeele, T. J.
Affiliations
Harvard University (US)
Venue
International Journal of Epidemiology
Published
2012-01-09
Volume
41
Issue
2
Pages
514-520
ISSN
  • 0300-5771 (ISSN-L)
  • 1464-3685
OpenAlex
W2114647004
Cited by
1,494
References
35
Citations by year
2026: 292 · 2025: 187 · 2024: 161 · 2023: 153 · 2022: 125 · 2021: 123 · 2020: 92 · 2019: 89 · 2018: 64 · 2017: 74 · 2016: 43 · 2015: 32 · 2014: 26 · 2013: 21 · 2012: 7
FWCI
24.79
Type
journal article
Language
en
Open access
Locations
  • International Journal of Epidemiology (publishedVersion)
  • Europe PMC (PubMed Central) (publishedVersion)
  • PubMed Central (publishedVersion)
  • PubMed (publishedVersion)
  • HighWire Press Open Archive (submittedVersion)
Topics
  • Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
  • Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
Keywords
effect modification, interaction, reporting recommendations, statistical tables, relative risk, odds ratio, risk difference, confounding adjustment, causal inference
MeSH terms
Humans, Risk Factors, Models, Statistical, Causality, Effect Modifier, Epidemiologic, Epidemiologic Studies

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-07 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Recommendations for presenting analyses of effect modification and interaction. https://app.cochat.ai/papers/10.1093/ije/dyr218
The paper
Knol, M. J., & VanderWeele, T. J. (2012). Recommendations for presenting analyses of effect modification and interaction. International Journal of Epidemiology. https://doi.org/10.1093/ije/dyr218

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.