How to give a journal club presentation: a 12-slide template, appraisal by study design and a worked example

Disclosure: NOXI publishes this guide and is one of the tools it mentions; the method works in any presentation software.

This guide is for medical residents and students, nurses, pharmacists, allied-health trainees and life-science researchers who have to present someone else's paper and then lead the room. It gives a 12-slide template, timings for 30, 45 and 60-minute sessions, an appraisal map that pairs each study design with its current reporting guideline and its CASP checklist, the effect-size slide with every formula, five discussion questions that work in any room, and one fictional trial appraised from start to finish. The trial and every number in it were written for teaching; nothing in this guide is medical advice. If you are presenting your own research rather than someone else's, our guide to turning a research paper into a conference presentation is the better fit.

What a journal club presentation is, and what the room expects

A journal club presentation is a session in which one person presents a recently published study to colleagues, critically appraises its methods and results, and leads a discussion on whether the findings should change what the group does. The room expects two things from the presenter: an accurate summary and an independent judgement.

Most weak journal club talks have the same shape: twenty slides retelling the paper in the authors' order and words, then "any questions?" with four minutes left. The fix is structural. A summary tells the room what the authors did and found; an appraisal tells the room whether to believe it and whether it applies here. Both are needed, and they should be visibly separate in the deck so the audience knows when you stop speaking for the authors and start speaking for yourself.

Local conventions differ, so ask before you build anything. A University of Alabama at Birmingham guide for visiting students, based on a ten-step approach published in the British Journal of Hospital Medicine in 2013, notes that some programmes want a presentation under five minutes and some do not use PowerPoint at all, and advises asking the residents how journal club usually runs in their department (UAB, step-by-step approach to presenting at journal club). The template below is for the common case: a 30 to 60-minute session with slides.

The 12-slide journal club presentation template

A journal club presentation fits 12 slides: citation and why this paper, the question as PICO, background, study design, participants, intervention and comparison, outcomes, main results, effect size in plain numbers, appraisal, applicability, and discussion questions. Slides 1 to 9 are the summary, slides 10 and 11 are your judgement, and slide 12 hands the room over.

Three details make the template work. Slide 1 carries the reason for choosing the paper, so the room knows from the first minute why this study deserves 45 minutes. Slide 9 exists because almost every paper reports its main result in relative terms, and almost every practice decision needs absolute ones. And slide 11 puts your verdict on its own slide, in one sentence, so the discussion has something specific to agree or disagree with. For a 30-minute slot, merge background into slide 2 and outcomes into slide 6 (10 slides); for 60 minutes, add a second results slide for a key secondary outcome or subgroup and a harms slide (14 slides).

Write every slide title as a sentence that states what the slide shows, not as a label. "Readmission was 8.0% with oral vs 12.0% with IV" can be read and remembered from the back of the room; "Results" cannot. Read only your titles aloud in order: if they tell the story of the paper and your judgement, the deck is organised.

Choosing the paper for journal club

Choose one recent piece of primary research that addresses a question your group actually faces, reports its methods fully enough to appraise, and has a main result you can show in one figure or table. If the paper fails any of those three tests, the session will turn into a lecture or a guessing game.

• A real question. The best papers answer something the group argues about on the ward, at the pharmacy counter or in the lab meeting. "Why this paper" on slide 1 should be a sentence about your practice, not about the journal's impact factor.
• Primary research first. A single trial, cohort study or diagnostic accuracy study teaches appraisal best. Systematic reviews work well once the group is used to appraising the studies inside them.
• Appraisable methods. Check that the methods section, or its supplement, tells you how participants were selected, how groups were formed and how outcomes were measured. If you cannot answer the first three questions of the matching checklist from the paper, choose another.
• Presentable data. One primary outcome, one main figure. A paper with twelve co-primary outcomes is hard to summarise in a third of the session.
• Circulate it early. Stanford's Community College Outreach Program suggests sending the paper to your audience beforehand so they arrive familiar with the field and methods, and keeping background to no more than three slides (Stanford CCOP, How to give a journal club presentation, 2025). One week ahead is a common courtesy.

Read the paper three times: for the story, with the checklist open, and for the numbers on slides 8 and 9. Read the supplement too: missing-data handling and harms often live there.

