Alpha¹ by KaimenConfidential

Kaimen Rigor review

Artificial intelligence guided screening for cardiomyopathies in an obstetric population: a pragmatic randomized clinical trial.

Adedinsewo DA, Morales-Lara AC, Afolabi BB, Kushimo OA, Mbakwem AC, Ibiyemi KF, Ogunmodede JA, Raji HO, Ringim SH, Habib AA, Hamza SM, Ogah OS, Obajimi G, Saanu OO, Jagun OE, Inofomoh FO, Adeolu T, Karaye KM, Gaya SA, Alfa I, Yohanna C, Venkatachalam KL, Dugan J, Yao X, Sledge HJ, Johnson PW, Wieczorek MA, Attia ZI, Phillips SD, Yamani MH, Tobah YB, Rose CH, Sharpe EE, Lopez-Jimenez F, Friedman PA, Noseworthy PA, Carter RE, SPEC-AI Nigeria Investigators

DOI: 10.1038/s41591-024-03243-9
Record issued: 2026-08-16
Engine: 7.39.0
Exported: 2026-09-22

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/5b722775-1751-437e-b1cc-c9e210475ddb is authoritative.

Kaimen Rigor review

Kaimen Rigor reviewed

2.75 / 5

Weak — serious problems found

Interventional Study

Critical

High-priority

How this rating was calculated

Started at 5★ — deductions

IntegrityIntegrity concern ×2−1★
ClaimsEfficacy rests on an unvalidated surrogate endpoint−0.5★
ClaimsTreatment effect not shown to be clinically meaningful−0.5★
ReportingData & code availability partially met−0.25★

Model-reported confidence: highReviewer agreement: highEngine: 7.39.0Source: full text

Biggest issues

01Efficacy rests on an unvalidated surrogate endpoint
highOverreachreviewer-assessedFull detail
The primary endpoint is detection of left ventricular systolic dysfunction (LVSD) defined as LVEF < 50% on echocardiography, which is a surrogate for clinical outcomes. The paper does not provide evidence linking LVSD detection to improved clinical outcomes, and target engagement at the tested dose is not applicable since it's a screening intervention. The claim of improved diagnosis is based on a surrogate biomarker without demonstrating that this leads to improved patient outcomes.
“The primary end point was identification of LVSD during the study period.”
Find in source
02Treatment effect not shown to be clinically meaningful
highOverreachreviewer-assessedFull detail
The primary effect is an increase in detection of LVSD from 2.0% to 4.1% (absolute difference 2.1 percentage points). While statistically significant, the clinical meaningfulness is not anchored to any established minimal clinically important difference or patient-centered outcome. The number needed to screen is 47, but the benefit of detecting these cases is not shown to translate into improved outcomes.
“Using the AI-enabled digital stethoscope, the primary study end point was met with detection of 24 out of 587 (4.1%) versus 12 out of 608 (2.0%) patients with LVSD (intervention versus control odds ratio 2.12, 95% CI 1.05–4.27; P = 0.032).”
Find in source

Findings by category

101ArithmeticNumerical inconsistencies 302OverreachOverstated conclusions 03AuthenticityNone found 104DisclosureReporting gaps 05AvailabilityNone found 06Copyedit6 minor suggestions

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 and transparently reported pragmatic randomized trial with a strong scientific premise, rigorous design, and appropriate statistical analysis. The main weakness is the vague data availability statement and lack of code sharing, which limits reproducibility.

Both reviewers independently scored all eight dimensions and agreed on every status; no divergence to reconcile. The statistics verification recomputed 6 tests (all consistent) but does not cover all reported statistics; the citation check found no retracted or unresolved references; the reproducibility check found both links live.

Numerical inconsistencies

1 finding · worst low

Values 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.

Internal contradictions in the reported numbersAssessed

Recomputed 6 tests: 6 consistent, 0 inconsistent; 5 recomputed directly from the reported test statistics, 1 via agent-written checks.

Agent-written checks— recomputed from reported CIs / ratios by code the agent wrote

