Semaglutide on liver fibrosis and heart outcomes in patients at high risk of liver fibrosis: a prespecified analysis of the SELECT randomized trial.
Meyhöfer SM, Cariou B, Cercato C, Colhoun HM, Deanfield J, Long MT, Jeppesen OK, Lincoff AM, Lingvay I, Plutzky J, Newsome PN, Nicholls SJ, Quiroga M, Santini F, Sanyal AJ, Kahn SE, SELECT Trial Investigators
- DOI
- 10.1038/s41591-026-04281-1
- Record issued
- 2026-08-15
- Engine
- 7.39.0
- Exported
- 2026-09-20
Prepared by Alpha1. This document is confidential: it is intended for the recipient it was shared with and must not be redistributed. The live record at alpha1science.com/verify/3ad6e115-11e7-4645-b3fd-f5bb30ad5bd1 is authoritative.
How this rating was calculated
- ClaimsEfficacy rests on an unvalidated surrogate endpoint−0.5★
- StatisticsPrinted percentage does not match its own count (capped) ×2−0.25★
- The numeric-impossibility checks (GRIM/GRIMMER/DEBIT/SPRITE) did not run: 18 reported means were read, and their group size is not stated where the values are printed (this source has no machine-readable table structure). These checks need the count the mean was averaged over, so none was performed.
- 01Efficacy rests on an unvalidated surrogate endpoint
The primary efficacy claim is reduction in MACE, a hard clinical outcome, so the surrogate assessment is not applicable to the primary claim. However, the paper also claims improvements in liver health based on surrogate markers (ALT, AST, GGT, FLI) without demonstrating target engagement or a validated link to clinical outcomes. The primary MACE outcome is adequate, but the liver-related claims rely on surrogates.
“Semaglutide led to a 28% greater decrease in fatty liver index versus placebo (HR 0.72; 95% CI 0.71–0.73; P < 0.0001).”
- 02Printed percentage does not match its own count
22.7% does not match the reported count 59/263
“59 (22.7)”
Table 1 - 03Printed percentage does not match its own count
4% does not match the reported count 128/3293
“128 (4.0)”
Table 1
This Kaimen Rigor review uses Kaimen Rigor reviewers trained on a curated corpus of high-fidelity and retracted papers, with expert supervision and curation. It can still make mistakes; verify each finding against the source before relying on it.
This is a well-conducted prespecified secondary analysis of a large randomized controlled trial, with strong reporting of design, ethics, and data availability. The main weaknesses are minor: lack of explicit power analysis for subgroups, incomplete randomization details, and no explicit reporting guideline. The statistics verification found 2 inconsistencies among 9 recomputed tests, but no decision errors, so the core findings appear robust.
Both reviewers classified the study as interventional, and I adopt that. The evaluation covered the full text, with verification components for citations, statistics, reproducibility, and preregistration. The statistics component only recomputed a subset of tests (9 of many), so unverified tests should be treated with caution. The 2 inconsistent recomputations are not detailed and did not affect decisions, but they warrant a reader's attention.
Numerical inconsistencies
1 finding · worst mediumValues that contradict each other or are impossible for the stated sample: recomputed p-values and test statistics, GRIM/GRIMMER checks on summary numbers, percentages against their own counts, totals against their parts, and estimates against their own confidence intervals.
- Printed percentage does not match its own countRecomputed
Recomputed 6 tests: 6 consistent, 0 inconsistent; 3 recomputed directly from the reported test statistics, 3 via agent-written checks. 2 printed percentages that do not match their own count.