How long is a journal club presentation? The thirds rule

A journal club presentation usually fills a 30 to 60-minute session, and the time splits into thirds: summarise the paper in the first third, appraise it in the second, and discuss it in the last. In a 45-minute slot that means 15 minutes of summary, 15 of appraisal including your verdict, and 15 of discussion.

The rule protects the part most presenters lose. Summaries expand to fill whatever time they are given, because every detail of the methods feels important when you have just read it three times. Write the two checkpoints on your notes (15:00 and 30:00 in a 45-minute slot) and, if you reach the effect-size slide late, cut summary detail, never the appraisal or the discussion. The summary is the part the room could have read for itself.

Slide count follows from the minutes, but less tightly than in a conference talk, because the discussion third uses one slide for 10 to 20 minutes. Our slides-per-minute guide covers the general rule; for journal club, plan roughly one slide per minute for the first two thirds only.

Put the question on one slide: PICO

The PICO slide states the study's question as four short lines: the population studied, the intervention given, the comparison and the outcome measured. Putting it second, straight after the citation, lets everyone in the room judge the rest of the paper against the question the authors actually asked.

Write PICO in the paper's own terms first, then add a fifth line in yours: "Our question: does this apply to [our patients]?" That line becomes the thread that runs to the applicability slide. For the fictional STEP-DOWN trial the slide reads:

• P: adults aged 18 to 80 admitted with community-acquired pneumonia and clinically stable by day 3.
• I: switch to oral antibiotics on day 3.
• C: continue intravenous antibiotics for at least 5 days.
• O: readmission for any cause within 30 days.
• Our question: would an early switch be safe for the patients on our unit, a third of whom are over 80?

If you cannot write the PICO from the abstract, the question is vaguer than the title suggests, and that belongs on the appraisal slide.

Explain the design in one diagram: the participant flow

The design slide should show how participants moved through the study as a single diagram, not as paragraphs: how many were screened, how many were included and allocated, how many were lost or stopped treatment, and how many were analysed in each group. For randomised trials, the paper's CONSORT flow diagram is that diagram, and you should redraw it rather than paste it.

CONSORT 2025, the reporting guideline for randomised trials, pairs a 30-item checklist with that flow diagram. It was published in April 2025, simultaneously in BMJ, JAMA, The Lancet, Nature Medicine and PLOS Medicine, and it supersedes CONSORT 2010. Its authors added seven new checklist items, revised three, deleted one, and created a new open science section covering trial registration, access to the protocol and statistical analysis plan, sharing of de-identified data, and funding and conflicts of interest (CONSORT 2025 statement, PLOS Medicine, 14 April 2025; CONSORT-SPIRIT, published statements). When you present a trial, the open science items are quick wins for the appraisal slide: was the trial registered before enrolment, is the protocol available, and who paid for it?

Redraw the diagram with only the boxes that matter for your argument. For STEP-DOWN that is four rows: 1,000 randomised; 500 to oral and 500 to IV; 46 and 21 without a 30-day outcome; 500 and 500 in the primary analysis. Highlight the row the room should notice, here the missing outcomes, because the rest of the appraisal hangs on it. Observational studies rarely publish a flow diagram, but drawing one from the methods section (source population, eligible, exposed and unexposed, followed up, analysed) is one of the clearest ways to spot selection problems.

Results: redraw one figure, don't paste five

The results slide should show the primary outcome in one redrawn figure or a small table, with the numbers written on it, the time frame in the title and the source in a footer. Pasting every figure from the paper hands the room a reading task during the part of the talk where they most need to listen.

• Primary outcome first. Show the outcome the trial was designed and powered for. Secondary outcomes and subgroups go on a backup slide unless the discussion depends on them.
• Counts and denominators. Write "40 of 500 (8.0%)" rather than "8.0%". The room needs the counts for slide 9, and denominators reveal analysis choices.
• Redraw, then cite. Rebuild the chart with large labels and only the bars you are discussing, and put "Redrawn from [authors, journal, year], Figure 2" at the foot. A reproduced figure needs its source on the slide.
• One idea per figure. Stanford's CCOP guide suggests one slide per figure or even per panel, noting for each its aim, method, result and contribution to the paper's question. For laboratory papers with multi-panel figures this is the right unit.

