TRIPOD+AI at a glance
- Full title
- TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods
- Citation
- Collins GS, Moons KGM, Dhiman P, et al. BMJ 2024;385:e078378. doi:10.1136/bmj-2023-078378. First published 16 April 2024; open access (CC BY 4.0).
- What it covers
- Studies that develop a diagnostic or prognostic prediction model, evaluate (validate) its performance, or both, using regression or any machine learning method. The authors note that it is aimed primarily at non-generative models.
- What it contains
- A 27-item checklist (52 sub-items in total) and a separate 13-item TRIPOD+AI for Abstracts checklist.
- Status
- Supersedes the TRIPOD 2015 checklist, which the authors say should no longer be used.
- TRIPOD-AI or TRIPOD+AI?
- The 2021 development protocol used the working name TRIPOD-AI. The published guideline is TRIPOD+AI, where the "+" signals one set of recommendations for both regression and machine learning models.
- Official sources
- The BMJ paper · TRIPOD+AI checklist (PDF) · TRIPOD+AI for Abstracts (PDF) · tripod-statement.org
If you have spent time submitting clinical prediction model papers in the past year or two, you have probably encountered the TRIPOD+AI checklist, either because a journal required it at submission or because a reviewer flagged missing items in a critique. The checklist is now the baseline reporting standard for any study that develops or validates a prediction model, whether the underlying method is logistic regression, a random forest, or a deep neural network trained on electronic health records. It replaced the TRIPOD 2015 checklist, which focused mainly on regression models.
The checklist is long. Its 27 items span the title, abstract, introduction, methods, open science practices, patient and public involvement, results, and discussion, and many split into sub-items. Several of the requirements that are new compared with TRIPOD 2015, such as describing how model fairness was addressed (item 14), saying where the protocol and registration can be found (items 18c and 18d), and stating data and code availability (items 18e and 18f), are far easier to report well when they were planned at the design stage. A protocol or a registration cannot be created after the analysis is done.
This guide maps all 27 items, explains what changed from TRIPOD 2015, covers where papers tend to fall short, and shows how to integrate the checklist into your workflow before the manuscript exists rather than after it is already drafted.
Working Principle
Treat TRIPOD+AI the way you treat your ethics approval: a decision that shapes the study design, not an administrative box to tick after data collection is complete. The checklist items that authors most often omit, including study registration, sample size justification, and fairness analysis, cannot be retrofitted after the analysis is done.
From TRIPOD 2015 to TRIPOD+AI: Why the Update Was Needed
TRIPOD, which stands for Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis, has been the standard reporting framework for prediction model studies since its original publication in 2015. The original 22-item checklist covered the essential elements of study design, outcome definition, predictor selection, and model performance reporting. For the logistic regression and Cox proportional hazard models that made up most clinical prediction research at the time, it was a workable standard.
The problem was that by the early 2020s, machine learning methods had entered clinical research at scale. Gradient-boosted trees, support vector machines, neural networks trained on imaging data, and natural language processing models applied to clinical notes all pose reporting challenges that a regression-focused checklist cannot handle. How should a paper describe feature importance in a method with hundreds of predictors? How do you document preprocessing steps for a model trained on chest radiographs? What should readers know about calibration when the model outputs probabilities from a deep learning classifier? TRIPOD 2015 offered no guidance on any of these questions.
The TRIPOD+AI statement, published in the BMJ on 16 April 2024 by Gary Collins, Karel Moons, Paula Dhiman, Richard Riley, Andrew Beam, Ben Van Calster, and collaborators, addressed this gap directly. It is explicitly designed for both regression and machine learning prediction model studies, applies to development, validation, and combined development-and-validation designs, and the new checklist supersedes the 2015 version entirely. If you are writing a prediction model paper today, TRIPOD 2015 is no longer the appropriate reference: the authors state that the 2015 checklist should no longer be used. The EQUATOR Network, which maintains the most comprehensive directory of health research reporting guidelines, now describes TRIPOD+AI in its library entry for TRIPOD and lists TRIPOD 2015 as a previous version.