- PERCENT22.7% does not match the reported count 59/263
“59 (22.7)”
Table 1 - PERCENT4% does not match the reported count 128/3293
“128 (4.0)”
Table 1
- CONSISTENTreported p = .035 · recomputed p = .036Recomputed HR 0.79 (95% CI 0.63–0.98), reported p=0.035
“HR 0.79; 95% CI 0.63–0.98; P = 0.035”
Taken as given: 0.63–0.98 is a two-sided 95% confidence interval for the HR of 0.79, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.035 is the p for THIS estimate, not for another comparison in the same sentenceMethod: back the two-tailed p out of the log-scale CI width and compare it against the printed pHow we recomputed it: pCI(0.79, 0.63, 0.98, 1) - CONSISTENTreported p = .110 · recomputed p = .116Recomputed HR 0.66 (95% CI 0.39–1.10), reported p=0.11
“HR 0.66; 95% CI 0.39–1.10; P = 0.11”
Taken as given: 0.39–1.10 is a two-sided 95% confidence interval for the HR of 0.66, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.11 is the p for THIS estimate, not for another comparison in the same sentenceMethod: back the two-tailed p out of the log-scale CI width and compare it against the printed pHow we recomputed it: pCI(0.66, 0.39, 1.1, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Recomputed HR 0.74 (95% CI 0.63–0.88), reported p=0.0004
“HR 0.74; 95% CI 0.63–0.88; P = 0.0004”
Taken as given: 0.63–0.88 is a two-sided 95% confidence interval for the HR of 0.74, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.0004 is the p for THIS estimate, not for another comparison in the same sentenceMethod: back the two-tailed p out of the log-scale CI width and compare it against the printed pHow we recomputed it: pCI(0.74, 0.63, 0.88, 1) - UNCOMPUTABLEreported p = .480 · recomputed p = .141Recomputed HR 0.88 (95% CI 0.74–1.04), reported p=0.48
“HR 0.88; 95% CI 0.74–1.04; P = 0.48”
Taken as given: 0.74–1.04 is a two-sided 95% confidence interval for the HR of 0.88, not a range, an IQR, or a different interval level; the HR is a RATIO measure, so the interval is symmetric on the log scale; p=0.48 is the p for THIS estimate, not for another comparison in the same sentenceMethod: back the two-tailed p out of the log-scale CI width and compare it against the printed pHow we recomputed it: pCI(0.88, 0.74, 1.04, 1) - CONSISTENTreported p < .001 · recomputed p = <.001Reviewers 1, 2HR for MACE in FIB-4 >= 1.3 subgroup
“HR 0.74; 95% CI 0.63–0.88; P = 0.0004”
Taken as given: The HR is 0.74 and the 95% CI is 0.63-0.88.; The CI is two-sided at 95%.; The p-value is two-sided.Method: Recomputed p-value from the reported HR and CI using the normal approximation for the log hazard ratio.How we recomputed it: pCI(0.74, 0.63, 0.88, 1) - CONSISTENTreported p = .035 · recomputed p = .036Reviewers 1, 2HR for MACE in age-specific FIB-4 subgroup
“HR 0.79; 95% CI 0.63–0.98; P = 0.035”
Taken as given: The HR is 0.79 and the 95% CI is 0.63-0.98.; The CI is two-sided at 95%.; The p-value is two-sided.Method: Recomputed p-value from the reported HR and CI using the normal approximation for the log hazard ratio.How we recomputed it: pCI(0.79, 0.63, 0.98, 1) - CONSISTENTreported p = .110 · recomputed p = .116Reviewers 1, 2HR for MACE in FIB-4 > 2.67 subgroup
“HR 0.66; 95% CI 0.39–1.10; P = 0.11”
Taken as given: The HR is 0.66 and the 95% CI is 0.39-1.10.; The CI is two-sided at 95%.; The p-value is two-sided.Method: Recomputed p-value from the reported HR and CI using the normal approximation for the log hazard ratio.How we recomputed it: pCI(0.66, 0.39, 1.10, 1)
Overstated conclusions
1 finding · worst highConclusions that reach past what the paper's own results support — including a significance claim that no longer holds when the statistic is recomputed, and efficacy resting on an unvalidated surrogate endpoint.
- Efficacy rests on an unvalidated surrogate endpointAssessed
6 major claims checked against the paper's own evidence: all adequately supported.
- supportedReviewer 1Semaglutide reduces MACE in patients at high risk for substantial liver fibrosis.The claim is supported by the reported HRs and CIs in the FIB-4 >= 1.3 subgroup, which show a statistically significant reduction.Evidence: HR 0.74; 95% CI 0.63–0.88; P = 0.0004
“MACE was reduced by 26% (hazard ratio (HR) 0.74; 95% confidence interval (CI) 0.63–0.88; P = 0.0004)”
AbstractFind in source - supportedReviewer 1Semaglutide leads to a greater decrease in fatty liver index versus placebo.The claim is supported by the reported treatment difference and p-value.Evidence: 28% greater decrease in FLI (95% CI 0.71–0.73; P < 0.0001)
“Semaglutide led to a 28% greater decrease in fatty liver index versus placebo (HR 0.72; 95% CI 0.71–0.73; P < 0.0001).”