CONSISTENTreported p = .032 · recomputed p = .036
Recomputed odds ratio 2.12 (95% CI 1.05–4.27), reported p=0.032
“odds ratio 2.12, 95% CI 1.05–4.27; P = 0.032”
Find in source
Taken as given: 1.05–4.27 is a two-sided 95% confidence interval for the odds ratio of 2.12, not a range, an IQR, or a different interval level; the odds ratio is a RATIO measure, so the interval is symmetric on the log scale; p=0.032 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(2.12, 1.05, 4.27, 1)
CONSISTENTreported p = .125 · recomputed p = .130
Recomputed odds ratio 1.75 (95% CI 0.85–3.62), reported p=0.125
“odds ratio 1.75, 95% CI 0.85–3.62; P = 0.125”
Find in source
Taken as given: 0.85–3.62 is a two-sided 95% confidence interval for the odds ratio of 1.75, not a range, an IQR, or a different interval level; the odds ratio is a RATIO measure, so the interval is symmetric on the log scale; p=0.125 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(1.75, 0.85, 3.62, 1)
CONSISTENTreported p = .227 · recomputed p = .232
Recomputed odds ratio 1.57 (95% CI 0.75–3.29), reported p=0.227
“odds ratio 1.57, 95% CI 0.75–3.29; P = 0.227”
Find in source
Taken as given: 0.75–3.29 is a two-sided 95% confidence interval for the odds ratio of 1.57, not a range, an IQR, or a different interval level; the odds ratio is a RATIO measure, so the interval is symmetric on the log scale; p=0.227 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(1.57, 0.75, 3.29, 1)
CONSISTENTreported p = .976 · recomputed p = 1.000
Recomputed OR 1.00 (95% CI 0.74–1.35), reported p=0.976
“OR 1.00, 95% CI 0.74–1.35 P = 0.976”
Find in source
Taken as given: 0.74–1.35 is a two-sided 95% confidence interval for the OR of 1.00, not a range, an IQR, or a different interval level; the OR is a RATIO measure, so the interval is symmetric on the log scale; p=0.976 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(1, 0.74, 1.35, 1)
CONSISTENTreported p = .621 · recomputed p = .639
Recomputed OR 1.10 (95% CI 0.74–1.64), reported p=0.621
“OR 1.10, 95% CI 0.74–1.64, P = 0.621”
Find in source
Taken as given: 0.74–1.64 is a two-sided 95% confidence interval for the OR of 1.10, not a range, an IQR, or a different interval level; the OR is a RATIO measure, so the interval is symmetric on the log scale; p=0.621 is the p for THIS estimate, not for another comparison in the same sentence
Method: back the two-tailed p out of the log-scale CI width and compare it against the printed p
How we recomputed it: pCI(1.1, 0.74, 1.64, 1)
CONSISTENTreported p = .026 · recomputed p = .026Reviewers 1, 2
All-cause mortality: HR 4.20, 95% CI 1.18-14.87, P=0.026
“All-cause mortality d | 12/587 | 3/608 | 4.20 (1.18, 14.87) | 0.026”
Find in source
Taken as given: The hazard ratio is 4.20 with 95% CI 1.18-14.87.; The CI is a 95% confidence interval for the hazard ratio.; The p-value is two-sided.
Method: Recomputed p-value from the reported hazard ratio and 95% CI using the normal approximation for the log hazard ratio.
How we recomputed it: pCI(4.20, 1.18, 14.87, 1)

lowinternal contradiction
The number of LVSD cases at study end for the digital stethoscope is 24, but the text says 'three additional LVSD cases were identified with two in the intervention arm and one in the control arm' after 22 at entry, which sums to 24, but the control arm total is 12, which is 11+1, consistent.
“At the study end, three additional LVSD cases were identified with two in the intervention arm and one in the control arm, resulting in a total of 24 cases (4.1%) of LVSD (LVEF < 50%) identified in the intervention arm compared to 12 (2.0%) in the control arm”
ResultsFind in source
lowinternal contradiction
The abstract reports 1,232 randomized, but the main text says 1,232 randomized and 1,195 completed baseline. The number 1,196 appears only in the summary paragraph.
“A total of 1,232 (616 in each arm) participants were randomized and 1,195 participants (587 intervention arm and 608 control arm) completed the baseline visit”
AbstractFind in source

Overstated conclusions

3 findings · worst high

Conclusions 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
Treatment effect not shown to be clinically meaningfulAssessed
Conclusions only partially backed by the presented evidenceAssessed

7 major claims checked against the paper's own evidence: 2 only partially supported (evidence backs part of the claim; gaps or caveats remain); the rest adequately supported.