For chart choice, labelling and how to highlight the comparison that matters, our guide on how to present data covers the rules in detail. The one rule specific to journal club: show the result the way the authors defined it before you show it any other way, so the room can see what you changed and why.

Effect sizes in plain numbers: ARR, RRR and NNT on one slide

The effect-size slide converts the paper's result into five numbers the room can act on: the control event rate, the experimental event rate, the absolute risk reduction (the difference between them), the relative risk reduction (the difference as a share of the control rate) and the number needed to treat (one divided by the absolute risk reduction). Show each formula next to its value.

The definitions and conventions follow the Cochrane Handbook's chapter on interpreting results, which also recommends presenting absolute risks as natural frequencies, for example "120 per 1,000 to 80 per 1,000", rather than percentages alone, and reminds readers that relative effects tend to be more stable across risk groups than absolute ones (Cochrane Handbook, Chapter 15: Interpreting results and drawing conclusions). That last point matters at the applicability slide: the RRR of 33.3% may carry over to a sicker population, but the NNT will not, because it depends on the baseline risk.

Add the confidence interval

An effect without its precision is half a result. The second CASP question under results asks exactly this: was the precision of the estimate reported? If the paper gives a confidence interval for the risk difference, show it; if it gives only a p-value, you can compute an approximate 95% interval from the counts, and show the working on a backup slide:

• Standard error: √(0.12 × 0.88 ÷ 500 + 0.08 × 0.92 ÷ 500) = √(0.0002112 + 0.0001472) = √0.0003584 = 0.0189.
• Margin: 1.96 × 0.0189 = 0.0371.
• ARR interval: 0.040 − 0.0371 = 0.0029 to 0.040 + 0.0371 = 0.0771, that is 0.3 to 7.7 percentage points.
• NNT interval: 1 ÷ 0.0771 = 12.97, rounded up to 13; 1 ÷ 0.0029 = 344.8, rounded up to 345. So the NNT is 25 (95% CI 13 to 345).

This is a simple normal-approximation (Wald) interval; the paper's own method may differ slightly, and the paper's figure wins if it reports one. The point for the room is the width: the data are compatible with a large benefit (one readmission avoided for every 13 patients) and with a trivial one (one for every 345). The fictional paper's p-value of 0.035 says the difference is unlikely to be chance alone; the interval says how little we know about its size. Our guide to presenting uncertainty on slides shows how to draw ranges like this so they are read correctly.

Two cautions keep this slide honest. First, NNT always needs its time frame and outcome: "25 over 30 days to prevent one readmission". Second, if the confidence interval for the ARR crosses zero, the NNT interval runs from benefit to harm through infinity, and you should present the ARR interval instead of an NNT range.

Appraisal by study design: reporting guideline and CASP checklist

Appraise a paper with the checklist built for its design: identify the study design, check the paper against that design's reporting guideline (what it should report), then work through the matching CASP checklist (what you should ask). Using a randomised-trial checklist on a cohort study, or no checklist at all, is the most common reason a journal club appraisal sounds like opinion.

The reporting guidelines come from the EQUATOR Network library, which lists each guideline with its current version and history (PRISMA, STROBE, STARD, SRQR, COREQ, CHEERS, TRIPOD+AI). The checklists come from the Critical Appraisal Skills Programme, which offers eleven free downloadable checklists, including the two meta-analysis variants (CASP checklists). A reporting guideline is not a quality score: a paper can report every CONSORT item clearly and still describe a flawed trial. Use the guideline to find what is missing and the checklist to judge what is there.

The CASP RCT checklist, mapped to your slides

A CASP checklist is a free set of appraisal questions for one study design, each answered yes, no or can't tell. The checklist for randomised controlled trials (2024 version) groups 11 questions into four sections, and each section maps onto a part of the deck. Working through it with the slide numbers next to it is the quickest way to make sure the appraisal slide covers validity, results and relevance rather than only the weaknesses you happened to notice.

The checklist ends with an appraisal summary in three columns: methodologically sound, relatively poor methodology, and unknowns. That is a ready-made layout for slide 10. The third column is the one presenters skip and the one the room finds most useful, because "can't tell" items show where the paper's reporting failed. CASP's own instructions say a large number of can't-tell answers should make you interpret the findings with caution (CASP RCT Checklist, 2024).

