Contents
On this page
- Historical development of the IMRAD structure
- Functions of each manuscript section
- The title and abstract as structural components
- Common structural errors
- Example of a results paragraph structured according to IMRAD
- Reporting guidelines as a structural overlay
- Administratively checked structural requirements
- Allocating the word limit across sections
- Key points
- Application to your study
- Frequently asked questions
Academy · Research Writing · Structure of a medical paper
The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines
IMRAD (Introduction, Methods, Results and Discussion) is not merely a formatting convention. Each of its four sections addresses a different question, and most structural failures in submitted manuscripts arise from placing content in the wrong section. Sollaci and Pereira’s survey of 1,297 articles demonstrates how recently, and how completely, this structure became standard.
Authors instructed to “follow IMRAD” frequently treat it as four headings to be completed. This produces papers in which the introduction reviews the literature rather than posing a question, the results interpret rather than report, and the discussion introduces numbers that appear nowhere else. Reviewers may interpret such features as carelessness or, more seriously, as selective reporting. This article sets out the function of each section, the most frequently recurring structural failures, and the correspondence between reporting guidelines and manuscript structure, so that the checklist serves as a drafting tool rather than a submission formality.
Learning objectives
- The question each IMRAD section is intended to answer, and the content that belongs exclusively to it
- The recency of the convention: 0% of 1935 articles against over 80% by the 1970s, across 1,297 papers
- The structural failures that reviewers interpret as selective reporting rather than as poor organisation
- The use of CONSORT, PRISMA 2020 or STROBE as drafting aids rather than as checklists completed after drafting
- Converting a journal’s word limit into a per-section plan before drafting
- The ICMJE structural requirements that are checked administratively and may delay a submission
Common misconceptions
- “IMRAD is the natural shape of a scientific paper.” It has been the norm for only approximately fifty years. Sollaci and Pereira found no article in their 1935 sample using it; the New England Journal of Medicine adopted it fully in 1975, the BMJ in 1980, JAMA and The Lancet in 1985.
- “The introduction is where the literature is reviewed.” Its function is to narrow the topic to one answerable question. A literature review that does not narrow the topic is the most common reason that an introduction is overlong and uninformative.
- “Results and discussion can be combined if the writing flows.” Interpretation within the results removes the reader’s ability to disagree with the authors’ reading of their own data, which is precisely what a reviewer seeks to do.
- “Reporting guidelines are a submission checkbox.” STROBE has 22 items, 18 of which are common to cohort, case-control and cross-sectional designs. Each unmet item is likely to be raised by a reviewer as a criticism if it is left unaddressed.
- “Structure is cosmetic; the science is what counts.” Structure is the means by which the science becomes verifiable. For the purposes of review, a result that a reader cannot locate is equivalent to a result that was not reported.
Historical development of the IMRAD structure
The four-part structure is a twentieth-century standardisation, not a timeless form. Sollaci and Pereira, in the Journal of the Medical Library Association, surveyed 1,297 original articles from the BMJ, JAMA, The Lancet and the New England Journal of Medicine, sampled at five-year intervals from 1935 to 1985, and coded each for structure with an interobserver agreement of 0.95 20. No article in the 1935 sample used IMRAD. By the 1970s the structure accounted for more than 80% of original papers 20. Full adoption arrived journal by journal: the New England Journal of Medicine in 1975, the BMJ in 1980, and both JAMA and The Lancet in 1985 20.
Historians of the form describe the same development from the other direction. Atkinson, in Applied Linguistics, traced 250 years of the Edinburgh Medical Journal and showed research writing moving from narrative author-centred report toward impersonal, method-centred argument 23. Bazerman documented the same shift in physics in Social Studies of Science, and Meadows, in the Journal of Information Science, read the paper as an artefact shaped by the needs of its readers 25,26. Vickery’s account in the Journal of Documentation of the 1948 Royal Society Scientific Information Conference records the moment when the pressure for a standard, scannable format became explicit 22.
