Lancet Writing

---
name: lancet-writing
description: Use to structure and hold the length of a Lancet Article main text — Introduction, Methods, Results, Discussion (~3000–3500 words), the Research in context panel, a limited reference list (~30), and a cautious, globally minded Discussion that does not overstate.
---

# Main-Text Writing & Structure (lancet-writing)

## When to trigger

- Drafting or trimming a Lancet original-research **Article**.
- The main text is over length, or the Discussion overstates the findings.
- The structure wanders (lab/clinic chronology instead of IMRaD).
- Co-authors keep adding paragraphs and references and the paper is creeping past budget.

## Article structure (IMRaD + the panel)

The Lancet original-research **Article** uses standard **IMRaD** with the Research in context panel:

| Section | Holds |
|---------|-------|
| **Introduction** | The problem, the gap, and the specific objective — brief (the systematic search lives in the panel, not a long literature review). |
| **Methods** | Design, setting/countries, participants, randomisation/allocation/blinding (or observational design), outcomes, sample size, statistical analysis (pre-specified), registration, ethics. Methods are detailed and **stay in the main text** (clinical journals expect full Methods, unlike Science's supplement model). |
| **Results** | Recruitment/flow (with the flow diagram), Table 1, primary outcome, secondary outcomes, subgroups, harms — in a logical order. |
| **Discussion** | Principal findings → comparison with other studies → strengths and limitations → implications. |
| **Research in context panel** | The mandatory three-part box (see `lancet-research-in-context`). |

## Length and reference budget

- **Main text ~3000–3500 words** (Articles; confirm the current cap — formats and limits change).
- **References limited (~30)** for an Article — cite the key and the systematic-search-anchored literature, not everything.
- Tables/figures within the journal's display-item allowance (see `lancet-figures-tables`); extended methods/results → appendix.
- An **appendix / supplementary material** carries the protocol summary, additional analyses, full subgroup tables, and the reporting checklist.

## The Discussion — four moves, cautiously

Write the Discussion in this order, and keep it cautious:

1. **Principal findings** — restate the main result plainly, without re-listing every number.
2. **Comparison with other studies** — situate against prior evidence (consistent with the panel's systematic search); explain agreement/disagreement.
3. **Strengths and limitations** — be candid; name confounding, generalisability, missing data, power.
4. **Implications** — for practice, policy, and research, with calibrated causal language.

> The Lancet prizes **caution, not overstatement**. Match claims to design: an observational study shows association, not causation; a single trial rarely "proves" — it "supports" or "provides evidence for." A non-inferiority trial cannot claim superiority.

## House style

- Accessible, globally minded prose: avoid US-only framing; define context for an international readership; spell out abbreviations on first use.
- British spelling is the house style; use the journal's conventions for numbers, units (SI), and dates.
- Lead Results paragraphs with the clinical finding, then the supporting statistics.

## Output format

```
【Sections present】 Introduction / Methods / Results / Discussion / Research-in-context panel — all? yes/no
【Methods in main text?】 yes (good) / pushed to appendix (FIX)
【Main-text word count】 N → over/under ~3000–3500 by M
【Reference count】 N → vs ~30 budget
【Discussion four moves】 principal findings / comparison / strengths-limitations / implications — all? yes/no
【Overstatement check】 causal language calibrated to design? yes/no
【Next】 lancet-ethics
```

## Anti-patterns

- **Do not** write a long literature-review Introduction — the systematic search belongs in the panel.
- **Do not** overstate: no causal claims from observational designs, no "proves," no superiority claim from a non-inferiority trial.
- **Do not** let the Discussion re-list Results numbers instead of interpreting them.
- **Do not** blow the reference budget by citing exhaustively; cite the load-bearing literature.