partialReviewers 1, 2
AI-based electrocardiogram screening proved accurate in detecting cardiomyopathies.
The 12-lead AI-ECG did not reach statistical significance for the primary outcome (OR 1.75, P=0.125), but the diagnostic performance (AUC) was high.
Evidence: Primary outcome for 12-lead ECG: OR 1.75, 95% CI 0.85-3.62, P=0.125; AUC for LVEF<50% was 0.928.
“AI-based electrocardiogram screening proved accurate in detecting cardiomyopathies and suggests that it could improve detection of these conditions.”
AbstractFind in source
partialReviewer 1
The higher all-cause mortality in the intervention arm is likely due to differential ascertainment.
The paper offers plausible explanations but does not provide direct evidence to confirm differential ascertainment.
Evidence: Discussion of potential explanations: increased contact with healthcare services, mortality ascertainment prone to error.
Potential explanations or hypothesis for the observed higher mortality in the intervention arm include (1) a differential observation of mortality in the intervention arm due to increased contact with healthcare services...
Discussion ¶7reviewer’s wording
supportedReviewers 1, 2
AI-guided screening using a digital stethoscope improved the diagnosis of pregnancy-related cardiomyopathy.
The primary outcome was met with a statistically significant odds ratio of 2.12 (95% CI 1.05-4.27, P=0.032) for the digital stethoscope.
Evidence: Primary outcome analysis: 24/587 vs 12/608, OR 2.12, 95% CI 1.05-4.27, P=0.032.
“In pregnant and postpartum women, AI-guided screening using a digital stethoscope improved the diagnosis of pregnancy-related cardiomyopathy.”
AbstractFind in source
supportedReviewer 1
The study demonstrates a high prevalence of LVSD in an obstetric population in Nigeria.
The prevalence of LVSD in the intervention arm was 4.1% (24/587), which is higher than previously reported estimates.
Evidence: Primary outcome: 24/587 (4.1%) in intervention arm.
“We demonstrate a high prevalence of LVSD, in an obstetric population in Nigeria that supports the need for screening.”
Discussion ¶1Find in source
supportedReviewers 1, 2
The AI-enabled digital stethoscope detected cardiomyopathy with high sensitivity, specificity and negative predictive value.
The diagnostic performance metrics reported in Table 3 show high sensitivity (95.7%) and NPV (99.8%) for LVEF<50% using max prediction.
Evidence: Table 3: Max prediction for LVEF<50%: sensitivity 95.7%, specificity 82.0%, NPV 99.8%.
“We found that an AI-enabled digital stethoscope that analyzes single-lead ECG and phonocardiogram recordings, as well as an AI-enabled 12-lead ECG detected the presence of cardiomyopathy with high sensitivity, specificity and negative predictive value.”
Discussion ¶1Find in source
supportedReviewer 2
The NNS to detect one additional case of LVSD was 47.
The NNS is directly reported in the results.
Evidence: Results: 'The estimated number needed to screen (NNS) to detect one additional case of LVSD was 47.'
“The estimated number needed to screen (NNS) to detect one additional case of LVSD was 47.”
ResultsFind in source
supportedReviewer 2
AI-guided screening doubled the diagnosis of pregnancy-related cardiomyopathy compared to usual care.
The odds ratio of 2.12 indicates a doubling of detection, and the result was statistically significant.
Evidence: Primary outcome OR 2.12 (95% CI 1.05-4.27).
“AI-guided screening with a digital stethoscope doubled the diagnosis of pregnancy-related cardiomyopathy when compared to usual obstetric care”
DiscussionFind in source

Premise concern: surrogate not validated for clinical benefit; effect size not shown to be clinically meaningful.

INADEQUATESurrogate endpoint
The primary endpoint is detection of left ventricular systolic dysfunction (LVSD) defined as LVEF < 50% on echocardiography, which is a surrogate for clinical outcomes. The paper does not provide evidence linking LVSD detection to improved clinical outcomes, and target engagement at the tested dose is not applicable since it's a screening intervention. The claim of improved diagnosis is based on a surrogate biomarker without demonstrating that this leads to improved patient outcomes.
“The primary end point was identification of LVSD during the study period.”
Find in source
INADEQUATEEffect size
The primary effect is an increase in detection of LVSD from 2.0% to 4.1% (absolute difference 2.1 percentage points). While statistically significant, the clinical meaningfulness is not anchored to any established minimal clinically important difference or patient-centered outcome. The number needed to screen is 47, but the benefit of detecting these cases is not shown to translate into improved outcomes.
“Using the AI-enabled digital stethoscope, the primary study end point was met with detection of 24 out of 587 (4.1%) versus 12 out of 608 (2.0%) patients with LVSD (intervention versus control odds ratio 2.12, 95% CI 1.05–4.27; P = 0.032).”
Find in source

Data authenticity concerns

None found

An 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

1 finding · worst medium

Required detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.

Data/code availability incompleteAssessed

Eight Dimensions7 pass · 1 warn

The introduction cites multiple prior studies (retrospective and pilot prospective) showing AI-ECG effectiveness, and notes the high incidence of peripartum cardiomyopathy in Nigeria. The rationale for the trial is clearly stated: to determine whether AI-guided screening improves detection beyond standard care. Limitations of prior work (e.g., small pilot samples, retrospective designs) are implicitly addressed by conducting a large prospective randomized trial.