The practical consequence is that IMRAD is a convention designed to allow a reader to locate one thing quickly: what was done, to whom, and what happened. It was not designed to make writing easier. It was designed to make reading and verification easier, which is why deviations are penalised even when the prose is good.
Articles from the BMJ, JAMA, The Lancet and the New England Journal of Medicine, sampled at five-year intervals. Source: Sollaci LB, Pereira MG. The Introduction, Methods, Results, and Discussion (IMRAD) structure: a fifty-year survey. Journal of the Medical Library Association.
Functions of each manuscript section
Each section answers one question and should answer no other. The introduction answers why this question, and why now. It moves from the accepted position to the specific gap, and ends with the question or objective in one sentence. It does not summarise the methods, does not preview the findings, and is not a literature review. Kliewer, writing in the American Journal of Roentgenology, describes the typical shape as three paragraphs before the purpose statement — context, gap, purpose — which is a useful discipline because it requires a decision about what the gap is 19.
The methods section answers could someone else do this and judge whether it was done well. It is the section that reviewers scrutinise most closely, and the one in which vagueness is most costly: “appropriate statistical tests” asks a reviewer to accept the analysis on trust at precisely the point at which scrutiny is expected. Everything a reader needs to assess bias belongs here — design, setting, eligibility, how participants were selected or allocated, what was measured and with what instrument, the analysis population, and how missing data were handled.
The results section answers what happened, and nothing else. No interpretation, no comparison with other studies, no speculation about mechanism. Numbers appear here first and only here; a figure in the discussion that is not in the results is read as the product of a post hoc search. The discussion answers what it means and how far it can be trusted: the principal finding stated plainly, its relation to existing work, the plausible mechanisms, the genuine limitations, and the next steps. Kliewer describes around five discussion paragraphs before the closing summary, on a manuscript of roughly 18 paragraphs across five sections 18,19.
In summary: the introduction poses, the methods enable judgement, the results report, and the discussion interprets. A sentence that performs a different function from its section is misplaced.
The title and abstract as structural components
For most readers the abstract is the paper. It is also, in practice, where editorial triage occurs. Huth’s short piece in the Annals of Internal Medicine made the case for structured abstracts in trial reports for precisely this reason: an unstructured abstract allows an author to omit the design, the population or the effect size without the omission being visible 24. A structured abstract makes the gap conspicuous.
Journal guidance on this point is less consistent than authors assume. Fujita and colleagues, in European Science Editing, examined 64 journals indexed in Journal Citation Reports and found abstracts addressed in author guidance by 97% of Communication Science journals but 72% of Library and Information Science journals, titles by 72% against 53%, and keywords by 88% against 66% 8. The appropriate inference is not that guidance is optional but that its absence does not mean the expectation is absent — the expectation must instead be inferred from recently published papers in the target journal. The abstract writing page covers the structured forms in detail.
Two rules apply in every variation. The primary outcome should be stated with a number, not a direction. The conclusion should not exceed what the results support — the most common reason an otherwise sound abstract reads as overclaiming.
Common structural errors
These are not stylistic preferences. Each one either makes a claim uncheckable or makes it look as though something is being hidden.
- Interpretation inside results. “Importantly, mortality was reduced” is a discussion sentence placed in the results. It pre-empts the reader’s judgement.
- New numbers in the discussion. Any figure appearing first in the discussion reads as a post-hoc analysis, whether or not it was one.
- An introduction that does not end in a question. If the objective is not stated in one sentence, no reviewer can tell whether the design answers it.
- Methods reconstructed from the results. Analyses described in the order in which they were eventually run rather than as pre-specified or exploratory. If an analysis was exploratory, this should be stated; otherwise a reader may assume it and distrust the rest.
- Limitations as a disclaimer paragraph. A list of generic weaknesses (“single centre, modest sample”) with no statement of the direction in which each would bias the result. Reviewers expect the direction to be stated.
- Results that do not map onto outcomes. A primary outcome named in the methods but reported third, after two secondary outcomes that were statistically significant.
- Tables and text duplicating each other. Every number reported in both places doubles the opportunity for a discrepancy, and reviewers detect discrepancies.