What the 27-Item Checklist Actually Covers
The 27 main items are grouped into the title (item 1), abstract (item 2), introduction (items 3 and 4), methods (items 5 to 17), open science (item 18), patient and public involvement (item 19), results (items 20 to 24), and discussion (items 25 to 27). Counting sub-items, there are 52 things to report. Some apply only to model development (D) or only to model evaluation (E); a study that does both must address all of them. The table below summarises each group in plain language. Use the official checklist for the exact wording when you fill it in.
| Section | Items | What to report (summary) |
|---|---|---|
| Title | 1 | Identify the study as developing or evaluating a multivariable prediction model, and name the target population and the outcome |
| Abstract | 2 | Follow the separate 13-item TRIPOD+AI for Abstracts checklist |
| Introduction | 3a–3c, 4 | Healthcare context and rationale (with existing models), target population, intended purpose and users, known health inequalities, and whether the study covers development, evaluation, or both |
| Methods: data and participants | 5a–6c | Data sources for development and evaluation, their representativeness and dates, study setting and centres, eligibility criteria, and treatments received |
| Methods: data preparation, outcome, predictors | 7–9c | Pre-processing and quality checks, outcome definition and time horizon, outcome and predictor assessors and blinding, choice and definition of predictors |
| Methods: sample size and missing data | 10–11 | How the study size was arrived at and why it is sufficient, and how missing data were handled |
| Methods: analysis | 12a–12g | Data partitioning, predictor handling, model type and building steps (including hyperparameter tuning and internal validation), heterogeneity across clusters, performance measures such as discrimination, calibration, and clinical utility, model updating, and how predictions were calculated |
| Methods: other | 13–17 | Class imbalance methods, approaches to model fairness, model output and classification thresholds, differences between development and evaluation data, and ethics approval |
| Open science | 18a–18f | Funding, conflicts of interest, protocol, registration, data sharing, and code sharing |
| Patient and public involvement | 19 | Any patient and public involvement in design, conduct, reporting, interpretation, or dissemination, or a statement that there was none |
| Results | 20a–24 | Participant flow and characteristics, participants and outcome events in each analysis, the full model specification, performance estimates with confidence intervals (including key subgroups), heterogeneity, and any model updating |
| Discussion | 25–27c | Interpretation (including fairness), limitations, and usability in current care: handling poor-quality or missing inputs, the expertise users need, and next steps for research |
Summarised from Table 2 of Collins et al., BMJ 2024;385:e078378 (CC BY 4.0).
The title section deserves attention because it is one of the most frequently missed items. TRIPOD+AI asks that the type of study, whether development, external validation, or combined, be identifiable from the title. A title like "A machine learning model for predicting 30-day hospital readmission" passes only if it is clear whether the paper develops a new model, validates an existing one, or does both. Many titles still omit this. Reviewers looking at the title alone cannot tell whether they are reading a methodological contribution or a validation of someone else's prior work, and that distinction matters for how they evaluate the paper.
The methods section carries the heaviest load, and it is where most omissions cluster. TRIPOD+AI requires researchers to specify the study setting and time frame in enough detail to permit replication. It asks for explicit statements about how missing data were handled, because imputation method, complete case analysis, and listwise deletion can each produce substantially different estimates of model performance. It requires authors to describe all preprocessing steps for predictors, particularly for imaging or text-based inputs where normalization, augmentation, and feature extraction choices are often undisclosed. It asks for the predictor selection approach, whether variables were chosen a priori based on clinical reasoning or empirically during model training, and how any variable selection or dimensionality reduction step was performed and evaluated.