Bias and limitations: five questions to ask of any study

Whatever the design, five questions expose most of the problems a journal club should discuss: who was studied and how they were chosen, whether the groups were comparable, whether outcomes were measured the same way for everyone, who was lost along the way, and who funded the work. Ask all five before you write the appraisal slide.

Write each limitation on slide 10 as three parts: the limit, what it means for the result, and how big a problem it is. "Missing outcomes: 9.2% with oral vs 4.2% with IV, counted as not readmitted; if 20 of the 46 missing oral patients were readmitted, the benefit disappears" is a limitation the room can discuss. "Some missing data" is not. Three to four limitations is enough; more than five starts to read as a dismissal of a paper the group chose to read.

Balance them with strengths: naming what a study did well makes your criticism more credible.

Applicability: from "statistically significant" to "would we change practice?"

The applicability slide answers one question in one sentence: would this result change what we do, for which patients, and what would we need first? A result can be statistically significant, clinically small and inapplicable to your setting all at once, and the verdict slide is where you say which of those it is.

Work through four checks and put the answers on the slide:

• Similar patients. Compare the trial's inclusion criteria with your own population. If a large share of your patients would have been excluded, the result does not yet cover them. CASP's question 10 asks this directly.
• Size that matters. Is the ARR, at its lower confidence limit, still worth the change? An NNT of 25 is attractive; an NNT of 345 might not justify retraining a ward.
• Harms and costs. Were harms reported for each group? What would the change cost in time, training or supplies, and what could you stop doing? CASP's question 11 asks about disinvestment as well as investment.
• Your baseline risk. If your patients are sicker or healthier than the trial's, recompute the NNT with your own control event rate and the paper's relative risk. With an RR of 0.67 (8.0 ÷ 12.0), a unit whose readmission rate is 18% would expect 18 × 8.0 ÷ 12.0 = 12.0%, an ARR of 6.0 points and an NNT of 1 ÷ 0.06 = 16.7, rounded up to 17.

Then write the verdict as one of four plain options: adopt, adopt for a defined group, wait for more evidence (say which), or do not adopt. "Promising for stable adults under 80; not yet for our oldest patients" is a verdict. "Interesting study, more research needed" is not, because every study needs more research.

Leading the discussion: five questions and how to handle silence

Lead the discussion with open questions that ask the room to make a judgement, not to recall facts from the paper. Five questions work in almost any journal club, and they follow the order of the appraisal: belief, size, applicability, change and next evidence.

Silence after a question is normal for about ten seconds; most presenters break it after three. Count to ten before you say anything. If the room is still quiet, narrow the question ("Hands up if you would switch a stable 75-year-old on day 3"), ask a named person a low-stakes version ("From the pharmacy side, what would worry you?"), or state your own view and invite disagreement ("I would wait for the sensitivity analysis. Who thinks I'm being too cautious?"). Close the discussion by restating where the room landed in one sentence and, if there is an action, who owns it.

Ask the consultant or attending in advance to speak last, so trainees answer first.

Nursing, pharmacy and lab-science journal club variants

The 12-slide structure holds across professions, but each group leans on a different slide: nursing journal clubs on the applicability slide and the change to practice, pharmacy journal clubs on the effect sizes, dosing and harms, and laboratory journal clubs on the methods and each figure's experiment.

Nursing and allied-health groups often appraise qualitative studies, and the slide order shifts accordingly: the "results" become themes with supporting quotes, the effect-size slide becomes a slide on how the themes were derived, and the appraisal turns on reflexivity and whether the sample suits the question. The CASP qualitative checklist and COREQ or SRQR cover those points. For lab groups, the PICO slide becomes a hypothesis slide: model system, manipulation, control and readout.

Worked example: appraising the fictional STEP-DOWN trial

The worked example applies every table in this guide to one fictional paper presented at a 45-minute internal medicine journal club. The trial, the journal, the hospital, the people and all the numbers are invented for teaching; the arithmetic is real so you can rerun it on your own paper.