The remedy for most of these is mechanical. Print the manuscript, mark each sentence with the section to which it belongs, and relocate the mismatches. This is a tedious exercise that removes several reviewer comments.
Example of a results paragraph structured according to IMRAD
The difference rarely concerns language. The undisciplined version mixes reporting with interpretation and leaves the reader unable to separate the data from the authors’ reading of them. The disciplined version reports, and reserves the interpretation for the section designed for it — which is also where a reviewer is prepared to debate it.
The same findings, reported two ways (a constructed specimen — the figures are illustrative, not results from a real study)
Results doing the discussion’s job
Our findings were encouraging. Patients in the intervention arm did considerably better, with a clear reduction in symptom scores that was statistically significant, which supports the hypothesis that early physiotherapy is beneficial. This is consistent with previous reports. Interestingly, the effect appeared strongest in younger patients, suggesting that age may modify the response, although older patients also seemed to improve somewhat. Adverse events were not a problem.
No numbers. “Considerably better” and “clear reduction” cannot be checked, pooled or replicated.
Interpretation inside results. “Supports the hypothesis” and “consistent with previous reports” belong in the discussion; placed here they foreclose the reader’s judgement.
A subgroup claim with no test. “Appeared strongest in younger patients” reads as selective reporting unless the interaction was pre-specified and tested.
Hedged safety. “Not a problem” is an interpretation of harms data, not a report of it.
Results reporting, discussion interpreting
Mean symptom score at 12 weeks was 24.1 in the intervention arm and 31.6 in the control arm (difference 7.5; 95% CI 4.2 to 10.8; P<0.001). The pre-specified age interaction was not significant (P=0.21); subgroup estimates are reported in Table 3 and are exploratory. Adverse events occurred in 14 of 203 intervention participants and 11 of 199 controls. The discussion then opens: early physiotherapy reduced symptom scores by a margin exceeding the 5-point minimum clinically important difference, consistent with earlier trials in this population.
The primary outcome is reported first, with effect size, interval and the test.
Interpretation moves to the discussion, where it can be debated.
The subgroup analysis is named as pre-specified, tested, and labelled exploratory.
Harms are given as counts and denominators, so a reader can judge them.
The second version is not better written; it is better organised. Each change anticipates an objection: the numbers make the claim verifiable, the exploratory label removes the suspicion of data dredging, and the counts allow a reader to assess harms independently. Every element also maps to a CONSORT 2010 item, so the checklist can be used as the drafting prompt instead of a submission formality.
Reporting guidelines as a structural overlay
Reporting guidelines are design-specific checklists of the items a reader needs in order to appraise a study 4,14. They do not prescribe a structure; they populate one. Used at the drafting stage, they indicate in advance which questions a reviewer is implicitly checking.
Schulz, Altman and Moher’s CONSORT 2010 statement in Trials covers parallel-group randomised trials. Page and colleagues’ PRISMA 2020 statement in the BMJ covers systematic reviews. Von Elm and colleagues’ STROBE statement, published in the BMJ and later in the International Journal of Surgery, covers observational designs: 22 items, of which 18 apply to all three designs it addresses — cohort, case-control and cross-sectional — and 4 are design-specific 12,18,19. Kuper, Reeves and Levinson’s BMJ introduction to appraising qualitative research sets out the equivalent expectations where the design is qualitative 16.
Several related tools accompany these and belong in specific sections. Sterne and colleagues’ RoB 2, in the BMJ, is the tool a reviewer will use to judge bias in a randomised trial, so the methods should supply what RoB 2 requires 7. Guyatt and colleagues’ GRADE papers in the BMJ and the Journal of Clinical Epidemiology govern how certainty of evidence is expressed in a review’s results and summary-of-findings table 13,17. Where outcomes are patient-reported, Prinsen and colleagues in Trials and Mokkink and colleagues’ COSMIN consensus in the Journal of Clinical Epidemiology set out what must be said about the measurement instrument, which belongs in methods and is routinely omitted 10,15.