Calibration reporting is another persistent gap. Discrimination metrics like AUROC or the C-statistic describe how well a model ranks patients by risk. Calibration describes whether the predicted probabilities are actually accurate. A model that assigns 20 percent predicted probability to a group should, if well-calibrated, see around 20 percent of those patients experience the outcome. Many prediction model papers report discrimination only. TRIPOD+AI asks authors to specify all measures and plots used to evaluate performance, naming discrimination, calibration, and clinical utility as examples (item 12e), and to report performance estimates with confidence intervals (item 23a). The statement's glossary adds that calibration is best assessed graphically, with a calibration plot. A paper that reports only AUROC leaves readers unable to judge whether its predicted risks are accurate, and reviewers familiar with the statement will notice.
Checklist items most commonly omitted in published studies
- 1.Study type not specified in the title (development vs. validation)
- 2.Calibration statistics and calibration plot absent from results
- 3.No confidence intervals reported alongside performance metrics
- 4.Study not registered in a public registry before data collection
- 5.Sample size justification absent or post hoc
- 6.Missing data handling not specified
- 7.Code or model not deposited anywhere accessible
- 8.No subgroup performance analysis for fairness evaluation
Free TRIPOD+AI Worksheet
Print this page or copy the table into your manuscript folder, then fill in the page or line where each item is reported as you write. Items with letters have sub-items (52 in total); check the official checklist for the exact wording of each. Write "not applicable" with a reason rather than leaving a row blank. To fill in every sub-item with the official wording and download the result as a Word file, use our free fillable TRIPOD+AI checklist.
| Item | Topic | Page / line | Notes |
|---|---|---|---|
| 1 | Title | ||
| 2 | Abstract (TRIPOD+AI for Abstracts) | ||
| 3a–3c | Background | ||
| 4 | Objectives | ||
| 5a–5b | Data sources | ||
| 6a–6c | Participants | ||
| 7 | Data preparation | ||
| 8a–8c | Outcome | ||
| 9a–9c | Predictors | ||
| 10 | Sample size | ||
| 11 | Missing data | ||
| 12a–12g | Analytical methods | ||
| 13 | Class imbalance | ||
| 14 | Fairness | ||
| 15 | Model output | ||
| 16 | Training versus evaluation data | ||
| 17 | Ethical approval | ||
| 18a–18f | Open science: funding, conflicts, protocol, registration, data, code | ||
| 19 | Patient and public involvement | ||
| 20a–20c | Participants (results) | ||
| 21 | Model development | ||
| 22 | Model specification | ||
| 23a–23b | Model performance, overall and in key subgroups | ||
| 24 | Model updating | ||
| 25 | Interpretation | ||
| 26 | Limitations | ||
| 27a–27c | Usability of the model in context |
The New Domains That Set TRIPOD+AI Apart
Beyond updating the existing TRIPOD items, the 2024 statement introduced domains that reflect where clinical prediction research is heading and what funders, journals, and ethics boards are starting to require.
Fairness is a new domain. TRIPOD+AI asks whether model performance was assessed across subgroups of the study population, including subgroups defined by age, sex, race, or other clinically relevant characteristics. A model that performs well overall may perform substantially worse in a subgroup that was underrepresented in the training data or whose data came from a different care setting. Including a fairness analysis is not about political signaling. It is about whether the model actually works for the patients it will be applied to. A sepsis prediction model trained predominantly on one tertiary hospital's data may perform poorly at community hospitals, and that gap should be documented.
Reproducibility and open science requirements have also been strengthened. TRIPOD+AI asks for details of the availability of the study data (item 18e) and the analytical code (item 18f), and for details of the full prediction model, such as a formula, code, object, or programming interface, that would let others make predictions and evaluate it, including any restrictions on access (item 22). Journal policies are moving the same way. PLOS Medicine requires authors to share the code behind their results for manuscripts submitted from 22 April 2026, with access described in the data availability statement. Springer Nature's code policy asks for a code availability statement, and the Nature journals require code to be submitted for peer review. A paper that offers no pathway for independent validation is harder to defend in review.