Sub-criteria

prior work citedADEQUATE

premise rationaleADEQUATE

limitations addressedADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

The paper establishes a strong scientific premise by citing prior research on AI-ECG for LVSD detection and the high burden of peripartum cardiomyopathy in Nigeria, and clearly links this to the study objective.

Evidence

“A retrospective study (area under the curve (AUC) = 0.89) and a pilot prospective study among pregnant and postpartum women in the United States showed AI-based screening to be effective (AUC = 1.00 using a 12-lead ECG and 0.98 using a digital stethoscope) in identifying pregnancy-related LVSD with LVEF < 45%.”

Introduction ¶2Find in source

“To address this question, we conducted an open-label, randomized, pragmatic clinical trial among pregnant and postpartum women to evaluate whether AI-guided screening (using a digital stethoscope and 12-lead ECG) improves the diagnosis of pregnancy-related LVSD in an obstetric population in Nigeria compared to usual care.”

Introduction ¶3Find in source

PASSReviewer 2· 90% conf

The paper establishes a clear scientific premise with extensive prior work and a logical rationale for the study.

Evidence

Introduction ¶2Find in source

“it remains unknown whether AI-guided screening improves cardiomyopathy detection in obstetric patients beyond the current standard of care.”

Introduction ¶3Find in source

“we conducted an open-label, randomized, pragmatic clinical trial among pregnant and postpartum women to evaluate whether AI-guided screening (using a digital stethoscope and 12-lead ECG) improves the diagnosis of pregnancy-related LVSD in an obstetric population in Nigeria compared to usual care.”

Introduction ¶3Find in source

Randomization used dynamic minimization with site stratification via a web-based application. The unit of randomization is the individual participant. A power analysis is reported (848 women for 80% power, increased to 1,200). Inclusion/exclusion criteria are clearly listed. The mITT analysis set is defined, and an ITT sensitivity analysis is performed. Blinding is not applicable as the trial is open-label; this is stated in the title and methods. Outlier handling is addressed through the mITT definition and conservative assumptions for poor-quality AI predictions.

Sub-criteria

controlsN/A

power analysisADEQUATE

blinding levelsADEQUATE

outlier handlingADEQUATE

randomization unitADEQUATE

inclusion exclusionADEQUATE

randomization methodADEQUATE

replicate distinctionN/A

independent replicationN/A

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

The trial design is rigorous: randomization method and unit are described, power analysis is provided, inclusion/exclusion criteria are pre-specified, and the analysis population (mITT) is defined. Blinding is not applicable due to the open-label pragmatic design, which is stated.

Evidence

“Randomization was performed in real time using dynamic minimization with the study site as a stratification factor, through a web-based application (iMedidata).”

MethodsFind in source

paraphrase[Methods, Statistical analysis]

“We assumed that the prevalence of LVEF < 50% would be 4% in the intervention arm and 1% in the control group, which yielded an estimate of 848 (424 per group) women to achieve 80% power at an α of 0.05.”

paraphrase[Methods, Study participants]

“Inclusion criteria were female, aged 18–49 years, pregnant or within 12 months postpartum, and currently receiving obstetric care... Exclusion criteria were complex congenital heart disease... and inability to provide informed consent.”

PASSReviewer 2· 90% conf

The trial design is rigorous with appropriate randomization, prespecified analysis populations, and a power calculation.

Evidence

“Randomization was performed in real time using dynamic minimization with the study site as a stratification factor, through a web-based application (iMedidata).”

MethodsFind in source

paraphrase[Methods, Statistical analysis]

paraphrase[Methods, Study participants]

“Inclusion criteria were female, aged 18–49 years, pregnant or within 12 months postpartum, and currently receiving obstetric care ... Exclusion criteria were complex congenital heart disease ... and inability to provide informed consent.”

The study reports age, race (all Black), ethnicity, weight, height, blood pressure, heart rate, hemoglobin, and comorbidities such as hypertensive disorders. Sex is inherently female due to the obstetric population, and the study is single-sex by design, which is justified by the condition being pregnancy-related. Demographics are comprehensive.

Sub-criteria

demographicsADEQUATE

sex reportedADEQUATE

sex justifiedADEQUATE

age weight healthADEQUATE

housing conditionsN/A

species strain sourceN/A

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

Demographic and clinical characteristics are well reported, including age, sex (all female), ethnicity, pregnancy status, and comorbidities. Sex is reported and justified by the study population (pregnant/postpartum women).

Evidence

“Age, years | 1,195 | 31 (27–35) | 31 (26–35)”

Table 1Find in source

“Inclusion criteria were female, aged 18–49 years, pregnant or within 12 months postpartum”

MethodsFind in source

PASSReviewer 2· 90% conf

Demographic and clinical characteristics are well reported, including age, ethnicity, pregnancy status, and comorbidities.