The guideline should be selected before drafting, not before submission. The reporting guidelines page maps designs to checklists, and developmental editing is the service that addresses structure rather than sentences.
Administratively checked structural requirements
A separate set of requirements is structural in the sense that a submission can be delayed or returned if they are missing, independently of scientific merit. The ICMJE Recommendations are the common reference point, and three items account for most of the delays.
Authorship rests on four criteria, all of which must be met by each listed author: substantial contribution to conception or design or to data acquisition, analysis or interpretation; drafting or critical revision; approval of the final version; and agreement to be accountable for the work 1. Contributors who do not meet all four are acknowledged rather than listed — and this includes professional writing support, covered on the outsourcing medical writing page.
Trial registration must be in a registry meeting the minimum 24-item data set, and the registration identifier belongs in the methods, not only in a submission form 1. Since 1 July 2018, manuscripts reporting clinical trial results must include a data sharing statement 1. Both are inexpensive to satisfy prospectively and impossible to satisfy retrospectively, which is why they should be settled at the protocol stage. The 1997 Uniform Requirements in the Annals of Internal Medicine are the precursor of these rules and remain informative regarding how little the core has changed 21.
Allocating the word limit across sections
Most journals publish a total word count and leave its allocation to the author. A few publish the allocation, and these are instructive even for submissions elsewhere, because they encode an editor’s view of how much each section warrants. Radiology’s author instructions cap an Original Research article at 3,000 words and then divide it: a 15-word title, a 30-word summary statement, 75 words of key results, a 300-word abstract, a 400-word introduction, 800 words of materials and methods, 1,000 words of results, 800 words of discussion, and no more than 35 references 27.
This allocation should be read as a statement of priorities. Results receive the largest share. Methods and discussion receive the same allowance, so a discussion twice the length of the methods is a warning sign. The introduction receives 400 words, which is sufficient for context, gap and purpose and nothing else. Bannas and Reeder, in European Radiology, describe the same discipline for radiological manuscripts from the author’s side 9.
The budget should be constructed before drafting: take the journal’s total, allocate it in roughly these proportions, and write to the allocation. This is considerably easier than removing 900 words from a finished draft, and Kliewer’s observation that a manuscript usually requires around four drafts is more manageable when each draft is the appropriate length 19. For radiology specifically, the guide to radiology article types covers the submission limits by article type.
One journal’s published allocation, and what each limit forces
| Element | Limit | What the limit requires the author to decide |
|---|---|---|
| Title | 15 words | Design or population — rarely room for both |
| Summary statement | 30 words | The single claim the paper makes |
| Key results | 75 words | Which numbers are load-bearing |
| Abstract | 300 words | Design, participants, primary outcome, effect size |
| Introduction | 400 words | Context, gap, purpose — no literature review |
| Materials and methods | 800 words | What a reader needs to judge bias, not every procedure performed |
| Results | 1,000 words | Primary outcome first, in the order the methods promised |
| Discussion | 800 words | Finding, context, mechanism, limitations, next step |
| References | 35 maximum | Which prior work the argument rests on |
Radiology scientific style guide and author instructions (Radiological Society of North America), for Original Research articles. Limits are journal-specific; the proportions, not the absolute numbers, are what transfer. Paragraph-level guidance from Kliewer MA, American Journal of Roentgenology: around 18 paragraphs across five sections, three introduction paragraphs before the purpose statement and five discussion paragraphs before the summary.
Key points
Key points
IMRAD is a reading convention rather than a filing system: the introduction poses one answerable question, the methods allow a reader to judge whether the study could answer it, the results report what occurred, and the discussion considers what the findings mean and how far they can be trusted. It became standard only in the 1970s and 1980s (no article in Sollaci and Pereira’s 1935 sample used it, against over 80% by the 1970s), and it is therefore a set of expectations rather than a natural form. Most structural failures are errors of placement: interpretation within the results, new numbers in the discussion, or an introduction without a question. The reporting guideline for the design should be selected before drafting, a per-section word budget should be derived from the journal’s total, and registration, authorship and data sharing should be settled at the protocol stage rather than at submission.