AbstractFind in source - supportedReviewer 1Semaglutide improves liver enzymes (ALT, AST, GGT) compared with placebo.The claim is supported by the reported estimated treatment differences and p-values.Evidence: ETD at week 104: ALT -1.9 U/l (95% CI -2.6 to -1.2; P < 0.0001), AST -0.9 (95% CI -1.4 to -0.5; P < 0.0001), GGT -8.2 (95% CI -9.5 to -6.8; P < 0.0001)
The corresponding estimated treatment difference versus placebo at week 104 was −1.9 U l−1 (95% CI −2.6 to −1.2; P < 0.0001), −0.9 U l−1 (95% CI −1.4 to −0.5; P < 0.0001) and −8.2 U l−1 (95% CI −9.5 to −6.8; P < 0.0001), respectively.
Resultsreviewer’s wording - supportedReviewers 1, 2The MACE reduction in the high-risk fibrosis subgroup is consistent with the overall SELECT population.The claim is supported by the HRs in the subgroups, which are similar to the overall 20% reduction, though not all subgroups reach statistical significance.Evidence: HR 0.74 (0.63-0.88) in FIB-4 >= 1.3; HR 0.79 (0.63-0.98) in age-specific; HR 0.66 (0.39-1.10) in FIB-4 > 2.67
“As shown here, the primary results from the SELECT trial of a 20% MACE reduction in patients with established ASCVD and overweight or obesity is also seen in the subgroup of these patients with concurrent MASLD and at risk for substantial fibrosis.”
DiscussionFind in source - supportedReviewer 2Semaglutide reduces MACE in patients at high risk of liver fibrosis.The claim is supported by the reported HRs and CIs in the FIB-4 ≥1.3 subgroup (HR 0.74, 95% CI 0.63-0.88, P=0.0004).Evidence: HR 0.74; 95% CI 0.63–0.88; P = 0.0004
“In the cohort with an FIB-4 score ≥ 1.3, the primary MACE endpoint occurred in 244 of 3,233 patients (7.5%) in the semaglutide group and in 321 of 3,274 patients (9.8%) in the placebo group (HR 0.74; 95% CI 0.63–0.88; P = 0.0004; Fig. ).”
ResultsFind in source - supportedReviewer 2Semaglutide improves liver enzymes and fatty liver index.The claim is supported by the reported reductions in ALT, AST, GGT, and FLI with significant p-values.Evidence: Mean ± s.d. FLI decreased by 18.63 ± 0.21 with semaglutide treatment versus 2.49 ± 0.19 with placebo (Fig. ), representing a treatment difference of 16.1 (95% CI −16.7 to −15.6; P < 0.0001)
“Mean ± s.d. FLI decreased by 18.63 ± 0.21 with semaglutide treatment versus 2.49 ± 0.19 with placebo (Fig. ), representing a treatment difference of 16.1 (95% CI −16.7 to −15.6; P < 0.0001)”
ResultsFind in source
Premise concern: surrogate not validated for clinical benefit.
- INADEQUATESurrogate endpointThe primary efficacy claim is reduction in MACE, a hard clinical outcome, so the surrogate assessment is not applicable to the primary claim. However, the paper also claims improvements in liver health based on surrogate markers (ALT, AST, GGT, FLI) without demonstrating target engagement or a validated link to clinical outcomes. The primary MACE outcome is adequate, but the liver-related claims rely on surrogates.
“Semaglutide led to a 28% greater decrease in fatty liver index versus placebo (HR 0.72; 95% CI 0.71–0.73; P < 0.0001).”
- ADEQUATEEffect sizeThe primary effect size for MACE reduction is 20% (HR 0.80) in the overall trial, and in the high-risk fibrosis subgroup it is 26% (HR 0.74). These are clinically meaningful reductions in hard outcomes, anchored to established cardiovascular risk reduction.
“MACE was reduced by 26% (hazard ratio (HR) 0.74; 95% confidence interval (CI) 0.63–0.88; P = 0.0004)”
Data authenticity concerns
None foundAn adversarial read for patterns associated with data that may not be genuine: results that look too clean, implausibly large effects, duplicated data or images, and methods that do not match the results reported.
Checked — nothing surfaced.
Reporting gaps
None foundRequired detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.
Checked — nothing surfaced.
