Why this discipline needs its own guide
Background
Epidemiology sits between the structured trials world and the more interpretive social-science world. A typical observational study involves a small core team that defines the exposure and outcome variables, a larger team that curates and harmonises data from registries or biobanks, a statistician (often shared across studies) and clinical-discipline experts who anchor the substantive interpretation. CRediT maps reasonably onto that structure provided Data Curation is recognised as the substantive role it is.
Reporting against the STROBE checklist (or its disease-specific variants such as RECORD for routinely collected health data) supplements but does not replace the contributor statement. Pre-registration of the analysis plan on the Open Science Framework or in a journal-specific Registered Report is increasingly expected and is recorded under Methodology.
Key considerations
How to assign the roles
- Data Curation in epidemiology is substantial: harmonising variables across cohorts, defining inclusion windows, deriving phenotypes from coded health-record data. Name the contributors who did this work.
- Linkage of routine health-record datasets carries governance obligations that are themselves a Methodology contribution, not just a Resources one.
- Sensitivity analyses, multiple-imputation schemes and instrumental-variable approaches are Formal Analysis. Where these were pre-specified versus post-hoc, the manuscript should state so but the CRediT role does not change.
- Cohort principal investigators who provide access to the underlying data without further intellectual contribution belong under Resources, not Investigation.
- For meta-analyses, the protocol registration (PROSPERO) and the search strategy belong under Methodology; screening and extraction are Investigation; analysis is Formal Analysis.
Reporting Guideline Integration
STROBE & PRISMA to CRediT Crosswalk
Mapping Observational Epidemiology and Systematic Reviews to CRediT Roles
Observational epidemiology and systematic reviews require highly systematic metadata collection, literature searching, and statistical modeling. This crosswalk aligns the STROBE checklist and PRISMA meta-analysis steps with CRediT contributor roles.
| Checklist Item / Phase | Mapped CRediT Role(s) | Guidance & Practical Allocation |
|---|---|---|
| STROBE ObjectivesFormulating scientific research hypotheses, background questions, and aims. | ConceptualizationWriting – Original Draft | The intellectual framing and epidemiological hypotheses are Conceptualization. Writing the rationale is Writing – Original Draft. |
| STROBE Study DesignSetting up observational cohort, case-control, or cross-sectional structures. | ConceptualizationMethodology | Designing variables, defining exposure timelines, and selecting case-control matching criteria is Methodology. |
| STROBE Setting & SubjectsCohort recruitment, geographical enrollment, database sources, and follow-up. | MethodologyInvestigation | Structuring sampling frameworks is Methodology. Interviewing subjects, reviewing electronic health records, and measuring exposures maps to Investigation. |
| STROBE Variables & Data SourcesDefining exposures, outcomes, and confounders. Extracting electronic health records. | MethodologyData Curation | Developing definitions for outcome variables is Methodology. Harmonizing, cleaning, and cataloging electronic health records or biobank databases maps to Data Curation. |
| STROBE Statistical MethodsControlling for confounders, propensity scores, and sensitivity analyses. | Formal AnalysisSoftware | Running regression models or matching programs is Formal Analysis. Writing custom script pipelines in SAS, R, or Stata maps to Software. |
| PRISMA Search StrategyDeveloping and running exhaustive search strings across databases. | MethodologyInvestigation | Designing the query strings and selecting medical registries is Methodology. Running the searches and screening abstracts against inclusion criteria is Investigation. |
| PRISMA Data ExtractionExtracting study features, effect sizes, risk ratios, and demographics. | Data CurationInvestigation | Extracting raw clinical trial data maps to Investigation. Assembling these extractions into systematic repositories and checking quality is Data Curation. |
| PRISMA Synthesis & PoolingMeta-analysis statistics, forest plots, funnel plots, and bias analyses. | Formal AnalysisSoftwareVisualization | Running pooled risk ratios and modeling heterogeneity is Formal Analysis. Generating the final forest and funnel diagrams maps to Visualization. |
Worked example
A representative CRediT statement
Author Contributions (CRediT) S. Kowalski: Conceptualization, Methodology, Formal analysis, Writing – original draft. R. Mehta: Methodology, Data curation, Software. M. Ndiaye: Data curation, Investigation. J. Park: Formal analysis, Visualization. A. Bianchi: Conceptualization, Supervision, Writing – review & editing. Pre-registered analysis plan: OSF osf.io/XXXXX.
The role names above match the canonical wording at casrai.org/credit. Most publishers accept exactly this format.
Further reading
Discipline-specific sources
Common questions
Frequently asked
How do I assign CRediT roles for data harmonisation in an epidemiological study?
Data Curation in epidemiology is substantial work and should be recognised explicitly: harmonising variables across cohorts, defining inclusion windows, and deriving phenotypes from coded health-record data all belong under Data Curation. Name the contributors who did it. Note that linkage of routine health-record datasets carries governance obligations that are themselves a Methodology contribution, not merely Resources.
Does the STROBE checklist replace the CRediT statement?
No — STROBE (and disease-specific variants such as RECORD) is a reporting checklist for observational studies; it supplements but does not replace the contributor statement. CRediT records who did what, while STROBE governs how the study is reported. Pre-registration of the analysis plan, for example on the OSF, is recorded under Methodology.