Evidence

“The median age was 31 years and all women identified as Black. Fifty-five percent were of Yoruba ethnicity, 28% Hausa, 11% Igbo and 6% were from other ethnic groups.”

Table 1Find in source

“At baseline, 150 (12.5%) women had been diagnosed with a hypertensive disorder of pregnancy (which includes chronic hypertension, gestational hypertension, pre-eclampsia and eclampsia) during the index pregnancy.”

ResultsFind in source

The methods state that the study was approved by the Mayo Clinic institutional review board and local ethics research committees at all participating sites. Informed consent is described as written or oral, in accordance with local approvals. Regulatory compliance is implied through adherence to local ethics committee approvals and the Declaration of Helsinki is not explicitly named, but the approval statements are sufficient.

Sub-criteria

iacuc statementN/A

informed consentADEQUATE

irb ethics statementADEQUATE

regulatory complianceADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

Ethical approvals are adequately reported: the study was approved by the Mayo Clinic IRB and local ethics committees in Nigeria, and informed consent was obtained from all participants.

Evidence

“The study was approved by the Mayo Clinic institutional review board as well as local ethics research committees at all participating sites in Nigeria.”

MethodsFind in source

“All study participants provided written or oral informed consent in accordance with local ethics research committee approvals.”

MethodsFind in source

PASSReviewer 2· 90% conf

Ethical approval and informed consent are adequately documented.

Evidence

“The study was approved by the Mayo Clinic institutional review board as well as local ethics research committees at all participating sites in Nigeria.”

MethodsFind in source

“All study participants provided written or oral informed consent in accordance with local ethics research committee approvals.”

MethodsFind in source

The digital stethoscope (Eko DUO) and 12-lead ECG machine (GE Marquette 2000) are identified. The AI algorithms are named (US FDA-cleared 12-lead AI-ECG, Mayo Clinic model, digital stethoscope model) with version numbers (e.g., lvef_v2.2.0, ELEFT 7.2.0). Statistical software R v.4.1.2 is identified. Since this is a device trial, antibodies, cell lines, mycoplasma, and organisms are not applicable.

Sub-criteria

mycoplasma testingN/A

reagents identifiedADEQUATE

organisms identifiedN/A

antibodies identifiedN/A

cell line authenticationN/A

software tools identifiedADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

The investigational products (AI algorithms and devices) are identified with manufacturer and version, and the software used for analysis is named. No wet-lab reagents are used.

Evidence

paraphrase[Methods, Trial procedures and AI algorithms]

“The 12-lead ECGs were acquired in a standard fashion... using a GE Marquette 2000 ECG machine (GE Healthcare).”

“Statistical analyses were performed using R v.4.1.2.”

Statistical analysisFind in source

PASSReviewer 2· 90% conf

The investigational products (AI algorithms and devices) are identified with sources and versions.

Evidence

“The US FDA-cleared model had an in-built ECG data quality check. As such ECGs deemed to be of insufficient quality did not have AI predictions generated (algorithm version lvef_v2.2.0).”

MethodsFind in source

“The 12-lead ECGs were acquired in a standard fashion as with routine clinical care, in a supine or semi-recumbent position for 10 s on standard ECG paper, at a sampling rate of 500 Hz using a GE Marquette 2000 ECG machine (GE Healthcare).”

MethodsFind in source

The primary analysis uses logistic regression with Pearson chi-squared test, and p-values are reported exactly (e.g., P = 0.032). Effect sizes are reported as odds ratios with 95% CIs. Assumptions are handled by design (logistic regression, Cox for mortality). Data presentation includes CONSORT diagram, forest plots, and tables with per-group n. Mathematical plausibility checks: the reported percentages and counts appear consistent (e.g., 24/587 = 4.1%, 12/608 = 2.0%).

Sub-criteria

tests namedADEQUATE

exact p valuesADEQUATE

effect sizes ciADEQUATE

data presentationADEQUATE

software identifiedADEQUATE

assumptions verifiedADEQUATE

mathematical plausibilityPASS

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

Statistical methods are appropriate and well-reported: tests are named, effect sizes with CIs are provided, software is identified, and data presentation follows clinical trial norms. Exact p-values are reported for primary outcomes.

Evidence

“The odds ratio and 95% large sample CI was estimated using a logistic regression model and statistical significance was assessed with a Pearson chi-squared test at the α = 0.05 level of significance (two-sided).”