The introduction cites prior research on the interplay between MASLD, fibrosis, and MACE, and the rationale for using FIB-4 is well explained. The limitations of prior work (e.g., reliance on liver biopsy) are acknowledged, and the study addresses this by using noninvasive markers. The hypothesis follows logically from the cited evidence.
“The coexistence of overweight and obesity with metabolic dysfunction-associated steatotic liver disease (MASLD) underscores a substantial and intricate interplay between metabolic dysregulation and hepatic pathology.”
“The aim of this prespecified analysis was to evaluate the impact of semaglutide on liver enzymes and suspected MASLD in patients from the SELECT trial, and to explore the CV benefits of semaglutide in a subgroup at risk for substantial fibrosis as defined by FIB-4.”
“The coexistence of overweight and obesity with metabolic dysfunction-associated steatotic liver disease (MASLD) underscores a substantial and intricate interplay between metabolic dysregulation and hepatic pathology.”
Randomization method and unit are not detailed in this paper but are referenced from the primary SELECT publication. Blinding is described as double-blind. Power analysis is not reported for this subgroup analysis, but the primary trial was event-driven and the analysis is prespecified. Inclusion/exclusion criteria are clearly stated. Outlier handling is addressed through multiple imputation for missing data. Controls are inherent in the placebo group. Independent replication is not applicable for a single trial.
“The current work reports a prespecified analysis of SELECT, a randomized, double-blind, placebo-controlled, event-driven CV outcome trial”
“Missing data at week 104 were imputed using a multiple imputation model, conducted separately for each treatment arm and including baseline value as a covariate”
“The current work reports a prespecified analysis of SELECT, a randomized, double-blind, placebo-controlled, event-driven CV outcome trial”
“Missing data at week 104 were imputed using a multiple imputation model, conducted separately for each treatment arm and including baseline value as a covariate”
Sex is reported for all subgroups. Age, weight, BMI, and health status (e.g., prediabetes, CVD history) are reported. Demographics are reported in Table 1. Species/strain and housing conditions are not applicable for a human trial.
“Female, n (%) | 765 (23.2) | 743 (22.7) | 354 (19.3) | 350 (19.1) | 37 (15.0) | 59 (22.7)”
“Age, years | 66.7 (7.9) | 66.6 (7.8) | 63.5 (8.4) | 63.4 (8.2) | 71.8 (8.1) | 70.5 (8.6)”
“Pre-diabetes, n (%) | 2,061 (62.6) | 2,033 (62.1) | 1,120 (61.1) | 1,108 (60.5) | 147 (59.8) | 145 (55.1)”
“Female, n (%) | 765 (23.2) | 743 (22.7) | 354 (19.3) | 350 (19.1) | 37 (15.0) | 59 (22.7)”
“Age, years | 66.7 (7.9) | 66.6 (7.8) | 63.5 (8.4) | 63.4 (8.2) | 71.8 (8.1) | 70.5 (8.6)”
The paper states that the protocol was approved by the institutional review board and ethics committee at each participating center, and that all patients provided written informed consent. This satisfies both irb_ethics_statement and informed_consent. Regulatory compliance is implied by the trial registration and adherence to Good Publication Practice.
“All patients provided written informed consent before commencement of any trial-specific activity.”
“ClinicalTrials.gov registration no. NCT03574597”
“All patients provided written informed consent before commencement of any trial-specific activity.”
Semaglutide is identified as once-weekly subcutaneous 2.4 mg, with manufacturer implied (Novo Nordisk). The statistical software SAS v9.4 is identified. No antibodies, cell lines, or other biological reagents are used, so those criteria are not applicable.
“The trial evaluated once-weekly subcutaneous semaglutide 2.4 mg versus placebo”
“All statistical analyses were performed using the SAS software v.9.4 (SAS Institute).”
“The trial evaluated once-weekly subcutaneous semaglutide 2.4 mg versus placebo”
“All statistical analyses were performed using the SAS software v.9.4 (SAS Institute).”
Tests are named (Cox proportional hazards, ANCOVA). Assumptions are handled by design (e.g., competing risk model). Exact p-values are reported for primary outcomes. Effect sizes with CIs are reported. Software is identified. Data presentation includes Kaplan-Meier curves and forest plots. Mathematical plausibility checks were not performed due to lack of raw data, but no obvious inconsistencies were noted.