Patient and public involvement is another new item. For prediction models that will inform clinical decisions affecting patients directly, TRIPOD+AI asks whether patients or members of the public were involved in identifying the study question, designing evaluation criteria, or interpreting findings. Most academic research teams have not been doing this routinely, and the item is aspirational for many groups at present. Its inclusion in the checklist signals where funder expectations are moving, and grant applications for clinical AI work are increasingly assessed on whether patient involvement is planned.
A note on class imbalance
TRIPOD+AI added a specific item on class imbalance (item 13), which matters for the many clinical prediction models where the outcome is rare. If you used oversampling, undersampling, or synthetic data generation (such as SMOTE), the checklist asks you to state why and how, and any methods used afterwards to recalibrate the model or its predictions. The recalibration part is there for a reason: in a simulation study, van den Goorbergh and colleagues (2022) found that imbalance corrections left logistic regression models poorly calibrated, overestimating the probability of the minority class, without improving discrimination. One further point is good practice rather than a checklist requirement: say whether resampling was applied only within the training data (or training folds) or to the full dataset before splitting, because resampling before the split lets information leak into the evaluation data and can inflate performance estimates.
The Abstracts Checklist Is Not Optional
One of the most consistently overlooked components of TRIPOD+AI is the separate 13-item checklist for the abstract. Early evidence suggests it is rarely met in full: in a systematic review of 53 COVID-19 prognostic model studies, no study fully adhered to the abstract reporting requirements (Partheniadis et al., 2026), and among 522 orthopaedic AI abstracts checked against four of the 13 items, none met all four (Singh, 2025). This matters because the abstract is often the only part of a paper that most readers see, and it is the section reviewers read first to decide whether the full manuscript is worth their time.
The abstract checklist asks authors to identify the study as developing or evaluating a multivariable prediction model and name the target population and outcome, explain the healthcare context, state whether the study covers development, evaluation, or both, describe the data sources, eligibility criteria, and setting, specify the outcome (with the prediction time horizon for prognostic models), the type of model, and the performance measures used, report the number of participants and outcome events, summarise the predictors in the final model (for development studies), report performance estimates with confidence intervals, give an overall interpretation, and give the registration number and registry name. These requirements sit alongside any word limits the target journal imposes, which creates real tension. A 250-word structured abstract does not have a lot of room for all of this.
The practical answer is to prioritize metrics over prose. An abstract that opens with a single background sentence, spends three sentences on the study population and outcome, reports performance estimates such as discrimination and calibration with confidence intervals, and closes with the registration number and registry will satisfy reviewers better than one with eloquent framing but missing numbers. Background is the part of an abstract that readers understand from context. Performance with confidence intervals is the part they cannot infer. If space is tight, trim the background.
What the Adherence Studies Tell Us About the Gaps
The first studies to measure reporting against TRIPOD+AI point in the same direction. A bibliometric study in Cureus (Singh, November 2025) compared 522 orthopaedic AI prediction model abstracts published in the 18 months before and the 18 months after TRIPOD+AI appeared, scoring four items from the abstract checklist. Confidence intervals around performance estimates appeared in 18.7% of abstracts before and 16.6% after. Registration details appeared in 0.5% and 1.0%. No abstract met all four items in either period, none of the changes was statistically significant, and the author concluded that 18 months on, TRIPOD+AI had not measurably improved reporting in orthopaedic AI abstracts. The study looked only at abstracts and used a single rater, so it says nothing about full-text reporting.
A systematic review of COVID-19 prognostic models in Diagnostic and Prognostic Research (Partheniadis et al., February 2026) assessed 53 studies, applying TRIPOD+AI to the machine learning studies and TRIPOD 2015 to the conventional ones. Overall adherence was 28.4% for the machine learning studies against 38.1% for the conventional studies. Appropriate titles were rare (13.6% of the machine learning studies), and no study fully reported a sample size calculation. The authors flag an important caveat: their search ran to July 2024, so most of the machine learning studies were published before TRIPOD+AI existed. Read the 28% as a baseline, not a verdict on the guideline.