Table 2Find in source

“intervention versus control odds ratio 2.12, 95% CI 1.05–4.27; P = 0.032”

AbstractFind in source

PASSReviewer 2· 90% conf

Statistical methods are appropriate and well described, with exact p-values and effect sizes with CIs.

Evidence

“The odds ratio and 95% large sample CI was estimated using a logistic regression model and statistical significance was assessed with a Pearson chi-squared test.”

Statistical analysisFind in source

“odds ratio 2.12, 95% CI 1.05–4.27; P = 0.032”

ResultsFind in source

“Statistical analyses were performed using R v.4.1.2.”

Statistical analysisFind in source

The data availability statement says data can be made available upon request with an analysis plan, but it does not specify a repository or a managed-access platform. It mentions a Material Transfer Agreement and a 1-month response timeframe, which is somewhat concrete but still relies on contacting the corresponding author. Code is not shared because it is proprietary, which is understandable but limits reproducibility.

Sub-criteria

code sharingINADEQUATE

accession numbersN/A

repository depositN/A

data availability statementINADEQUATE

Per reviewer (2) — the votes behind the verdict

WARNReviewer 1· 90% conf

A data availability statement is present but vague, and code is not shared due to proprietary reasons. The data access mechanism is described but lacks a clear platform or independent review process.

Evidence

“The underlying data supporting the findings of this study can be made available to clinical investigators and researchers upon request. Written requests for data sharing including an analysis plan will be required before approval.”

Data availabilityFind in source

“The code itself cannot be shared because it is proprietary intellectual property that has been licensed to Anumana and Eko Health.”

Data availabilityFind in source

WARNReviewer 2· 90% conf

Data availability is stated but with a vague access mechanism; code is not shared due to proprietary reasons.

Evidence

Data availabilityFind in source

“The code itself cannot be shared because it is proprietary intellectual property that has been licensed to Anumana and Eko Health.”

Data availabilityFind in source

The trial is registered (NCT05438576). The CONSORT-AI extension is mentioned. All prespecified outcomes are reported, including exploratory ones. Limitations are thoroughly discussed, including referral bias, attrition, and the definition of LVSD. Conclusions are proportional, acknowledging the higher mortality in the intervention arm and the need for further study. Funding sources and competing interests are disclosed.

Sub-criteria

funding coiADEQUATE

trial registrationADEQUATE

reporting guidelineADEQUATE

methods completenessADEQUATE

all outcomes reportedADEQUATE

limitations discussedADEQUATE

conclusions proportionalADEQUATE

Per reviewer (2) — the votes behind the verdict

PASSReviewer 1· 90% conf

The manuscript is transparent: methods are detailed, the trial is registered, CONSORT-AI is referenced, all outcomes are reported, limitations are discussed, and funding/COI are disclosed.

Evidence

“ClinicalTrials.gov registration: NCT05438576”

AbstractFind in source

“The CONSORT-AI Extension guideline was utilized in reporting study design and results.”

MethodsFind in source

“The key limitations of this study were introduced by the pragmatic clinical trial design and enrolling study participants at teaching hospitals with a licensed cardiologist and echocardiography capabilities.”

Discussion ¶8Find in source

PASSReviewer 2· 90% conf

The manuscript is transparent with trial registration, reporting guideline adherence, and discussion of limitations.

Evidence

“ClinicalTrials.gov registration: NCT05438576”

AbstractFind in source

“The CONSORT-AI Extension guideline was utilized in reporting study design and results.”

MethodsFind in source

Discussion ¶8Find in source

Registered (1 ID: ClinicalTrials.gov). Reporting guidelines cited: CONSORT, STARD.

ClinicalTrials.gov

NCT05438576

Reporting guidelines: CONSORT, STARD

Broken references and links

None found

References 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 38 references by DOI: 3 verified — 35 no DOI (shown, not verified).

