A multicenter pragmatic implementation study of AI-ECG-based clinical decision support software to identify low LVEF: Clinical trial design and methods
Open accessAbstract
Background: Artificial intelligence (AI) enabled algorithms can detect or predict cardiovascular conditions using electrocardiogram (ECG) data. Clinical studies have evaluated ECG-AI algorithms, including a recent single-center study which evaluated outcomes when clinicians were provided with ECG-AI results. A Multicenter Pragmatic IMplementation Study of ECG-AI-Based Clinical Decision Support Software to Identify Low LVEF (AIM ECG-AI) will evaluate clinical impacts of clinical decision support software (CDSS) integrated within the electronic health record (EHR) to provide point-of-care ECG-AI results to clinicians during routine outpatient care. Methods: AIM ECG-AI is a multicenter, cluster-randomized trial recruiting and randomizing clinicians to receive access to the CDSS (intervention) or provide usual care. Clinicians are recruited from 5 geographically distinct health systems and clustered at the care team level. AIM ECG-AI will evaluate clinical care provided during >32,000 eligible clinical encounters with adult patients with no history of low LVEF and who have a digital ECG documented within the health system's EHR, with 90 day follow up. Results: Study data includes clinician surveys, study software metrics, and EHR data as a read-out for clinician decision-making. AIM ECG-AI will evaluate detection of left ventricular ejection fraction ≤40 % by echocardiography, with exploratory endpoints. Subgroup analyses will evaluate the health system, clinician, and patient-level characteristics associated with outcomes (NCT05867407). Conclusion: AIM ECG-AI is the first multisite clinical evaluation of an EHR-integrated, point-of-care CDSS to provide ECG-AI results in the clinical workflow. The findings will provide valuable insights for clinically focused software design to bring AI into routine clinical practice.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
A safety-net model may be useful when positive AI-ECG results enrich echocardiography yield and lead to timely confirmatory evaluation.
“A safety-net model may be useful when positive AI-ECG results enrich echocardiography yield and lead to timely confirmatory evaluation.”
From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [38]
The Discussion (page not provided) reports that providing AI-ECG results in the EAGLE study increased new low-LVEF diagnoses despite similar echocardiogram-ordering rates, which is consistent with improved diagnostic yield. The Methods > Endpoints section describes echocardiographic evaluation within 90 days, but this design paper does not establish that positive results led to timely confirmatory evaluation or assess a safety-net model. (LLM, conf=0.83)
Experimental — accuracy not yet measuredLow-LVEF detection is the most mature use case, having already progressed from model development to randomized trials and early implementation studies.
“Low-LVEF detection is the most mature use case, having already progressed from model development to randomized trials and early implementation studies.”
From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [38]
The Introduction describes retrospective and prospective low-LVEF detection studies and a randomized clinical trial. The Discussion describes the earlier EAGLE trial and this paper’s planned multicenter implementation study. These passages support the progression described, but the paper does not compare low-LVEF detection with other AI-ECG use cases to establish that it is the most mature. (LLM, conf=0.86)
Experimental — accuracy not yet measuredGatekeeper use carries a different clinical risk profile because false-negative or poorly calibrated low-risk outputs may delay diagnosis.
“By contrast, gatekeeper use carries a different clinical risk profile because false-negative or poorly calibrated low-risk outputs may delay diagnosis.”
From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [38]
The Methods > Intervention section reports 84.5% sensitivity for the low-LVEF algorithm and says clinicians receive alerts for actionable results such as positive findings. This supports the possibility of false negatives, but the paper does not describe gatekeeper use, poorly calibrated low-risk outputs, or delayed diagnosis resulting from them. (LLM, conf=0.78)
Experimental — accuracy not yet measuredThe regulatory maturity of AI-ECG for SHD screening remains uneven.
“The regulatory maturity of AI-ECG for SHD screening remains uneven.”
From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [38]
The Methods (Intervention; page unavailable) state that one AI-ECG algorithm for detecting low ejection fraction received FDA clearance in 2023. The Introduction (page unavailable) describes limited real-world randomized validation of cardiovascular AI, but neither passage compares regulatory status across AI-ECG tools for structural heart disease screening. The claim that regulatory maturity is uneven is therefore not established by this paper. (LLM, conf=0.76)
Experimental — accuracy not yet measuredThe main failure modes of safety-net AI-ECG deployment are excess downstream testing, incomplete follow-up, and alert fatigue.
“Its main failure modes are excess downstream testing, incomplete follow-up, and alert fatigue.”