Application to your study
Application to your study
- State the objective in a single sentence before drafting. Without such a sentence, the introduction has no end point and the design cannot be judged against it.
- Construct a word budget for each section. Allocate the journal’s total in approximately the proportions used by journals that publish their own allocation: results largest, methods and discussion equal, and introduction smallest.
- Select the reporting guideline before drafting. CONSORT applies to parallel-group trials, PRISMA 2020 to systematic reviews and STROBE to observational designs. Drafting should follow the checklist rather than auditing against it later.
- Identify interpretive sentences in the results and relocate them. Any sentence containing “suggests”, “importantly”, “consistent with” or “encouraging” belongs in the discussion.
- Confirm that every number in the discussion has already appeared in the results. Any number that has not is likely to be read as post hoc, whether or not it is.
- Report the primary outcome first and identify it as primary. If it appears after a significant secondary outcome, a reviewer is likely to notice the order.
- State the direction of each limitation. Indicate the direction in which each limitation would bias the estimate; a list without directions reads as a disclaimer.
- Resolve the ICMJE requirements at the protocol stage. These include the four authorship criteria for each contributor, registration in a registry meeting the 24-item minimum data set, and a data sharing statement for trial results.
Frequently asked questions
Introduction, Methods, Results and Discussion. Almost no journal states it as a requirement, because it is assumed: Sollaci and Pereira, in the Journal of the Medical Library Association, surveyed 1,297 original articles from the BMJ, JAMA, The Lancet and the New England Journal of Medicine and found the structure in no 1935 article but in over 80% of original papers by the 1970s, with each journal adopting it fully between 1975 and 1985. It should be treated as the default expectation for original research. Case reports, narrative reviews, editorials and letters follow different structures, which the journal usually specifies.
Authors should work from the journal’s total and allocate by priority. Radiology publishes its split for Original Research: 3,000 words overall, divided into a 300-word abstract, 400-word introduction, 800 words of materials and methods, 1,000 words of results and 800 words of discussion, plus a 15-word title, a 30-word summary statement, 75 words of key results and a maximum of 35 references. The absolute numbers are journal-specific but the proportions transfer well — results largest, methods and discussion equal, introduction smallest. Kliewer, in the American Journal of Roentgenology, describes a typical manuscript as around 18 paragraphs across five sections.
No, and it is the structural error reviewers flag most often. The results report what happened; the discussion states what it means. Sentences containing “suggests”, “importantly” or “consistent with previous work” are discussion sentences. The cost of combining them is not stylistic: interpretation placed within the results removes the reader’s ability to separate the data from the authors’ reading of them, and that is precisely the separation peer review exists to make. The converse error is equally damaging — a number appearing first in the discussion reads as an analysis conducted after the data were seen, whether or not it was.
It depends on the design. CONSORT 2010, from Schulz, Altman and Moher in Trials, covers parallel-group randomised trials. PRISMA 2020, from Page and colleagues in the BMJ, covers systematic reviews. STROBE, from von Elm and colleagues in the BMJ, covers observational designs and contains 22 items, 18 of them common to cohort, case-control and cross-sectional studies, with 4 design-specific. Kuper, Reeves and Levinson in the BMJ set out the equivalent expectations for qualitative work. Related tools sit alongside these: RoB 2 for bias in trials, GRADE for certainty of evidence in reviews, and COSMIN where the outcome is patient-reported.
Last, but the single claim should be drafted first. The abstract is the part most readers see and, in practice, where editorial triage occurs, so it must convey the design, population, primary outcome with a number, and a conclusion that does not exceed the result. Huth, in the Annals of Internal Medicine, argued for structured abstracts precisely because an unstructured abstract allows an omission to pass unnoticed. Journal guidance varies more than authors expect: Fujita and colleagues, in European Science Editing, found abstracts addressed in author guidance by 97% of Communication Science journals against 72% of Library and Information Science journals; where guidance is silent, the expectation should therefore be inferred from recent papers in the journal.
Authors should name the specific weaknesses and, for each, the direction in which it would bias the estimate. “Single centre, modest sample size” with no further comment is a disclaimer, not a limitation: a reviewer needs to know whether the bias would inflate or attenuate the effect, and whether the estimate could plausibly cross the threshold of importance. Two or three genuine limitations with directions are more informative than six generic ones. Anything an appraiser of the paper would want to know should be included — loss to follow-up, an unvalidated instrument, an outcome assessed unblinded — because otherwise a reviewer may identify it and treat the omission as the problem.
Three recur. Authorship rests on four criteria that each listed author must meet, and contributors who do not meet all four are acknowledged rather than listed — which includes professional writing support. Trial registration must be in a registry meeting the minimum 24-item data set, with the identifier in the methods and not only in a submission form. Since 1 July 2018, manuscripts reporting clinical trial results must include a data sharing statement. All three are straightforward if settled at the protocol stage and effectively impossible to remedy afterwards, which is why they should be addressed early.
Not on its own, and this should be stated clearly. Structure makes a study verifiable; it does not improve the study. Jawaid and Jawaid, in Pakistan Journal of Medical Sciences, report that between 70% and 80% of submissions to their journal are not processed further after initial screening and the editor’s triage, and list 17 distinct reasons — most of which concern the question, the design and the reporting rather than the prose. Good structure removes a category of avoidable objection and makes the genuine strengths of a study visible to a reader working quickly. This is valuable, but it cannot compensate for a design that is unable to answer the question.
References
- International Committee of Medical Journal Editors (ICMJE). Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. 2026. read 9 October 2026. icmje.org
- Bloothooft M, Pennings HJM, van der Heyden MAG. The Utrecht University Honours Program review project: example based scientific publishing training aimed at bachelor medical students. BMC Medical Education. 2024;24(1):1-12. doi:10.1186/s12909-024-05098-7 read 9 October 2026. doi.org/10.1186/s12909-024-05098-7
- Borowiec BG. Ten simple rules for scientists engaging in science communication. PLOS Computational Biology. 2023;19(7):e1011251. doi:10.1371/journal.pcbi.1011251 read 9 October 2026. doi.org/10.1371/journal.pcbi.1011251
- Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021:n71. doi:10.1136/bmj.n71 read 9 October 2026. doi.org/10.1136/bmj.n71
- Jawaid SA, Jawaid M. Common reasons for not accepting manuscripts for further processing after editor's triage and initial screening. Pakistan Journal of Medical Sciences. 2019;35(1):1-3. doi:10.12669/pjms.35.1.28 read 9 October 2026. doi.org/10.12669/pjms.35.1.28
- Pasarica M, Bailey M, Cendan JC. Increasing students' publication productivity: could launching a university scientific journal be a catalyst? Cureus. 2019;11(1):e3953. doi:10.7759/cureus.3953 read 9 October 2026. doi.org/10.7759/cureus.3953
- Sterne JAC, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898. read 9 October 2026. frontiersin.org
- Fujita MSL, Agustin-Lacruz MdC, Terra AL. Journals' guidelines about title, abstract and keywords: an overview of Information Science and Communication Science areas. European Science Editing. 2018;44(4):76-79. read 9 October 2026. europeanscienceediting.org.uk
- Bannas P, Reeder SB. How to write an original radiological research manuscript. European Radiology. 2017;27(11):4455-4460. doi:10.1007/s00330-017-4879-8 read 9 October 2026. doi.org/10.1007/s00330-017-4879-8
- Prinsen CA, Vohra S, Rose MR, et al. How to select outcome measurement instruments for outcomes included in a 'Core Outcome Set': a practical guideline. Trials. 2016;17(1):449. read 9 October 2026. link.springer.com