NO DOI
Epidemiology of peripartum cardiomyopathy: incidence, predictors, and outcomes
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Clinical features and outcomes of peripartum cardiomyopathy in nigeria
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Incidence, clinical characteristics, and risk factors of peripartum cardiomyopathy in Nigeria: results from the PEACE Registry
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Worldwide incidence of peripartum cardiomyopathy and overall maternal mortality
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Pregnancy-related cardiovascular deaths in California: beyond peripartum cardiomyopathy
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Peripartum cardiomyopathy: from genetics to management
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Team-based care of women with cardiovascular disease from pre-conception through pregnancy and postpartum: JACC Focus Seminar 1/5
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
An artificial intelligence-enabled ECG algorithm for the identification of patients with atrial fibrillation during sinus rhythm: a retrospective analysis of outcome prediction
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Detection of hypertrophic cardiomyopathy using a convolutional neural network-enabled electrocardiogram
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Electrocardiogram screening for aortic valve stenosis using artificial intelligence
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Deep learning electrocardiographic analysis for detection of left-sided valvular heart disease
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Screening for cardiac contractile dysfunction using an artificial intelligence-enabled electrocardiogram
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Artificial intelligence-enabled electrocardiograms for identification of patients with low ejection fraction: a pragmatic, randomized clinical trial
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Detecting cardiomyopathies in pregnancy and the postpartum period with an electrocardiogram-based deep learning model
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Artificial intelligence based screening for cardiomyopathy in an obstetric population: a pilot study
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Development and validation of an electrocardiographic artificial intelligence model for detection of peripartum cardiomyopathy
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
An artificial intelligence electrocardiogram analysis for detecting cardiomyopathy in the peripartum period
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Point-of-care screening for heart failure with reduced ejection fraction using artificial intelligence during ECG-enabled stethoscope examination in London, UK: a prospective, observational, multicentre study
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Automated detection of low ejection fraction from a one-lead electrocardiogram: application of an AI algorithm to an electrocardiogram-enabled digital stethoscope
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Cardiovascular diseases in nigeria: current status, threats, and opportunities
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
First trimester preeclampsia screening and prediction
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Screening for depression among the general adult population and in women during pregnancy or the first-year postpartum: two systematic reviews to inform a guideline of the canadian task force on preventive health care
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
The mathematical limitations of fetal echocardiography as a screening tool in the setting of a normal second-trimester ultrasound
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Effect of second-trimester sonographic cervical length on the risk of spontaneous preterm delivery in different risk groups: a prospective observational multicenter study
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
The Mobile Economy: Sub Saharan Africa 2022
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Peripartum cardiomyopathy: JACC state-of-the-art review
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Vital signs: pregnancy-related deaths, United States, 2011–2015, and strategies for prevention, 13 states, 2013–2017
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Racial and ethnic disparities in maternal mortality in the united states using enhanced vital records, 2016‒2017
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Long-term outcomes of women with peripartum cardiomyopathy having subsequent pregnancies
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Consequences of maternal mortality on infant and child survival: a 25-year longitudinal analysis in Butajira Ethiopia (1987–2011)
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Clinical trials overview: from explanatory to pragmatic clinical trials
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Eko Low Ejection Fraction Tool (ELEFT): Reduced Ejection Fraction Machine Learning-based Notification Software
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Screening for peripartum cardiomyopathies using artificial intelligence in Nigeria (SPEC-AI Nigeria): clinical trial rationale and design
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
NO DOI
Age and sex estimation using artificial intelligence from standard 12-lead ECGs
No DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).

2 data/code links checked; 2 live.

datahttps://clinicaltrials.gov/study/NCT05438576LIVEHTTP 200
Resolves, but the content could not be matched to the paper.
datahttps://clinicaltrials.gov/ct2/show/NCT05438576LIVEHTTP 200
Resolves, but the content could not be matched to the paper.

Copyediting

6 minor

Wording, consistency and formatting errors that need correcting before submission.

No major wording or formatting errors. 6 minor suggestions below.

6 copyedit issues flagged: mostly consistency, clarity, typo.

MINORconsistencyAbstract
“1,232 (616 in each arm) participants were randomized and 1,195 participants (587 intervention arm and 608 control arm) completed the baseline visit”
→ Ensure the numbers are consistent throughout the paper.
The abstract states 1,232 randomized, but the main text says 1,232 randomized and 1,195 completed baseline. This is consistent, but the abstract also mentions 1,196 in the summary paragraph, which is inconsistent.
MINORconsistencyAbstract, summary paragraph
“In this pragmatic, randomized clinical trial involving 1,196 pregnant and postpartum women”
→ Change to 1,195 to match the mITT analysis set.
The number 1,196 appears only in the summary paragraph and conflicts with the 1,195 reported elsewhere.
MINORclarityMethods, Statistical analysis
“The estimates were rounded up to 500 per group to account for uncertainties in the calculations.”
→ Clarify that the sample size was increased to 500 per group, and later to 1,200 total.
The sentence is slightly confusing because it says 'rounded up to 500 per group' but the total was later increased to 1,200.
MINORconsistencyAbstract
“1,232 (616 in each arm) participants were randomized and 1,195 participants (587 intervention arm and 608 control arm) completed the baseline visit”
→ Ensure consistent use of 'intervention arm' vs 'intervention group' throughout.
Minor inconsistency in terminology.
MINORtypoResults, Sample characteristics
“1,195 (587 in the intervention arm and 608 in the control arm) completed baseline assessments”
→ Check for duplicate '1,195' in text.
Potential duplication.
MINORclarityMethods, Statistical analysis
“The estimates were rounded up to 500 per group to account for uncertainties in the calculations.”
→ Clarify that the sample size was increased to 500 per group, then later to 600 per group.
Could be clearer.

