skip to main content
study systems

Paper Triage in 7 Minutes: Decide If It’s Worth Your Time

A time-boxed way to triage papers: abstract-first reading, methods sniff-test, results extraction, and a “trust and relevance” score you can apply on shift.

Review date
2026-01-28
Reading time
9 min read
Author
Dr. Kola Tytler MBBS CertHE MBA MSt MRCGP
On this page

Key points

  • You are not reading to be impressed — you are reading to decide.
  • Triage is a filter: relevance first, then trust, then applicability.
  • Most papers can be rejected quickly for the right reasons (wrong population, wrong outcome, weak design, biased reporting).

The 7-minute triage clock

1

Minute 1 — Title + last sentence of the abstract

Identify what the authors claim. If the claim is irrelevant to your question, stop immediately.
2

Minutes 2–3 — Population, design, outcomes (from the abstract)

Extract: who was studied, what was done, and what was measured. If the outcome is a proxy you don’t care about, be cautious.
3

Minutes 4–5 — Methods sniff-test

Look for allocation, comparators, follow-up, missing data handling, and whether outcomes were pre-specified. You are detecting obvious threats to validity.
4

Minute 6 — Results extraction

Find the actual effect size (not just p-values). If you cannot find an effect size quickly, that is itself a signal.
5

Minute 7 — Trust × Relevance score

Give a quick 1–5 for trust (design + bias risk) and 1–5 for relevance (population + outcome match). Anything <3 in either category needs a reason to proceed.

The point of triage

Triage protects your attention. It prevents you from spending 45 minutes on something that was never going to change your decision.

Fast validity red flags

  • No clear comparator (or comparator is not clinically meaningful).
  • Outcome switching or unclear primary outcome.
  • Heavy reliance on surrogate outcomes without justification.
  • Large loss to follow-up without transparent handling.
  • Effect sizes missing or buried; reliance on “significant/non-significant” language.
  • Conflicts of interest not disclosed clearly.
SourceBMJ: How to read a paper (series hub)
Open Link
Your learning path

Save your place and choose what comes next

sign in to save your learning
Work through this guide (11 steps)

Use the actions from this article to track what you have worked through.

Sign in using the reading controls above to save your steps.

  • Identify what the authors claim. If the claim is irrelevant to your question, stop immediately.

    Return to this section

  • Extract: who was studied, what was done, and what was measured. If the outcome is a proxy you don’t care about, be cautious.

    Return to this section

  • Look for allocation, comparators, follow-up, missing data handling, and whether outcomes were pre-specified. You are detecting obvious threats to validity.

    Return to this section

  • Find the actual effect size (not just p-values). If you cannot find an effect size quickly, that is itself a signal.

    Return to this section

  • Give a quick 1–5 for trust (design + bias risk) and 1–5 for relevance (population + outcome match). Anything <3 in either category needs a reason to proceed.

    Return to this section
  • Return to this section
  • Return to this section
  • Return to this section
  • Return to this section
  • Return to this section
  • Return to this section

Short study sessions · Article 2 of 3