- Wager E, Woolley K, Adshead V, Cairns A, Fullam J, Gonzalez J, Grant T, Tortell S. Awareness and enforcement of guidelines for publishing industry-sponsored medical research among publication professionals: the Global Publication Survey. BMJ Open. 2014;4(4):e004780. doi:10.1136/bmjopen-2013-004780 read 9 October 2026. doi.org/10.1136/bmjopen-2013-004780
- von Elm E, Altman D, Egger M, Pocock S, Gotzsche P, Vandenbroucke J. The STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement: guidelines for reporting observational studies. International Journal of Surgery. 2014. doi:10.1016/j.ijsu.2014.07.013 read 9 October 2026. doi.org/10.1016/j.ijsu.2014.07.013
- Guyatt G, Oxman AD, Akl EA, et al. GRADE guidelines: 1. Introduction - GRADE evidence profiles and summary of findings tables. Journal of Clinical Epidemiology. 2011;64(4):383-394. read 9 October 2026. link.springer.com
- Schulz KF, Altman DG, Moher D. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. Trials. 2010;11(1). doi:10.1186/1745-6215-11-32 read 9 October 2026. doi.org/10.1186/1745-6215-11-32
- Mokkink LB, Terwee CB, Patrick DL, et al. The COSMIN study reached international consensus on taxonomy, terminology, and definitions of measurement properties for health-related patient-reported outcomes. Journal of Clinical Epidemiology. 2010;63(7):737-745. doi:10.1016/j.jclinepi.2010.02.006 read 9 October 2026. doi.org/10.1016/j.jclinepi.2010.02.006
- Kuper A, Reeves S, Levinson W. An introduction to reading and appraising qualitative research. BMJ. 2008;337:a288. read 9 October 2026. link.springer.com
- Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924-926. read 9 October 2026. link.springer.com
- von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, et al. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335(7624):806-808. doi:10.1136/bmj.39335.541782.AD read 9 October 2026. doi.org/10.1136/bmj.39335.541782.AD
- Kliewer MA. Writing it up: a step-by-step guide to publication for beginning investigators. American Journal of Roentgenology. 2005;185:591-596. read 9 October 2026. tech.snmjournals.org
- Sollaci LB, Pereira MG. The Introduction, Methods, Results, and Discussion (IMRAD) structure: a fifty-year survey. Journal of the Medical Library Association. 2004;92(3):364-371. read 9 October 2026. genreacrossborders.uwaterloo.ca
- International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. Annals of Internal Medicine. 1997;126(1):36-47. read 9 October 2026. stat.cmu.edu
- Vickery B. The Royal Society Scientific Conference of 1948. Journal of Documentation. 1992;54(3):281-283. read 9 October 2026. stat.cmu.edu
- Atkinson D. The evolution of medical research writing from 1735 to 1985: the case of the Edinburgh Medical Journal. Applied Linguistics. 1992;13(4):337-374. read 9 October 2026. stat.cmu.edu
- Huth EJ. Structured abstracts for papers reporting clinical trials. Annals of Internal Medicine. 1987;106(4):626-627. read 9 October 2026. stat.cmu.edu
- Meadows AJ. The scientific paper as an archaeological artifact. Journal of Information Science. 1985;11(1):27-30. read 9 October 2026. stat.cmu.edu
- Bazerman C. Modern evolution of the experimental report in physics: spectroscopic articles in Physical Review, 1893-1980. Social Studies of Science. 1984;14:163-196. read 9 October 2026. stat.cmu.edu
- Radiology (Radiological Society of North America). Radiology scientific style guide and author instructions: scientific editing. Radiology. read 9 October 2026. pubs.rsna.org
- Fernandez E, Garcia AM, Seres E, Bosch F. Students' satisfaction and perceived impact on knowledge, attitudes and skills after a 2-day course in scientific writing: a prospective longitudinal study in Spain. BMJ Open. 8(1):e018657. read 9 October 2026. bmjopen.bmj.com
Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.
How to cite this article
Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.
Pubrica Academy. (2026, October 9). The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines. Pubrica. https://pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/
Pubrica Academy. “The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines.” Pubrica, 9 Oct. 2026, pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/.
Pubrica Academy (2026) The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines. Pubrica. Available at: https://pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/ (Accessed: 15 October 2026).
