Writing Conventions for Medicine
Abstracts are structured, and the headings are prescribed
Most clinical journals require labeled sections — Background, Methods, Results, Conclusions — inside a hard word limit, often 250 or 300. Every numeric result you claim in the abstract has to appear identically in the body, because reviewers check.
The reporting checklist depends on the design
Randomized trials follow CONSORT, observational studies follow STROBE, and systematic reviews follow PRISMA with its flow diagram. Journals ask for the completed checklist at submission, so it is cheaper to write the Methods against it than to retrofit later.
Report estimates with intervals, and register the trial first
Effect estimates carry confidence intervals; a p-value on its own is treated as thin reporting. Trials are expected to be registered before enrollment, and the registration number belongs in the abstract or the Methods.
ICMJE authorship and disclosure are checked at submission
The four ICMJE criteria decide who is an author and who belongs in acknowledgments. Conflict-of-interest, funding, and data-sharing statements are separate required sections, not optional courtesies.
Observational findings never get causal verbs
Cohort and case-control work says “was associated with” and lists the residual confounding you could not adjust for. Editors treat causal phrasing on non-randomized data as a substantive error, and it is one of the most common reasons for a revise decision.
Case reports are their own genre with a consent requirement
The structure is presentation, investigations, management, outcome, and a discussion of why this case adds something. Written patient consent for publication is required, and journals ask you to state that you obtained it.