“Time-to-event endpoints were analyzed using a Cox proportional hazards model with treatment group (semaglutide, placebo) as fixed factors together with the two-sided 95% CI and two-sided P values”
“HR 0.74; 95% CI 0.63–0.88; P = 0.0004”
“HR 0.74; 95% CI 0.63–0.88”
“Time-to-event endpoints were analyzed using a Cox proportional hazards model with treatment group (semaglutide, placebo) as fixed factors together with the two-sided 95% CI and two-sided P values”
“HR 0.74; 95% CI 0.63–0.88; P = 0.0004”
The data availability statement provides a clear route for requesting access via Novo Nordisk's data-sharing process, including conditions and contact. This is adequate for patient-level data. No code is shared, but no bespoke code is mentioned, so code_sharing is not applicable.
“Authorized researchers can request access to clinical trial data by submitting a research proposal for review and approval by Novo Nordisk and an internal independent review panel.”
“Authorized researchers can request access to clinical trial data by submitting a research proposal for review and approval by Novo Nordisk and an internal independent review panel.”
Methods are detailed enough for replication. Trial registration is provided. A reporting summary is mentioned. All outcomes are reported, including non-significant ones. Limitations are discussed. Conclusions are proportional. Funding and COI are disclosed.
“ClinicalTrials.gov registration no. NCT03574597”
“Although the findings of this analysis present compelling evidence for the clinical relevance of semaglutide in addressing ASCVD and MASLD in individuals living with overweight or obesity, it is essential to acknowledge potential methodological limitations.”
“ClinicalTrials.gov registration no. NCT03574597”
“As this analysis was not powered to reach significance because of the subcategory population sizes in the intermediate-risk and high-risk FIB-4 groups, further investigations into these populations are needed.”
Registered (1 ID: ClinicalTrials.gov). No reporting guideline cited.
Broken references and links
None foundReferences checked against Crossref, OpenAlex and Retraction Watch for retractions and resolvability, plus declared data and code links probed for whether they resolve to content matching the paper.
Checked — nothing surfaced.
Checked 39 references by DOI: 38 verified — 1 no DOI (shown, not verified).
- NO DOIObesity and obesity-induced inflammatory disease contribute to atherosclerosis: a review of the pathophysiology and treatment of obesityNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
3 data/code links checked; 3 live.
- datahttp://clinicaltrials.gov/study/NCT03574597LIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttps://clinicaltrials.gov/ct2/show/NCT03574597LIVEHTTP 200Resolves, but the content could not be matched to the paper.
- datahttp://novonordisk-trials.comLIVEHTTP 200Resolves, but the content could not be matched to the paper.
Copyediting
5 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 5 minor suggestions below.
5 copyedit issues flagged: mostly consistency, clarity.
- MINORconsistencyAbstract“MACE was reduced by 26% (hazard ratio (HR) 0.74; 95% confidence interval (CI) 0.63–0.88; P = 0.0004)”→ Ensure the percentage reduction (26%) is consistent with the HR (0.74 implies 26% reduction).The percentage is correctly derived from the HR, but the abstract states 'reduced by 26%' while the HR is 0.74, which is consistent.
- MINORclarityResults, Cardiovascular outcomes“In the second subgroup with an FIB-4 score ≥ 1.3 or aged ≥65 years with an FIB-4 score ≥ 2.0”→ Clarify the subgroup definition to avoid ambiguity: 'FIB-4 ≥ 1.3 for patients <65 years or FIB-4 ≥ 2.0 for patients ≥65 years'.The phrasing is slightly ambiguous but understandable.
- MINORconsistencyTable 1“Statins | 2,914 (88.5) | 2,925 (89.3) | 1,625 (88.7) | 1,646 (89.9) | 211 (85.8) | 224 (85.2)”→ Check that the percentages are consistent with the group sizes.The percentages appear consistent with the group sizes.
- MINORconsistencyAbstract“MACE was reduced by 26% (hazard ratio (HR) 0.74; 95% confidence interval (CI) 0.63–0.88; P = 0.0004)”→ Ensure consistent use of 'hazard ratio' vs 'HR' throughout.Minor stylistic inconsistency.
- MINORclarityResults, Cardiovascular outcomes“In the second subgroup with an FIB-4 score ≥ 1.3 or aged ≥65 years with an FIB-4 score ≥ 2.0”→ Clarify the subgroup definition to avoid ambiguity.The phrasing could be misinterpreted.