From AI-ECG for Echocardiography Triage in Structural Heart Disease: Evidence, Implementation, and Future Directions (2026) · cites this paper as [38]
The Methods (Data Collection and Endpoints; page not provided) describe tracking echocardiogram orders, dismissed notifications, and clinical outcomes. The Discussion (page not provided) identifies low software adoption as a possible limitation, but our copy does not identify excess downstream testing, incomplete follow-up, or alert fatigue as the main failure modes of safety-net deployment. Searched for those terms and related descriptions of unnecessary testing and missed follow-up. (LLM, conf=0.72)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure
Heidenreich, P. A., Bozkurt, B., Aguilar, D., Allen, L. A., Byun, J. J., Colvin, M. M., Deswal, A., Drazner, M. H., Dunlay, S. M., Evers, L. R., et al. (2022) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2021.12.012 · Cited by 2,862 · see citing works
Martin, S. S., Aday, A. W., Almarzooq, Z. I., Anderson, C. A. M., Arora, P., Avery, C. L., Baker-Smith, C. M., Barone Gibbs, B., Beaton, A. Z., Boehme, A. K., et al. (2024) · Circulation · doi:10.1161/cir.0000000000001209 · Cited by 2,800 · see citing works
Attia, Z. I., Kapa, S., Lopez-Jimenez, F., McKie, P. M., Ladewig, D. J., Satam, G., Pellikka, P. A., Enriquez-Sarano, M., Noseworthy, P. A., Munger, T. M., et al. (2019) · Nature Medicine · doi:10.1038/s41591-018-0240-2 · Cited by 1,389 · see citing works
Self-citationYao, X., Rushlow, D. R., Inselman, J. W., McCoy, R. G., Thacher, T. D., Behnken, E. M., Bernard, M. E., Rosas, S. L., Akfaly, A., Misra, A., et al. (2021) · Nature Medicine · doi:10.1038/s41591-021-01335-4 · Cited by 424 · see citing works
Self-citationMaddox, K. E. J., Elkind, M. S. V., Aparicio, H. J., Commodore-Mensah, Y. Y., de Ferranti, S. D., Dowd, W. N., Hernandez, A. F., Khavjou, O., Michos, E. D., Palaniappan, L. P., et al. (2024) · Circulation · doi:10.1161/cir.0000000000001256 · Cited by 397 · see citing works
Joshi, G., Jain, A., Araveeti, S. R., Adhikari, S., Garg, H., & Bhandari, M. (2024) · Electronics · doi:10.3390/electronics13030498 · Cited by 294 · see citing works
Attia, Z. I., Kapa, S., Yao, X., López-Jiménez, F. A., Mohan, T. L., Pellikka, P. A., Carter, R. E., Shah, N. D., Friedman, P. A., & Noseworthy, P. A. (2019) · Journal of Cardiovascular Electrophysiology · doi:10.1111/jce.13889 · Cited by 195 · see citing works
Self-citationAdedinsewo, D., Carter, R. E., Attia, Z., Johnson, P., Kashou, A. H., Dugan, J. L., Albus, M., Sheele, J. M., Bellolio, F., Friedman, P. A., et al. (2020) · Circulation: Arrhythmia and Electrophysiology · doi:10.1161/circep.120.008437 · Cited by 169 · see citing works
Self-citationMuehlematter, U. J., Blüthgen, C., & Vokinger, K. N. (2023) · The Lancet Digital Health · doi:10.1016/s2589-7500(23)00126-7 · Cited by 113 · see citing works
Yao, X., McCoy, R. G., Friedman, P. A., Shah, N. D., Barry, B. A., Behnken, E. M., Inselman, J. W., Attia, Z. I., & Noseworthy, P. A. (2020) · American Heart Journal · doi:10.1016/j.ahj.2019.10.007 · Cited by 98 · see citing works
Self-citationExternal validation of a deep learning electrocardiogram algorithm to detect ventricular dysfunction
Attia, I. Z., Tseng, A. S., Benavente, E. D., Medina‐Inojosa, J. R., Clark, T. G., Malyutina, S. K., Kapa, S., Schirmer, H., Kudryavtsev, A. V., Noseworthy, P. A., et al. (2021) · International Journal of Cardiology · doi:10.1016/j.ijcard.2020.12.065 · Cited by 82 · see citing works
Self-citationCluster-Randomized Trial Comparing Ambulatory Decision Support Tools to Improve Heart Failure Care
Mukhopadhyay, A., Reynolds, H. R., Phillips, L. M., Nagler, A. R., King, W. C., Szerencsy, A. C., Saxena, A., Aminian, R., Klapheke, N., Horwitz, L. I., et al. (2023) · Journal of the American College of Cardiology · doi:10.1016/j.jacc.2023.02.005 · Cited by 79 · see citing works
Kashou, A. H., Medina-Inojosa, J. R., Noseworthy, P. A., Rodeheffer, R. J., Lopez-Jimenez, F., Attia, I. Z., Kapa, S., Scott, C. G., Lee, A. T., Friedman, P. A., et al. (2021) · Mayo Clinic Proceedings · doi:10.1016/j.mayocp.2021.02.029 · Cited by 66 · see citing works
Self-citationArtificial Intelligence-Enabled Electrocardiography to Screen Patients with Dilated Cardiomyopathy
Shrivastava, S., Cohen, M. S., Attia, Z. I., Rosenbaum, A. N., Wang, L., Giudicessi, J. R., Redfield, M. M., Bailey, K. R., López-Jiménez, F. A., Lin, G., et al. (2021) · The American Journal of Cardiology · doi:10.1016/j.amjcard.2021.06.021 · Cited by 47 · see citing works
Self-citationAdedinsewo, D. A., Johnson, P. W., Douglass, E. J., Attia, I. Z., Phillips, S. D., GOSWAMI, R. M., Yamani, M. H., Connolly, H. M., Rose, C. H., Sharpe, E. E., et al. (2021) · European Heart Journal - Digital Health · doi:10.1093/ehjdh/ztab078 · Cited by 45 · see citing works
Self-citationRushlow, D. R., Croghan, I. T., Inselman, J. W., Thacher, T. D., Friedman, P. A., Yao, X., Pellikka, P. A., Lopez-Jimenez, F., Bernard, M. E., Barry, B. A., et al. (2022) · Mayo Clinic Proceedings · doi:10.1016/j.mayocp.2022.04.008 · Cited by 33 · cited 1× · see citing works
“To account for those Intervention group participants who do not interact with the study software, subgroup analyses are planned to evaluate clinical behaviors and decision making among low-adopters and high-adopters, as was previously described [14].”
Self-citationJentzer, J. C., