Document-Type Deep Dives

Section 01· 2 min read

Choose the document before you choose the template

A template can make a document look familiar before the team has decided what the document needs to do. Route the work through five questions first:

1. What decision or understanding must the output support? Publication, study conduct, authority review, safety assessment, reimbursement, professional education, patient choice, or another use?
2. Who is the primary reader? Specialist, general clinician, investigator, regulator, payer, author team, learner, patient, caregiver, or public audience?
3. What is the evidence object? One study, a development program, a case, a body of literature, a model, post-market information, expert discussion, or a controlled claim set?
4. What level of method must be visible? Narrative synthesis, reproducible evidence review, formal study reporting, regulated dossier integration, instructional design, or plain-language transformation?
5. What is the destination? Journal, congress, authority, sponsor system, payer body, field team, website, participant, or internal decision forum?

The answers identify the family and subtype. A study-based manuscript and a narrative review may both be journal articles, but their evidence engines differ. A plain-language summary of a paper and a lay summary of clinical-trial results may share communication techniques while serving different source and process requirements. A clinical protocol manuscript describes a study to the scientific community; a regulated clinical protocol controls how the study will be conducted.

Routing table

Section 02· 4 min read

Dossier patterns by document family

Every project includes a brief, governing material, working decisions, and the deliverable. The evidence roles differ.

Original clinical-research manuscript

Core sources commonly include the final protocol and amendments, statistical analysis plan, validated tables/figures/listings, clinical study report where available, registry information, and relevant literature. Add target-journal instructions, authorship and disclosure records, prior abstracts or presentations, and author input.

Questions to resolve:

• Which analyses are primary, secondary, sensitivity, subgroup, or exploratory?
• Which output version is authoritative?
• Which study population governs each result?
• Are manuscript claims consistent with the registry, CSR, and prior disclosures?
• Which results have already been published or presented?

Observational or real-world study manuscript

Add the study protocol or analysis plan, data-source description, cohort definitions, codebook, variable and outcome definitions, confounding strategy, missing-data approach, analysis outputs, and data-quality information. Literature should cover current knowledge and design-specific context.

Questions to resolve:

• How were patients, exposure, comparators, and outcomes identified?
• What is known about completeness, measurement error, and missingness?
• Which confounders were measured and adjusted, and which may remain?
• Does the language remain associative where causal inference is not justified?

Narrative review

The literature set is the core evidence. Include the review concept, audience, time and topic boundary, target, search notes, key and contradictory papers, clinical-practice context, and any author expertise used to shape interpretation.

Questions to resolve:

• What organizing problem does the review solve?
• How was coverage made balanced?
• Which sources are seminal, current, direct, or secondary?
• Is the method described honestly without implying systematic completeness?

Systematic review and meta-analysis

Include protocol or prespecified review plan, exact search strategies, dates and sources searched, deduplication record, screening decisions, exclusion reasons, included-study data, appraisal, extraction tables, synthesis plan, meta-analysis outputs where applicable, and flow information.

Questions to resolve:

• Can the search and selection process be reproduced?
• Are eligibility decisions consistent?
• Is study quality considered before synthesis?
• Are heterogeneity and limitations represented?
• Does statistical pooling have suitable specialist ownership?

Case report

Include the clinical record approved for use, chronology, diagnostic and treatment information, outcome and follow-up, relevant literature, patient consent documentation where applicable, privacy review, and target instructions.

Questions to resolve:

• What is the teaching or hypothesis-generating value?
• Which details are necessary and which increase identifiability without adding value?
• Does the conclusion stay inside the evidence of one case?

Congress abstract, poster, and oral presentation

Include the parent study sources, current analysis cutoff, validated outputs, target-congress instructions, prior disclosure status, author information, and the approved evidence interpretation. Posters and talks add visual assets and presentation constraints.

Questions to resolve:

• Are results final or interim?
• Which details must survive compression?
• Does the visual preserve scale, denominator, and uncertainty?
• Is the derivative consistent with the later or parallel manuscript?

Clinical protocol

Include development strategy, background evidence, product information, prior nonclinical and clinical findings, study concept, objectives and endpoints, design decisions, operational input, statistical input, safety monitoring, ethics requirements, data-management plans, and current template/guidance.

Questions to resolve:

• Does every objective have an endpoint, assessment, and analysis path?
• Are eligibility, treatment, safety, withdrawal, and follow-up operationally clear?
• Do amendments propagate to consent, registration, SAP, and related files?

Clinical study report

Include final protocol and amendments, final SAP, validated outputs, data-review and deviation information, patient narratives, exposure and disposition data, prior reports, and applicable template and guidance. A draft report or slide is context, not automatically the numerical authority.

Questions to resolve:

• What differed between plan and conduct?
• Are all objectives and endpoints accounted for?
• Do synopsis, body, tables, narratives, and appendices agree?
• Are safety interpretation and patient privacy under appropriate review?

Investigator's brochure and integrated regulatory summaries

Include current product identity, nonclinical and clinical reports, safety information, pharmacology and pharmacokinetics, prior versions, development decisions, and the current structural requirements. Integrated summaries add program-level study tables, cross-study analyses, and evaluative strategy.

Questions to resolve:

• What is new since the prior version?
• Which risks and uncertainties are material to investigators or reviewers?
• Are conclusions integrated across sources rather than copied from individual reports?

Safety documents

Include case data or aggregate outputs, exposure, safety database information, prior reports, reference safety information, coding and listedness context, signal evaluation, risk-management information, and current reporting instructions.

Questions to resolve:

• Is the case or aggregate period defined?
• Are dates, seriousness, outcome, causality, and coding reconciled?
• Is personal information limited to what the document needs?
• Which recurring conclusions changed, and why?

Device and IVD evidence work

Include device identity and intended purpose, classification and lifecycle context, prior plans/reports, clinical or performance evidence, state-of-the-art literature, risk-management material, investigations, post-market information, complaints and vigilance where applicable, labeling, and current regulatory framework.

Questions to resolve:

• Is every claim tied to the correct device configuration and intended purpose?
• Is the literature search reproducible and current enough for a living evaluation?
• Which linked plan or report changes when new evidence arrives?

HEOR, value, and HTA work

Include the decision problem, population, interventions and comparators, clinical and economic evidence, epidemiology, resource use, utilities, costs, model files, assumptions, scenarios, sensitivity analyses, jurisdictional methods, and payer template.

Questions to resolve:

• Which perspective and time horizon govern the result?
• Can every narrative claim be traced to evidence or a model input?
• Are uncertainty and scenario dependence clear?

Patient and public material

Include the authoritative scientific or regulatory source, intended-reader profile, user questions, glossary or terminology decisions, privacy constraints, applicable requirements, translation plan, visual assets, patient or public input, and user-testing results.

Questions to resolve:

• What decision or understanding does the reader need?
• Which scientific qualifications must survive simplification?
• What did intended users misunderstand?
• Does translation preserve meaning and cultural relevance?

Section 03· 2 min read

Structure patterns by document family

Structure is not a universal list of headings. It is the ordered logic of the document. The patterns below are starting maps derived from the corpus; live targets may require different labels and order.

Research manuscript pattern

Title and authors
Abstract
Keywords / registration / required statements
Introduction: problem → current knowledge → gap → objective
Methods: design → population → intervention/exposure → outcomes → analysis → ethics
Results: flow/population → primary → secondary → other → safety
Discussion: synthesis → comparison → explanation → strengths/limitations → implication
Conclusion
Acknowledgments, contributions, funding, disclosures
References
Tables, figures, supplements

Review pattern

Title and abstract
Why the review is needed
Question and scope
Search and selection method at the rigor actually used
Thematic or evidence-based synthesis
Areas of agreement, disagreement, and uncertainty
Implications and gaps
Conclusion bounded by the evidence set
Disclosures and references

Systematic reviews add reproducible methods, study flow, included-study characteristics, appraisal, and synthesis results as formal research components.

Protocol pattern

Administrative identity and synopsis
Background and rationale
Objectives and estimands/endpoints as applicable
Design
Population and eligibility
Interventions and procedures
Assessments and schedule
Safety monitoring
Data handling and quality
Statistical approach
Ethics, consent, registration, and dissemination
References and appendices

CSR pattern

Title and synopsis
Ethics and administrative information
Investigators and study centers
Introduction and objectives
Investigational plan and analyses
Study population and conduct
Efficacy evaluation
Safety evaluation
Conclusions
Tables, figures, narratives, appendices, and references

Integrated summary pattern

Decision context and development strategy
Evidence inventory
Cross-study methods and comparability
Integrated efficacy / performance / value evidence
Integrated safety / risk evidence
Subgroups and special populations
Uncertainty, limitations, and missing information
Overall conclusions and implications
Supporting tables and source map

Patient-information pattern

What this information is and who it is for
Why the topic or study matters
What will happen / what happened
Benefits, risks, and uncertainty
Choices or actions for the reader
Privacy and support information where relevant
Where to learn more
Glossary or visual explanation as needed

Section 04· 1 min read

Publications and Medical Communications

Section 05· 1 min read

Primary clinical-trial manuscript

The job

Report the trial's prespecified question and findings for a journal audience with enough methodological transparency, numerical accuracy, and balanced interpretation for editorial and scientific evaluation.

Production sequence

1. Confirm target, authorship, prior disclosures, and study identity.
2. Reconcile protocol, amendments, SAP, registry, CSR, and validated outputs.
3. Map objectives and endpoints to Methods and Results.
4. Build the Introduction around the clinical problem and precise gap.
5. Plan tables and figures before narrative results.
6. Report primary, secondary, and safety findings in a prespecified order.
7. Build Discussion from findings, literature context, alternatives, strengths, limitations, and implications.
8. Check abstract and conclusion against the final body.
9. Complete contributions, acknowledgments, funding, disclosure, registration, data, and other required statements.

Dominant risks

• Manuscript becomes a compressed CSR rather than a reader-oriented paper.
• Primary and exploratory findings are mixed.
• Efficacy receives more interpretive space than safety.
• Results repeat displays without explaining the pattern.
• Conclusions extend beyond the population, endpoint, duration, or precision.
• Registry, abstract, manuscript, and source report disagree.

Section 06· 1 min read

Secondary and post-hoc manuscript

The job

Communicate a question derived from an existing study while preserving its relationship to the original design, primary analysis, and prior publications.

Extra controls

• State whether the question and analysis were prespecified.
• Identify multiplicity and subgroup limitations.
• Explain relation to primary objectives and publications.
• Prevent retrospective rationale from sounding prospective.
• Avoid fragmenting one evidence story into redundant papers without distinct value.

The conclusion should make the analysis useful without converting it into confirmation. “These findings support further study” may be more faithful than a practice recommendation.

Section 07· 1 min read

Observational and real-world-evidence manuscript

The job

Report evidence from routine care, registries, databases, cohorts, case-control designs, or other nonrandomized sources with explicit provenance, cohort construction, analytic methods, and design limitations.

Extra controls

• Define data source, coverage, quality, and linkage.
• Describe operational definitions and coding.
• Explain selection, confounding, adjustment, missingness, and sensitivity analyses.
• Report estimates and uncertainty.
• Keep association and causation distinct.
• Discuss directness and generalizability to the intended population.

The observational study can answer questions a trial may not, but its usefulness depends on transparent methods rather than causal-sounding prose.

Section 08· 1 min read

Reviews, evidence synthesis, and meta-analysis

Narrative review

Create a concept before collecting prose. The review should solve a reader problem: explain a changing field, compare approaches, clarify controversy, or identify a practical evidence gap. Use a transparent search description appropriate to the process, represent counterevidence, and organize around themes or decisions rather than paper-by-paper summaries.

Systematic review

Treat the protocol, search, screening, extraction, appraisal, and synthesis as the study. Preserve exact search strategies, decision reasons, and flow. Report included evidence and limitations. The writer should not retrofit eligibility to produce a preferred result.

Meta-analysis

Work with statistical expertise. Confirm compatibility of populations, interventions, outcomes, designs, and effect measures. Explain heterogeneity and sensitivity. A pooled estimate is not automatically more truthful than the studies it combines.

Dominant risks

• Convenience selection described as systematic.
• Search methods omitted because they occurred “before writing.”
• Quality appraisal mentioned but not used in interpretation.
• Conclusions broader than eligibility.
• Statistical pooling used to hide conceptual heterogeneity.

Section 09· 1 min read

Case report and case series

The case report's strength is specificity. Establish why the observation matters, provide a coherent chronology, explain diagnostic reasoning and intervention, report outcome and follow-up, discuss relevant literature, and state the lesson modestly.

Protect privacy by removing unnecessary detail and following applicable consent and target requirements. Do not generalize efficacy or safety from one case. A case can alert, illustrate, or generate a hypothesis; it rarely settles one.

Section 10· 1 min read

Editorial, commentary, and letter

Choose one primary point. State the position early, support it with proportionate evidence, acknowledge the strongest alternative, and make the requested action or interpretation clear. A short format magnifies unsupported assertion.

For response letters to editors and reviewers, answer every item, show the change, and explain disagreements. Keep tone factual and respectful. The purpose is not to win against the reviewer but to make the manuscript and reasoning inspectable.

Section 11· 1 min read

Congress abstract

Follow the target headings and limits. Include sufficient methods to interpret the result. Report the main estimate and uncertainty, define the analysis population, and state data status. Do not cut limitations and safety merely because space is scarce.

Before submission, compare the abstract with protocol, registry, outputs, and any prior disclosure. After later analyses, preserve a record of changes so the full publication can explain material differences.

Section 12· 1 min read

Poster and oral presentation

Build around one communication objective. Use visual hierarchy to answer: why, how, what happened, what it means. Keep type legible at expected viewing distance. Use figures and tables that reveal patterns; remove decorative complexity.

For talks, design slides to support speech. One slide should make one main point. Put the conclusion where the audience can see the evidence behind it. Rehearse timing and questions. A presentation is a live reader journey, not a manuscript broken into rectangles.

Section 13· 1 min read

Plain-language summary of a publication

Translate the research question, method, main findings, safety context, limitations, and meaning for a non-specialist reader. Use familiar terms, explain necessary concepts, and present absolute numbers with denominators where possible. Avoid describing the intervention as proven, safe, or effective beyond the study.

Involve intended readers or patient experts and perform user testing when the workflow allows. Record what changed because of their input. Readability scoring can support review but cannot replace comprehension testing.

Section 14· 1 min read

Medical-information response

Define the question precisely, search and appraise current relevant evidence, synthesize it fairly, and remain within the approved and applicable context. State limitations and include safety where material. Use controlled references and review status because responses may be reused.

Version the response. Set a review trigger based on new evidence, labeling, policy, or recurring use. Archive superseded content so it cannot be mistaken for current approved material.

Section 15· 1 min read

Scientific platform and core narrative

Organize the disease, mechanism, evidence, terminology, key messages, objections, safety, and references into a controlled source for derivative communication. Give every claim an owner, support, wording boundary, approval state, and review date.

Separate:

• Evidence statement
• Approved interpretation
• Permitted audience/context
• Required qualifier
• Supporting references or internal sources
• Status and version

The platform is valuable only when it reduces drift. It should not become an excuse to copy one sentence into every audience context.

Section 16· 1 min read

Advisory-board materials and reports

Before the meeting

Define the business or scientific objective, decisions sought, participant role, questions, pre-reading, evidence background, and facilitation plan. Avoid briefing material that pushes participants toward a predetermined answer while presenting the event as independent advice.

During the meeting

Capture statements, areas of agreement, disagreement, rationale, evidence gaps, and actions. Distinguish direct participant input from the writer's synthesis. Clarify attribution and confidentiality rules.