Kaimen Rigor Action ItemsFinal pass before journal submission

The published work is robust and well-reported; an informed reader should weigh the limited data/code availability and the minor internal inconsistency in the participant count (1,196 vs 1,195) as the main caveats. No erratum is warranted for the core findings, but the authors should consider clarifying the data access mechanism and correcting the count discrepancy.

1.
HIGHdata code
In the Data Availability section, specify a concrete managed-access platform (e.g., Vivli or YODA) or a named data access committee with clear conditions and a defined timeline for response.
The current statement relies on contacting the corresponding author and lacks a transparent, independent review process, which limits reproducibility.
2.
HIGHdata code
Consider depositing de-identified aggregate data or summary statistics in a public repository (e.g., Figshare or Zenodo) to enhance transparency while protecting patient privacy.
Even if raw data cannot be shared, aggregate data would improve the reproducibility of the reported results.
3.
HIGHreporting
Correct the participant count inconsistency in the Abstract summary paragraph: change '1,196' to '1,195' to match the mITT analysis set reported elsewhere.
The number 1,196 appears only in the summary paragraph and conflicts with the 1,195 reported in the abstract and main text, which could confuse readers.
4.
MEDIUMreporting
Clarify the sample size calculation in Methods, Statistical analysis: state that the estimate was rounded up to 500 per group and later increased to 600 per group (1,200 total) to account for uncertainties.
The current sentence 'rounded up to 500 per group' is confusing given the final sample size of 1,200.
5.
MEDIUMreporting
Ensure consistent terminology throughout the manuscript: use either 'intervention arm' or 'intervention group' consistently.
Minor inconsistency in terminology could distract readers and reduce clarity.
6.
MEDIUMreporting
In the CONSORT flow, report the number of participants screened but not enrolled, and the number who withdrew consent or were excluded for each reason.
Improves transparency and completeness of the participant flow, as recommended by CONSORT.
7.
MEDIUMreporting
Clarify the role of the data safety monitoring board (or lack thereof) and whether any interim analyses were performed.
Readers need to know how safety was monitored in this open-label trial.
8.
MEDIUMreporting
Provide more detail on the handling of missing data for secondary outcomes and any sensitivity analyses performed.
Missing data handling is a common reviewer concern and affects the robustness of secondary findings.
9.
MEDIUMreporting
Clarify the definition of 'clinical recognition' in the control arm and how it was ascertained to reduce potential bias.
The primary outcome depends on this definition, and ambiguity could affect interpretation.
10.
MEDIUMreporting
Report the intraclass correlation coefficient for echocardiogram readings between local and central review.
Provides evidence of the reliability of the outcome measurement.
11.
LOWreporting
Specify the exact version of the CONSORT-AI checklist used and where it is available.
Enhances transparency and allows readers to verify adherence to the reporting guideline.
12.
LOWreporting
Include a statement on whether any adverse events were related to the AI screening itself.
Addresses a potential safety concern specific to the intervention.
13.
LOWdata code
Consider sharing the statistical analysis code (even if the AI algorithms are proprietary) to improve reproducibility.
The statistical code is not proprietary and sharing it would allow independent verification of the analyses.

Full text (77,246 chars · analyzed ~97%, head + tail)Analyzed: Aug 16, 2026

RigorReviewer 1

RigorReviewer 2

RigorPre-Submission Reviewer

CopyeditAgent Scribe

CitationsCitation Integrity

StatisticsStatistics Consistency

ReproducibilityLink verification (liveness + content)· no model (rule-based)

PreregistrationDeterministic check· no model (rule-based)

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.”

How stars are deducted

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.

The eight reporting dimensions

Scientific PremisePrior work, rationale, and how it addresses gaps.

Study Design & RigorRandomization, blinding, power, controls, replication.

Biological VariablesSex, age, species/strain, housing, demographics.

Ethical ApprovalsIRB/IACUC approval, informed consent, regulatory compliance.

Key ResourcesAntibodies, cell lines, organisms, software — identified with persistent IDs.

Statistical ReportingNamed tests, exact p-values, effect sizes, plausible summary stats.

Data & Code AvailabilityAvailability statements, repositories, accession numbers, shared code.

Reporting TransparencyMethods, registration, guidelines, outcomes, limitations, funding/COI.

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.

Artificial intelligence guided screening for cardiomyopathies in an obstetric population: a pragmatic randomized clinical trial.

DOI

10.1038/s41591-024-03243-9

Record issued

2026-08-16

Engine

7.39.0

Exported

2026-09-22