None of this is new. The TRIPOD+AI statement itself cites more than a dozen systematic reviews, in fields from oncology and orthopaedics to stroke and emergency triage, that found prediction model studies incompletely reported, with key details missing.
To make studies like these comparable, the team behind the original TRIPOD adherence tool, which includes TRIPOD+AI authors Gary Collins and Karel Moons, has updated it for TRIPOD+AI (de Kanter et al., Journal of Clinical Epidemiology, published online 23 December 2025). The new tool breaks the guideline into 37 main items and 136 adherence elements, includes several automated scoring rules, and comes in separate versions for development studies, evaluation studies, and studies that do both. Its stated purpose is to make adherence assessments comparable across fields and over time. For your own manuscript, the fillable checklist on tripod-statement.org remains the right tool.
The implication for authors is straightforward. Several of the items these studies find missing, notably registration and sample size justification, along with fairness planning and data or code sharing, depend on decisions made before data collection begins. You cannot register a study after the analysis is done and pass that off as prospective registration. A sample size calculation done after the fact is not a justification. These items have to be in the protocol.
Writing the Methods Section Under TRIPOD+AI
The most practical advice is to read the checklist before you finalize your study protocol, not when you sit down to write the manuscript. Once you have the checklist in hand, go through the methods section item by item and write the corresponding paragraph. The result will be a longer methods section than many researchers are used to, but it will be the kind of methods section that allows a reader to understand exactly what was done.
The data sources subsection should cover not just where the data came from but when it was collected, who it covered, what exclusion criteria were applied, and whether the development and validation sets came from the same or different sources. If you are using retrospective electronic health record data from a single hospital, that is important context. If you are using a multi-site registry, name the sites or at least their characteristics. Reviewers for prediction model papers routinely look for generalizability, and a vague "data were collected from a tertiary care center" raises questions that a sentence or two of additional context can answer.
The predictors subsection should name every variable that entered the candidate predictor set, describe how each was measured or extracted, and note any transformations applied before model training. For imaging-based models, this includes the preprocessing pipeline. For models using clinical notes, it includes the text processing approach. For structured EHR data, it includes how continuous variables were handled (raw value, binned, log-transformed) and how categorical variables were encoded. These details are not incidental. Two models trained on the same clinical question with different preprocessing choices will perform differently, and a reader who cannot reproduce your preprocessing pipeline cannot validate your model.
Template: Model performance reporting under TRIPOD+AI
Model discrimination was assessed using the area under the receiver operating characteristic curve (AUROC) with 95% confidence intervals calculated by bootstrap resampling (1000 iterations). Calibration was assessed using the Brier score and visually inspected using a calibration plot comparing mean predicted probabilities against observed event rates across deciles of predicted risk. Subgroup performance was assessed in pre-specified strata by age (<65 vs. 65 years), sex, and primary diagnosis category.
Notice what this template does. It names the discrimination metric and how the confidence interval was calculated. It names the calibration metric and describes how calibration was visualized. It specifies that subgroup analyses were pre-specified, not exploratory, and names the strata. Each of these details answers a TRIPOD+AI checklist item. Writing your methods with the checklist open is slower than writing from memory, but it produces manuscripts that clear review faster.
Registering a Prediction Model Study and Depositing Your Code
Two TRIPOD+AI items cause the most difficulty for research teams that are encountering the checklist for the first time: study registration and code or model deposit. Both feel like bureaucratic overhead, and both require decisions that have to happen early.
Prediction model studies, unlike randomized controlled trials, do not have a mandatory registry. ClinicalTrials.gov accepts observational and prediction studies but is not routinely used for them. OSF (the Open Science Framework) is the most commonly used registry for prediction model and other observational study pre-registrations. Zenodo is another option. A pre-registration does not constrain your analysis permanently. Deviations from the pre-registered plan are acceptable and should be reported as deviations, because readers and reviewers can then distinguish confirmatory from exploratory findings. What the registration establishes is that the primary research question, outcome, and analysis approach were decided before you saw the data.