The paper in one paragraph. STEP-DOWN randomised 1,000 adults aged 18 to 80 who were admitted with community-acquired pneumonia and clinically stable by day 3, 1:1 at 14 hospitals, using central web-based randomisation. The intervention group (500) switched to oral antibiotics on day 3; the control group (500) continued intravenous antibiotics for at least 5 days. The trial was open-label, the primary outcome was readmission for any cause within 30 days, taken from hospital records by assessors who did not know the allocation, and the trial was registered before enrolment. It was funded by the manufacturer of the oral antibiotic.

Time split. 15 minutes of summary (slides 1–9) + 15 minutes of appraisal and verdict (slides 10–11) + 15 minutes of discussion (slide 12) = 45 minutes. Omar's checkpoints are 15:00 at the start of slide 10 and 30:00 at the start of slide 12.

Effect size. Readmission was 60 of 500 with IV and 40 of 500 with oral. CER = 60 ÷ 500 = 12.0%; EER = 40 ÷ 500 = 8.0%. ARR = 12.0 − 8.0 = 4.0 percentage points (95% CI 0.3 to 7.7). RRR = 4.0 ÷ 12.0 = 33.3%. NNT = 1 ÷ 0.04 = 25 (95% CI 13 to 345). As natural frequencies: 120 per 1,000 readmitted with IV vs 80 per 1,000 with oral. The paper reports p = 0.035.

Missing outcomes. 46 of 500 oral patients (46 ÷ 500 = 9.2%) and 21 of 500 IV patients (21 ÷ 500 = 4.2%) had no 30-day outcome, a gap of 9.2 − 4.2 = 5.0 percentage points. The primary analysis counted them as not readmitted, which flatters the arm with more missing data. Omar runs three checks for the appraisal slide:

• Complete cases only: oral 40 ÷ 454 = 8.81%; IV 60 ÷ 479 = 12.53%. ARR = 12.53 − 8.81 = 3.72 points; NNT = 1 ÷ 0.0372 = 26.9, rounded up to 27. The direction holds.
• Tipping point: the benefit disappears when the oral rate reaches the IV rate of 12.0%, that is 60 events of 500. That needs 60 − 40 = 20 of the 46 missing oral patients to have been readmitted (20 ÷ 46 = 43.5%), if no missing IV patient was.
• Worst case: if all 46 missing oral patients were readmitted and no missing IV patient was, the oral rate becomes (40 + 46) ÷ 500 = 17.2%, which is 17.2 − 12.0 = 5.2 points worse than IV. The paper reports no sensitivity analysis, so this goes in the unknowns column.

Appraisal summary (slide 10). Sound: concealed central randomisation, an objective outcome with blinded assessors, registration before enrolment, and similar groups at baseline. Weaker: open-label treatment, missing outcomes twice as common in the oral arm and handled by assuming no event, and industry funding. Unknown: what happened to the 67 patients without outcomes, and harms, which appear only in a supplementary table.

Applicability (slide 11). The trial excluded patients over 80. In the last quarter, 412 of 1,329 admissions to Omar's unit were over 80 (fictional): 412 ÷ 1,329 = 31.0%, so almost a third of the unit's patients fall outside the evidence. Verdict: "Promising for stable adults under 80; not yet for our oldest patients. Ask the authors for the missing-data sensitivity analysis."

Discussion (slide 12). Omar uses questions 1, 3 and 4 from the table above, rewritten for the paper: "Does the missing data change your view?", "Would you switch a stable 82-year-old on day 3?" and "What would our unit need to change to switch early, and who would own it?" The room's conclusion goes on a closing line in the minutes: no change for patients over 80; the pharmacy lead will check whether the paper's supplement reports harms by age.

Here are two slides from this article's worked example, as NOXI built them, unedited.

Figures, sources and copyright on journal club slides

Every figure or table you reproduce from a paper needs its source on the slide, and you should check your institution's rules and the journal's licence before reusing figures beyond the room. Redrawing a figure from the published numbers, with a "Redrawn from" credit, is usually cleaner to read and simpler to reuse than a screenshot.

