jama-writing-style
GitHub用于精简JAMA稿件语言，去除夸大表述（spin），匹配临床风格与AMA规范。优化措辞但不改动分析或数据。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill jama-writing-style -g -y
SKILL.md
Frontmatter
{
"name": "jama-writing-style",
"description": "Use when tightening prose and removing spin from a JAMA manuscript to match the journal's clinical house style and the AMA Manual of Style. Polishes language and framing; it does NOT change the analysis or the reported numbers."
}
Writing Style & House Style (jama-writing-style)
When to trigger
- Prose is verbose, hedged, or jargon-heavy for a general clinician audience
- The Conclusions overstate the data ("spin")
- Observational findings are written with causal verbs
- Late-stage polish before submission, after design/stats/exhibits are settled
House-style essentials (AMA Manual of Style)
- JAMA follows the AMA Manual of Style. Conform numbers, units, abbreviations, statistical notation, and reference format to it — verify current rules in the manual / author instructions.
- SI units with conventional units as needed; define every abbreviation at first use; avoid undefined acronyms.
- Person-first, non-stigmatizing language (e.g., "patients with diabetes," not "diabetics"); report and justify race/ethnicity and sex/gender per current JAMA guidance.
- Exact p-values (P = .04, not P < .05); leading zeros omitted for p-values per AMA convention; estimates always paired with 95% CIs.
- Active voice and short sentences where possible; the IMRaD structure (Introduction, Methods, Results, Discussion).
Avoiding "spin"
Spin is language that makes results sound better than the data support. JAMA reviewers and editors actively police it.
| Spin pattern | Fix |
|---|---|
| Null primary outcome, upbeat conclusion | State the primary result plainly; do not pivot to secondaries |
| Secondary/post hoc outcome framed as main finding | Foreground the pre-specified primary outcome |
| "Trend toward significance" for a non-significant result | Report the estimate + CI; avoid the phrase |
| Causal verbs ("reduced," "caused") for observational data | Use associational verbs ("was associated with") |
| "Safe and well tolerated" beyond what harms data show | Report harms quantitatively; calibrate the claim |
| Over-generalizing beyond the studied population | Bound the conclusion to the sample/setting |
Discussion structure JAMA expects
- Brief statement of the principal findings (tied to the primary outcome).
- Comparison with prior evidence; how this study adds.
- Possible mechanisms / explanations (without overreach).
- Limitations — candid, specific (bias, generalizability, power, missing data).
- Conclusions and clinical relevance, proportionate to the evidence.
How JAMA editors read for spin and clarity
At the Journal of the American Medical Association, manuscript editors and the AMA Manual of Style copy desk hold prose to a general-clinician readability standard and treat unsupported optimism as a fixable defect. The bar: a non-specialist clinician reads the Conclusions, knows what changes at the bedside, and trusts the wording does not exceed the data. Recurring fixes: Conclusions warmer than the Results get rewritten to the primary outcome; stigmatizing labels become person-first; a leading-zero "p=0.043" becomes exact P = .04 with the 95% CI; causal verbs on a cohort become "was associated with"; and generic limitations name the specific bias, population bound, and missing data.
Worked micro-example: de-spinning a conclusion (illustrative)
Vignette (illustrative): a prospective cohort, N = 12,500 adults, finds a dietary pattern associated with lower incident heart failure; adjusted hazard ratio 0.88 (95% CI, 0.79-0.98).
- Spun draft: "Adopting this diet reduces heart failure and should be recommended for all adults."
- JAMA-calibrated rewrite: "Greater adherence was associated with lower incident heart failure (adjusted HR, 0.88; 95% CI, 0.79-0.98); whether the association is causal requires randomized evaluation."
The rewrite swaps the causal verb for associational language, keeps the estimate with its CI, and bounds the recommendation.
Reviewer pushback and the fix
- "Conclusions overstate the data (spin)." Fix: tie the headline to the primary outcome; if null, say so and do not pivot to a secondary.
- "Clinical bottom line not actionable." Fix: end the Discussion with one restrained sentence on what a clinician should now do.
Calibration anchors (hedge where uncertain): person-first language, exact-P / no-leading-zero notation, and estimate-plus-CI pairing are durable AMA conventions; reference-formatting and race/ethnicity rules evolve — confirm against the latest AMA Manual of Style.
Checklist
- Abstract, Results, and Conclusions agree numerically and in tone
- Conclusion is bounded by the primary outcome; no spin
- Observational findings use associational language
- Harms reported quantitatively before any "safe/tolerated" claim
- Abbreviations defined at first use; SI units; AMA statistical notation
- Person-first, non-stigmatizing language; race/sex reported per guidance
- A specific, candid Limitations paragraph is present
- Reference format follows AMA style (verify current rules)
Anti-patterns
- Spin: a positive-sounding conclusion over a null primary outcome
- Causal language for an observational association
- "Trend toward significance" / "marginally significant"
- Burying limitations or making them generic ("as with all studies…")
- Jargon and undefined acronyms that exclude a general clinician reader
- Tone in the abstract that the body does not support
Output format
【Spin found】none / instances: ...
【Causal-vs-associational language correct】yes / fixes: ...
【Harms calibrated】yes / no
【Limitations specific】yes / no
【AMA style conformance】numbers/units/abbrev/refs: ...
【Next skill】jama-cover-letter
Version History
- 1839142 Current 2026-07-05 13:24