For code deposit, the practical options for most academic research teams are GitHub, OSF, and Zenodo. Depositing code does not require making patient data public. The model weights, preprocessing scripts, and evaluation code can be separated from the clinical records and shared independently. A GitHub repository with a README describing the data format required to run the model, even if the original data cannot be shared, is the kind of code availability items 18f and 22 ask you to describe, and it gives other researchers a starting point for validation on their own data. Archive a release in a repository that issues a DOI, such as Zenodo, as well: Springer Nature's code policy, for example, says a GitHub link alone is not sufficient because it does not assign a permanent identifier.
If your institutional review board prohibits data sharing and you cannot share code because doing so would expose identifiable information, TRIPOD+AI accepts a clear statement explaining this. The statement should name the specific restriction (IRB prohibition, data sharing agreement, commercial sensitivity) rather than simply saying "data are available on reasonable request," which on its own tells readers little about whether, or how, they could get access. If there is a genuine reason, write it plainly.
Getting the Journal Submission Right
The TRIPOD+AI authors recommend submitting a completed checklist that gives the page or line where each item is reported, and a fillable template is available on tripod-statement.org. Where a journal requires reporting checklists, the completed form goes in with the manuscript; BMJ's author guidance, for example, asks for a file titled "Research Checklist" that lists the page where each item appears. If an item cannot be reported because the information is unknown or irrelevant, the statement says to acknowledge this clearly. Simply entering N/A without context will flag a gap in peer review, and a reviewer familiar with the checklist will likely ask about it.
If you are submitting to a journal whose author instructions do not explicitly mention TRIPOD+AI but do require reporting guideline checklists for clinical research, include TRIPOD+AI regardless. The EQUATOR Network's library entry for TRIPOD now describes TRIPOD+AI, for studies that develop a prediction model or evaluate its performance, and lists TRIPOD 2015 as a previous version. A checklist submitted voluntarily is not a reason for rejection. A missing checklist when the journal expects one sometimes is, and a reviewer who knows the standard and finds the manuscript incomplete will say so.
For journals that have not yet explicitly adopted TRIPOD+AI, it is worth checking whether the journal follows EQUATOR recommendations or publishes guidance specifically for clinical prediction model papers. Author instructions often lag behind the guideline: as of October 2026, BMJ's own author guidance still lists TRIPOD, not TRIPOD+AI, for prediction model studies, even though TRIPOD+AI was published in The BMJ. When a journal's instructions name only TRIPOD 2015, using TRIPOD+AI and noting in your cover letter that you have applied the updated 2024 standard is an appropriate approach and is unlikely to generate editorial pushback.
Before you submit: a quick self-check
Run through each of the following before uploading your manuscript:
- Study type (development / validation / both) is stated in the title.
- Abstract names the performance measures used (for example, discrimination and calibration) and reports estimates with confidence intervals.
- Abstract gives the registration number and registry name.
- Methods describe missing data handling explicitly.
- All preprocessing steps for predictors are named.
- Calibration plot or calibration statistics appear in the results.
- Performance is reported by pre-specified subgroups (fairness evaluation).
- Registration details appear in the manuscript, or it states that the study was not registered (item 18d).
- Data and code availability are described (items 18e and 18f), with a justified exception where sharing is not possible.
- The completed TRIPOD+AI checklist with page references is attached to the submission.
What Poor Adherence Actually Costs You
A common response to reporting checklists is that they slow down writing without improving science. That misreads what the checklist is for. The reporting gaps that TRIPOD+AI is designed to close, missing calibration, undescribed preprocessing, absent confidence intervals, are not cosmetic problems. They are the gaps that make it impossible for other teams to validate a model, apply it to a different population, or determine whether it would work in their clinical setting.
The clinical AI literature already has a well-documented problem with models that perform impressively in development papers and then fail when used at other institutions. External validation studies routinely find much lower performance than development papers reported. Some of this is genuine generalizability failure. Some of it is preventable reporting omission: if the development paper had described preprocessing clearly, the validation team would have applied the same steps and compared models on a fairer basis.