After the meeting

Produce minutes, executive summary, or full report according to purpose. Do not manufacture consensus from a majority theme. Preserve important minority views. Route factual corrections separately from requests to change the meaning of what was said.

Section 17· 1 min read

CME, continuing education, and training

Start with the learner and practice gap. Define desired outcomes and measurable learning objectives, then choose an instructional format capable of producing them. Develop evidence-balanced content and activities that require application, not only recognition.

Connect evaluation to the objective. A satisfaction score cannot prove practice change. Maintain independence, disclosure, and promotional boundaries according to the governing context. For field training, confirm approved claims and give learners practice using evidence accurately in realistic situations.

Section 18· 1 min read

Regulatory, Safety, Devices, Market Access, and Public-Facing Work

Section 19· 1 min read

Clinical protocol

The job

Create the operational and scientific plan for a study. The protocol must align rationale, objectives, endpoints, design, population, intervention, assessments, safety, data, statistics, ethics, and dissemination.

Writer's control points

• Build an objective-endpoint-assessment-analysis matrix.
• Use consistent definitions across synopsis, body, schedule, and appendices.
• Make eligibility and withdrawal criteria executable.
• Record cross-functional decisions and unresolved feasibility issues.
• Assess amendment impact across related study documents.

Dominant risks

Ambiguous procedures, inconsistent timing, objectives without analyses, template boilerplate, and late changes that do not propagate.

Section 20· 1 min read

Clinical study report

The job

Provide a complete, traceable account of study design, conduct, analysis, results, and interpretation, including differences from the plan.

Writer's control points

• Build a protocol/SAP/CSR/output mapping before results drafting.
• Track deviations from plan and explain their treatment.
• Reconcile population counts and data cutoffs.
• Draft synopsis after the body is stable, then verify it independently.
• Plan patient narratives and privacy review early.
• Use focused QC passes for data, consistency, and cross-references.

Dominant risks

Thousands of source pages create false confidence that “the answer is somewhere.” The writer must route exact sources and control repeated information. Synopsis, efficacy, safety, narratives, and appendices can drift unless the source hierarchy is explicit.

Section 21· 1 min read

Investigator's brochure

The job

Integrate current nonclinical and clinical knowledge relevant to investigators and safe study conduct.

Writer's control points

• Compare the prior and current evidence base.
• Maintain product and formulation identity.
• Integrate rather than list study findings.
• Make risks, uncertainties, and safety guidance visible.
• Record why sections changed or remained stable.

An IB is a living synthesis. Reuse requires change control and source traceability, not annual copying.

Section 22· 1 min read

CTD overviews and summaries

The job

Explain the development program and integrate quality, nonclinical, clinical, efficacy, and safety evidence for authority evaluation.

Writer's control points

• Start from the decision and development logic.
• Build cross-study comparison tables.
• Define methods for integrated analyses.
• Explain differences among studies, populations, formulations, doses, and durations.
• Keep overview judgments supported by detailed summaries and reports.
• Use expert review for integrated interpretation.

The summary is not successful merely because it is shorter than its sources. It must show the pattern across them.

Section 23· 1 min read

Pharmacovigilance and safety writing

Case narrative

Construct a chronological, clinically coherent account from verified case data. Include necessary demographics, exposure, event, relevant history, tests, treatment, outcome, and assessment according to the applicable context. Reconcile discrepancies rather than smoothing them into a story. Minimize unnecessary identifying detail.

Aggregate and periodic report

Define period, product, exposure, data sources, signals, risks, and changes since the prior report. Integrate rather than list. Track recurring text, changed evidence, and downstream risk-management implications.

Dominant risks

Privacy loss, coding mismatch, inconsistent dates, incomplete follow-up, unsupported causality, missed change from prior cycle, and weak reconciliation across safety systems and narrative.

Section 24· 1 min read

Public disclosure and transparency

Create a cross-record matrix for protocol, registry, result posting, CSR, publication, and public summary. Compare study identity, objectives, endpoints, populations, dates, and results. Explain legitimate differences and correct errors.