• Cite on the slide. Put "Source: [first author], [journal], [year], Figure [n]" in a small footer on every slide that shows the paper's data, not only on slide 1. The deck may be shared without your voice.
• Check the licence. Open-access papers under a Creative Commons licence usually allow reuse with attribution; subscription papers depend on the publisher and your institution's teaching exceptions. Internal teaching use and posting the deck online are different cases.
• Checklists have licences too. The CASP RCT checklist is published under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 licence, so you can show its questions in a non-commercial teaching session with attribution.
• Patient data. If you add local data to the applicability slide, such as your unit's age mix, use aggregates only, and follow your organisation's rules on what can be uploaded to any online tool. Our guide to confidential material and AI presentation tools covers what to check.

Using AI to build a journal club deck from the paper's PDF

AI can turn the paper's PDF into a first-draft summary deck in a few minutes, which frees your time for the part only you can do: the appraisal, the effect-size arithmetic and the verdict. Use AI for slides 1 to 8, write slides 9 to 12 yourself, and check every number against the paper.

Three habits make AI drafts safe for journal club. Keep the paper as the single source of facts, and give the tool the paper itself rather than asking it to recall the study, so it does not mix in another trial. Check every count, percentage and confidence interval against the paper's tables, because a summary that rounds 8.8% to 9% breaks your NNT. And write the appraisal in your own words: the room wants your judgement, and your programme may require it. Our guide to making a presentation from a PDF with AI explains what to check after conversion, and the guide to AI presentations with research citations covers how to keep sources visible on the slides.

The prompts below work in NOXI or any AI presentation tool. Replace the brackets.

How to build the journal club deck in NOXI

NOXI turns the paper's PDF into a structured deck that you then complete with your own appraisal and check against the paper. It is free, with no card and no watermark; uploading a file needs a free account, and a full deck uses about 25 of the 1,000 monthly credits.

Frequently asked questions

How do you present a journal club?

Present a journal club in three equal parts. First summarise the paper: citation, the question as PICO, design, participants, intervention, outcomes and the main result. Then appraise it with the checklist for its design, convert the result to absolute numbers, and state your verdict on its own slide. Finally lead a discussion with open questions about belief, size, applicability and what would change practice.

What is the format of a journal club presentation?

The common format is a 12-slide deck: citation and why this paper, PICO, background, design and flow, participants, intervention and comparison, outcomes, main result, effect size in plain numbers, appraisal, applicability and verdict, and discussion questions. Some programmes use a handout or no slides, and some ask for a five-minute summary, so check your department's convention first.

How many slides should a journal club presentation have?

About 12 slides suits a 45-minute session: nine for the summary, one for the appraisal, one for the verdict and one for discussion questions. Use about 10 for a 30-minute slot by merging background into the question slide, and about 14 for 60 minutes by adding a secondary outcome or harms slide. Put extra detail on backup slides after the last one.

How long is a journal club presentation?

Most journal club sessions run 30 to 60 minutes, though some programmes ask for a short five-minute summary. Split the session into thirds: in 45 minutes, 15 minutes of summary, 15 of appraisal including the verdict, and 15 of discussion. In 30 minutes, use 10, 10 and 10. If you run late, cut summary detail, never the discussion.

How do you critically appraise a paper for journal club?

Identify the study design, check the paper against that design's reporting guideline, such as CONSORT 2025 for randomised trials or STROBE for observational studies, and answer the matching CASP checklist. Ask whether the results are valid, what they are in absolute terms and with what precision, and whether they apply to your patients. Summarise as sound points, weaker points and unknowns.

What questions should you ask at journal club?

Five questions work in most rooms: Do you believe the result, and what is the biggest threat to it? Is the effect big enough to matter, given the confidence interval? Would our patients have been eligible? What would we have to change to act on this, and at what cost? What study would change your mind? Wait about ten seconds after each before prompting.

What is a CASP checklist?

A CASP checklist is a free set of structured appraisal questions from the UK Critical Appraisal Skills Programme, written for one study design. CASP offers eleven, including randomised trials, systematic reviews, cohort, case-control, cross-sectional, diagnostic, qualitative, economic evaluation and clinical prediction rule studies. The 2024 RCT checklist has 11 questions in four sections, answered yes, no or can't tell.

How do you explain NNT in a presentation?