There is also a practical publication benefit. At journals with statistical review, the reviewer may know the checklist item by item: two of the TRIPOD+AI authors, Gary Collins and Richard Riley, are statistics editors at The BMJ. A paper with a complete checklist and methods that answer every item gives that reviewer far less to list as missing in the first round. Treating TRIPOD+AI as a writing framework rather than a post-hoc audit is, in the end, less work overall.
References
- Collins GS, Moons KGM, Dhiman P, et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ 2024;385:e078378. doi:10.1136/bmj-2023-078378
- Singh S. Effect of TRIPOD+AI guidelines on the reporting quality of artificial intelligence prediction models in orthopaedic surgery: an 18-month bibliometric study. Cureus 2025;17(11):e97176. doi:10.7759/cureus.97176
- Partheniadis I, Talimtzi P, Nikolakopoulou A, Haidich AB. Machine learning-based COVID-19 prognostic models lag behind in reporting quality: findings from a TRIPOD/TRIPOD + AI systematic review. Diagn Progn Res 2026;10:3. doi:10.1186/s41512-026-00218-x
- de Kanter E, Kaul T, Heus P, et al. Adherence to TRIPOD+AI guideline: an updated reporting assessment tool. J Clin Epidemiol 2026;191:112118. doi:10.1016/j.jclinepi.2025.112118
- van den Goorbergh R, van Smeden M, Timmerman D, Van Calster B. The harm of class imbalance corrections for risk prediction models: illustration and simulation using logistic regression. J Am Med Inform Assoc 2022;29(9):1525–1534. doi:10.1093/jamia/ocac093
- Journal and library policies cited (checked October 2026): PLOS Medicine code availability, Springer Nature code policy, BMJ author guidance on reporting guidelines, and the EQUATOR Network TRIPOD entry.
Further Reading
Publishing Medical AI Research in 2026
What journals now expect from AI research papers, from study design through to software and data availability.
CONSORT 2025 Clinical Trial Reporting
The parallel update to clinical trial reporting standards and what changed for intervention studies.
How to Disclose AI Use in Medical Manuscripts
Separate from TRIPOD+AI: the disclosure policies that apply when you use AI tools to write the manuscript itself.
How to Read Journal Author Guidelines
How to find the specific reporting guideline requirements before you choose a target journal.
Written by Dr. Meng Zhao
Physician-Scientist · Founder, LabCat AI
MD · Former Neurosurgeon · Medical AI Researcher
Dr. Meng Zhao is a former neurosurgeon turned medical-AI researcher. After years in the operating room, he moved into applied AI for clinical workflows and now leads LabCat AI, a medical-AI company working on decision support and research tooling for clinicians. He built Journal Metrics as a free resource for researchers who need reliable journal metrics without paid database subscriptions.
New on PubMed
What's new in your field?
Type a topic to see the newest PubMed papers from the past week and the big picture of what they found, summarized by Claude on the Pro plan. Get a free weekly email to keep up.
See a demo:GLP-1CRISPRLong COVID
Related Articles
Heterogeneity in Meta-Analysis: How to Interpret and Report I², Q and Tau²
How to interpret and report heterogeneity in a meta-analysis: what I², Cochran's Q and tau² each measure, why none of them tells you how different the studies really are, and why a prediction interval belongs in your results.
12 min readWriting GuideZotero vs Mendeley in 2026: Which Reference Manager Should Researchers Use?
Zotero vs Mendeley for medical researchers in 2026: cost and storage, citation styles for switching journals, Word and LaTeX integration, moving a library between tools, and how to choose.
11 min readWriting GuideFixed vs Random Effects Meta-Analysis: Which Model to Report
Fixed vs random effects meta-analysis explained for medical authors: what each model assumes, how to choose and pre-specify one, when to use Hartung-Knapp and prediction intervals, and how to write the methods sentence reviewers expect.
11 min read