The published work is robust and well-reported. An informed reader should weigh the minor reporting gaps (no explicit power analysis for subgroups, incomplete randomization details, no named reporting guideline) and the 2 inconsistent statistics recomputations, which are not decision-altering but suggest a need for careful scrutiny. No erratum is warranted based on the available evidence.
- 1.HIGHstatisticsIn the Methods/Statistical analysis, investigate and reconcile the 2 inconsistent statistics recomputations flagged by the verification component; if they reflect errors, issue a correction.Even though no decision errors were found, unexplained inconsistencies in reported statistics undermine reader confidence and may warrant an erratum.
- 2.HIGHreportingIn the Methods or Reporting Summary, explicitly state adherence to a reporting guideline (e.g., CONSORT) for this secondary analysis.Reviewer 2 noted the reporting guideline is not explicitly named, which is a minor transparency gap that is easily fixed.
- 3.MEDIUMrigorIn the Methods, add a statement about the power or sample size considerations for this subgroup analysis, acknowledging that it was not powered for these subgroups.Both reviewers flagged the lack of power analysis for subgroups; adding this context helps readers interpret null or marginal findings.
- 4.MEDIUMrigorIn the Methods, specify the randomization method and unit (e.g., central randomization, patient-level) as described in the primary SELECT publication.Reviewer 1 noted randomization details are incomplete in this paper; referencing the primary publication would fully satisfy the criterion.
- 5.MEDIUMreportingIn the Abstract, explicitly state that this is a prespecified secondary analysis to avoid ambiguity.Reviewer 2 suggested this to clarify the nature of the analysis for readers.
- 6.MEDIUMreportingIn the Acknowledgements or Competing Interests, clarify the role of the funder (Novo Nordisk) in the analysis and interpretation.Reviewer 2 suggested this to enhance transparency regarding funder involvement.
- 7.MEDIUMdata codeIn the Data Availability section, consider adding a statement about the availability of statistical analysis code, if any, to enhance reproducibility.Reviewer 1 suggested this; even if no code is shared, stating its unavailability is transparent.
- 8.MEDIUMreportingIn the Discussion, explicitly mention the limitations of using FIB-4 as a surrogate for fibrosis and the lack of histological confirmation.Reviewer 1 suggested this to strengthen the limitations discussion.
- 9.MEDIUMreportingIn the Results, report the number of patients with missing FIB-4 data and how missingness was handled in the analysis.Reviewer 1 suggested this to improve transparency about missing data.
- 10.LOWcopyeditIn the Results, clarify the subgroup definition: 'FIB-4 ≥ 1.3 for patients <65 years or FIB-4 ≥ 2.0 for patients ≥65 years'.The copyedit pass flagged the current phrasing as ambiguous.
- 11.LOWcopyeditIn the Abstract, ensure consistent use of 'hazard ratio' vs 'HR' throughout.The copyedit pass noted a minor stylistic inconsistency.
The star rating is the report’s one-glance summary. Every paper starts at 5★ and loses stars for the concrete problems the review finds — so a rating is never a vague average, it’s a running total you can read line by line under “How this rating was calculated.”
- Reporting — 8 dimensionseach dimension that fully fails−½★
- each dimension partially met−¼★
- Statistics · Integrity · Claimseach serious problem−1★
- each medium problem−½★
- Citationseach retracted or unverifiable reference−¼★
- Copyeditonly when the manuscript needs a full edit−½★
The rating never drops below 1★, and a demonstrable critical failure (an impossible statistic, a proven ethics violation) caps it at 1★ on its own — so the stars can never look healthy when the verdict is CRITICAL.
The rating draws on a panel of agents. Three independent Kaimen Rigor reviewers grade the eight dimensions below across several independent passes (the shown verdict is their majority vote — steadier than any single run), isolate the paper’s major claims and check its own evidence backs them, and flag integrity concerns. Alongside them, a citation agent resolves every reference against Crossref, OpenAlex, and Retraction Watch; a statistics agent recomputes reported tests; and rule-based checks verify that declared data/code links actually resolve. Full text is required — an abstract-only submission is not analyzed.
Graded against NIH, MDAR, ARRIVE 2.0, CONSORT, EQUATOR, and RRID guidelines. A dimension that doesn’t apply to the study type is skipped, never penalized.
This Kaimen Rigor review is model-assisted and is not a substitute for formal expert review. It complements human evaluation by surfacing potential methodological concerns — verify each finding against the source.