Pubrica Academy. “The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines.” Pubrica. October 9, 2026. https://pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/
Pubrica Academy, ‘The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines’ (Pubrica, 9 October 2026) <https://pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/> accessed 15 October 2026.
Pubrica Academy. The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines. [Internet]. Pubrica; 2026 Oct 9 [cited 2026 Oct 15]. Available from: https://pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/
Pubrica Academy, “The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines,” Pubrica, Oct. 9, 2026. [Online]. Available: https://pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/
@misc{structure_of_a_medical_paper_writing_e_2026,
author = {{Pubrica Academy}},
title = {The Structure of a Medical Paper: Essential Elements of IMRAD and Reporting Guidelines},
howpublished = {Pubrica},
year = {2026},
month = {oct},
url = {https://pubrica.com/academy/medical-writing/structure-of-a-medical-paper-writing-essential-elements-and-reporting-guidelines/},
note = {Accessed: <span data-cite-acc="iso">2026-10-15</span>}
}This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.
Section-by-section structure plan you can copy
This plan should be completed before drafting, not afterwards. The word targets are derived by dividing the journal’s total in roughly the proportions used by journals that publish their own allocation. Any line that cannot be completed represents a question a reviewer will ask.
MANUSCRIPT STRUCTURE PLAN
Working title (<=15 words): ______________________________________
Target journal: ____________________ Total word limit: ______
Study design: ____________________ Reporting guideline: ____________
THE ONE-SENTENCE CLAIM
______________________________________________________________
(If this does not fit on one line, the paper has more than one paper in it.)
INTRODUCTION target: ____ words (~13% of total)
Para 1 - accepted position: _________________________________
Para 2 - the specific gap: __________________________________
Para 3 - purpose, stated as a question or objective: ________
[ ] No methods previewed [ ] No findings previewed
METHODS target: ____ words (~27% of total)
[ ] Design and setting
[ ] Eligibility / sampling / allocation, described so it can be judged
[ ] Each outcome, with the instrument used and its validation status
[ ] Primary outcome named as primary, and pre-specified where applicable
[ ] Analysis population stated (e.g. intention to treat)
[ ] Missing-data handling stated
[ ] Named test per outcome - never "appropriate tests"
[ ] Registration identifier present (registry meets 24-item data set)
[ ] Ethics approval and consent
RESULTS target: ____ words (~33% of total)
Order of reporting:
1. Participant flow / sample characteristics
2. PRIMARY outcome, with effect size, interval and test
3. Secondary outcomes, in the order the methods promised
4. Harms, as counts with denominators
5. Exploratory analyses, labelled exploratory
[ ] No interpretation anywhere in this section
[ ] No number here that is absent from a table, and none duplicated loosely
DISCUSSION target: ____ words (~27% of total)
Para 1 - principal finding, restated without new numbers
Para 2 - how it sits against existing work
Para 3 - plausible mechanism or explanation
Para 4 - limitations, EACH with the direction of likely bias:
limitation ____________ would bias the estimate ____________
limitation ____________ would bias the estimate ____________
Para 5 - what should happen next, and what this does not establish
[ ] Every number here already appeared in the results
ABSTRACT (write last) target: ____ words
[ ] Design [ ] Setting [ ] Participants [ ] Primary outcome with a NUMBER
[ ] Conclusion does not exceed the result
ADMINISTRATIVE
[ ] Each listed author meets all four ICMJE criteria
[ ] Non-qualifying contributors acknowledged, writing support included
[ ] Data sharing statement present (required for trial results since 1 July 2018)
[ ] Funding and conflicts declared
[ ] Reference count within the journal's limit Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.
How Pubrica can help
Pubrica’s developmental editors address structure before sentences: whether the introduction ends in an answerable question, whether the results report the primary outcome in the order specified in the methods, and whether any material appears for the first time in the discussion. Pubrica maps each manuscript against the reporting guideline for its design and returns the unmet items as a list of questions a reviewer is likely to raise.