For public document release, apply the current disclosure and privacy process. Preserve scientific usefulness while protecting personal and legitimately confidential information. Redaction, anonymization, and document preparation need specialist review; they are not final-format editing.

Section 25· 1 min read

Medical-device and IVD evidence suites

The job

Maintain linked evidence about intended purpose, safety, performance, clinical benefit, state of the art, risk, and post-market experience across the lifecycle.

Writer's control points

• Lock device identity, variant, intended purpose, population, and claims.
• Maintain a reproducible literature method and search update history.
• Separate evidence types and appraisal.
• Connect conclusions to risk management and post-market plans.
• Assess change impact across every linked plan, report, and summary.
• Keep public summaries understandable without weakening risk information.

Dominant risks

Evidence for one configuration or intended use migrates to another; literature searches become irreproducible; old conclusions survive new post-market evidence; and linked documents contradict one another.

Section 26· 1 min read

HEOR and market-access dossiers

The job

Present clinical, economic, humanistic, and real-world evidence for a defined payer or assessment decision.

Writer's control points

• Define perspective, population, comparator, time horizon, and decision context.
• Map every model input and assumption to a source or approved rationale.
• Explain scenario and sensitivity analyses.
• Keep clinical evidence, modeled outputs, and value interpretation distinct.
• Tailor to the current jurisdictional method and template.

Dominant risks

Model precision hides assumption uncertainty; evidence from another jurisdiction is treated as directly transferable; narrative claims lose their connection to inputs; and favorable scenarios receive disproportionate emphasis.

Section 27· 1 min read

Informed consent and participant information

The job

Support an informed, voluntary decision by explaining the study or procedure, what participation involves, reasonably relevant risks and benefits, alternatives, privacy, choices, contacts, and rights in the required context.

Writer's control points

• Organize around participant questions.
• Use plain language without removing necessary information.
• Keep consent consistent with protocol and current study procedures.
• Involve patients or intended users.
• Test navigation, comprehension, and actionability.
• Control translation and versioning across amendments.

Dominant risks

Legal or scientific completeness becomes unreadable; simplification removes decision-critical detail; copied boilerplate does not fit the study; and translated versions drift from the approved source.

Section 28· 1 min read

Lay summary of clinical-trial results

Explain the study purpose, who participated, how it was done, main findings, safety or adverse reactions, limitations, and where more information is available. Clarify that group results do not determine an individual's treatment. Use balanced visuals and absolute numbers.

Phase I and early-development studies may require extra explanation because endpoints, healthy-volunteer context, and future product use may be unfamiliar or uncertain. State what was and was not learned.

Section 29· 1 min read

Patient education and public communication

Start from the audience's real questions and literacy context. Explain what the condition or intervention is, why the information matters, what actions are available, and when professional help is needed. Distinguish education from individual medical advice.

Build trust through accurate sourcing, transparent uncertainty, accessible design, and respectful tone. Counter misinformation by addressing the underlying concern and evidence, not by dismissing the reader. Use multiple formats when the audience benefits, but verify that adaptation has not changed meaning.

Section 30· 1 min read

Translation and localization

Provide translators with source context, purpose, audience, terminology, visuals, and decision-critical language. Use qualified medical translation and independent review. Back translation can identify differences but does not replace target-language comprehension and cultural review.

Check numbers, units, product names, risk terms, negation, dates, and action statements. Adapt examples and visuals where cultural meaning changes. Retest the final layout because text expansion and script direction can alter accessibility.

Section 31· 1 min read

One-page production card

Use this card for any document type:

Section 32· 1 min read

Sources

This deep-dive shelf is synthesized only from:

• Strategic Scientific and Medical Writing
• Writing High-Quality Medical Publications
• Medical Writing: A Guide for Clinicians, Educators, and Researchers
• The Complete Guide to Medical Writing
• EMWA Career Guide
• Ultimate Guide to Becoming a Medical Writer
• AMWA journal corpus
• EMWA journal corpus

Return to the Medical Writers' Handbook