Show the arithmetic and say it in people. If 12.0% of control patients and 8.0% of treated patients had the outcome, the absolute risk reduction is 4.0 percentage points, and the number needed to treat is 1 ÷ 0.04 = 25: treat 25 patients for the stated time to prevent one event. Always add the time frame, the outcome and the confidence interval.

Can I use AI to make a journal club presentation?

Yes, if your programme allows it, for the summary half. Give the AI tool the paper itself, ask for the citation, PICO, design, outcomes and main result slides, and check every number against the paper. Write the effect-size arithmetic, the appraisal and the verdict yourself, because the room is there to hear your judgement, and declare the AI use if your programme asks.

Sources & methodology

How this guide was made: the CONSORT 2025 details come from the statement in PLOS Medicine and the CONSORT-SPIRIT website; the other reporting guidelines and their versions from the EQUATOR Network's guideline pages; the checklist list, the RCT checklist's sections, questions and licence from CASP's own pages and its 2024 checklist; the NNT definition, rounding and presentation advice from the Cochrane Handbook; and local-convention and paper-choice advice from university journal club guides at UAB and Stanford. The 12-slide template, the thirds rule and timing table, the five bias questions, the five discussion questions and the profession table are editorial recommendations, not a standard. The STEP-DOWN trial, Kestrel Medical Review, Fairhaven Teaching Hospital, Dr Omar Saleh and every number in the example are fictional and written for teaching; the confidence intervals use a normal approximation and every calculation was checked by hand. Nothing in this guide is medical advice. All sources were checked on 16 October 2026.

• [1] CONSORT 2025: 30-item checklist plus flow diagram, supersedes CONSORT 2010, seven new items, three revised, one deleted, new open science section, published 14 April 2025 — CONSORT 2025 statement: updated guideline for reporting randomised trials. PLOS Medicine 22(4): e1004587 (2025).
• [2] Simultaneous publication of CONSORT 2025 in BMJ, JAMA, The Lancet, Nature Medicine and PLOS Medicine; flow diagram documents participant progress through the trial — CONSORT-SPIRIT, Published statements.
• [3] Eleven free downloadable CASP checklists by study design, including systematic reviews with meta-analysis of RCTs and of observational studies, cross-sectional and clinical prediction rule checklists — Critical Appraisal Skills Programme (CASP), CASP checklists.
• [4] CASP RCT checklist: 11 questions in sections A to D, yes / no / can't tell answers, appraisal summary columns, Creative Commons BY-NC-SA 4.0 licence — CASP, Checklist for Randomised Controlled Trials (2024).
• [5] PRISMA 2020 as the current version, updating PRISMA 2009 — EQUATOR Network, PRISMA 2020 statement.
• [6] STROBE (2007) for cohort, case-control and cross-sectional studies — EQUATOR Network, STROBE statement.
• [7] STARD 2015 for diagnostic accuracy studies — EQUATOR Network, STARD 2015.
• [8] SRQR (2014) for qualitative research — EQUATOR Network, Standards for Reporting Qualitative Research.
• [9] COREQ (2007), a 32-item checklist for interviews and focus groups — EQUATOR Network, COREQ.
• [10] CHEERS 2022 for health economic evaluations, updating CHEERS 2013 — EQUATOR Network, CHEERS 2022 statement.
• [11] TRIPOD+AI (2024) for clinical prediction models, replacing TRIPOD 2015 — EQUATOR Network, TRIPOD+AI statement.
• [12] NNT definition, 1 ÷ risk difference rounded up, NNTB and NNTH labels, natural frequencies, stability of relative vs absolute effects — Schünemann HJ, et al. Cochrane Handbook for Systematic Reviews of Interventions, Chapter 15 (version 6.5).
• [13] Ten-step approach to presenting at journal club (Bowles et al., British Journal of Hospital Medicine, 2013), local conventions including sub-five-minute formats and sessions without PowerPoint — University of Alabama at Birmingham, Step-by-step approach to presenting at journal club.
• [14] Sending the paper in advance, no more than three background slides, one slide per figure or panel — Stanford Community College Outreach Program, How to give a journal club presentation (2025).

Build your journal club deck in NOXI, free: upload the paper's PDF, review the outline, add your appraisal and export an editable PowerPoint or PDF for the meeting room.