Skip to content
All articles

Statstics

How to Write a Research Paper Using IMRaD: Structure, Reporting Guidelines and Reviewer Pitfalls

Most medical research papers follow the IMRaD structure: Introduction, Methods, Results and Discussion, plus a title, abstract and references. Each section answers one question: why was the study done, how was it done, what was found, and what does it mean?

Resora Research Team · 15 Nov 2026 · 5 min read

Most medical research papers follow the IMRaD structure: Introduction, Methods, Results and Discussion, plus a title, abstract and references. Each section answers one question: why was the study done, how was it done, what was found, and what does it mean? Write the Methods and Results first, because they are factual, then the Introduction and Discussion, and the abstract last. Follow the reporting guideline for your design (CONSORT for trials, STROBE for observational studies, STARD for diagnostic accuracy, PRISMA for systematic reviews). Reviewers most often criticise an unclear question, an inadequate sample, missing confounder adjustment, results that go beyond the data, and a discussion that ignores limitations. Write to be understood, not to impress.

The IMRaD structure at a glance

SectionQuestion it answersTypical share of a paper
IntroductionWhy did you do it?10–15%
MethodsHow did you do it?20–25%
ResultsWhat did you find?25–30%
DiscussionWhat does it mean?30–35%

Write in this order: Methods → Results → Introduction → Discussion → Abstract → Title. Before writing, decide your one key message in a single sentence and choose the journal and read its instructions to authors.

Title

Specific, accurate and searchable. Include the design and population, and avoid abbreviations and claims such as "a novel study".

  • Weak: "Study of diabetes in rural people"
  • Strong: "Prevalence and risk factors of type 2 diabetes among adults in rural Rajasthan: a cross-sectional study"

Abstract

Most people read only the title and abstract, so make it stand alone. Use the journal's format, usually structured: Background, Methods, Results, Conclusion, in 150–300 words.

  • Give actual numbers in the Results (with CIs or p-values), not "a significant difference was found".
  • The Conclusion must follow from the results and answer the question.
  • No references, no new information that is absent from the main text.

Introduction

Three short paragraphs, shaped like a funnel:

  1. What is known: the burden or background, cited.
  2. What is not known: the gap or controversy.
  3. What you did: the aim, ending with the objective in one sentence.

Keep it to about 400–600 words. Don't review the whole literature, and don't state results here.

Methods

Give enough detail for another team to repeat your study. Follow your reporting guideline's checklist. Typical subheadings:

  • Study design, setting and period
  • Participants: eligibility criteria and how they were selected
  • Sample size: formula and assumptions
  • Variables and measurements: exact definitions, instruments, and validity
  • Procedures: interventions, blinding, follow-up
  • Statistical analysis: tests, software and version, significance level, handling of missing data and confounders
  • Ethics: committee approval (with number), consent, and trial registration where applicable

Use the past tense, and report what was done, not what was planned.

Results

Report findings, not interpretation. Move from the general to the specific:

  1. Participant flow: how many were approached, eligible, enrolled and analysed (a flow diagram for trials).
  2. Baseline characteristics: Table 1, with mean (SD) or median (IQR), and n (%).
  3. Primary outcome first, then secondary outcomes.
  4. Effect size with 95% CI and exact p-value, not p-values alone.
  5. Adjusted analysis: unadjusted and adjusted estimates.

Tables and figures should be understandable without the text. Do not repeat every table number in prose: highlight the key findings. Each table needs a clear title, units and footnotes for abbreviations.

Discussion

The discussion is where most papers go wrong. Use this order:

  1. Principal findings: one paragraph restating the main result in words.
  2. Comparison with other studies: agree or disagree, and why (different populations, definitions, methods).
  3. Possible explanations: biological or social, kept modest.
  4. Strengths and limitations: be honest about sampling, bias, confounding, sample size and generalisability. Reviewers trust authors who name their own weaknesses.
  5. Implications: what should clinicians, programmes or researchers do or study next.
  6. Conclusion: one short paragraph that answers the original question and goes no further than the data.

Do not simply repeat the results, introduce new results, or claim causation from an observational study. Words such as "associated with" are safer than "causes".

References

Use the journal's style (usually Vancouver). Cite recent, primary sources, check every reference against the original, and use a reference manager. Avoid excessive self-citation and unpublished sources.

Reporting guidelines: use the one that fits

Study typeGuidelineWhat it gives you
Randomised trialCONSORTChecklist and participant flow diagram
Observational (cohort, case–control, cross-sectional)STROBE22-item checklist
Systematic review / meta-analysisPRISMAChecklist and flow diagram
Diagnostic accuracySTARDChecklist and flow diagram
Case reportCAREStructure for reporting cases
Qualitative researchCOREQ / SRQRReporting of interviews and focus groups

The EQUATOR Network (equator-network.org) lists all guidelines. Most journals now require the completed checklist at submission.

What reviewers criticise most often

  1. No clear research question or primary outcome.
  2. Inadequate or unjustified sample size, so the study is underpowered.
  3. Inappropriate design or sampling, so the findings cannot be generalised.
  4. Confounding not addressed in the analysis.
  5. Wrong statistical tests (for example t-tests on skewed data, multiple tests without correction).
  6. Over-interpretation: conclusions that go beyond the data.
  7. Limitations ignored in the discussion.
  8. Poor English and unclear tables.
  9. Missing ethics approval or consent statement.
  10. Duplicate or redundant publication, and plagiarism.

Submission tips

  • Pick the journal first, and match your topic, audience and study type with its scope. Read recent issues.
  • Check that the journal is genuine: indexed (for example in PubMed/MEDLINE or Scopus) and transparent about peer review and fees. Be cautious of unsolicited invitations and promises of very fast publication.
  • Follow authorship criteria (ICMJE): substantial contribution, drafting or revising, final approval, and accountability. A guide is not an automatic co-author, and nobody should be listed as a favour.
  • Write a short cover letter stating the key message and why it fits the journal.
  • Run a plagiarism check and have a colleague read the full manuscript before you submit.
  • Don't submit to more than one journal at a time.
  • When reviewers respond, answer every comment politely in a point-by-point letter, and mark the changes in the manuscript. A request for revision is good news; most papers are published after revision.

Keywords

IMRaD structure, how to write a research paper, medical manuscript writing, introduction methods results discussion, STROBE checklist, CONSORT checklist, reporting guidelines, structured abstract, reviewer comments, how to publish a research paper, journal submission tips, scientific writing for doctors, ICMJE authorship, manuscript for postgraduates, discussion section

References

  • International Committee of Medical Journal Editors. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. icmje.org.
  • von Elm E, et al. The STROBE statement. Lancet. 2007;370:1453–1457.
  • Schulz KF, et al. CONSORT 2010 statement. BMJ. 2010;340:c332.
  • Page MJ, et al. The PRISMA 2020 statement. BMJ. 2021;372:n71.
  • Day RA, Gastel B. How to Write and Publish a Scientific Paper. Cambridge University Press.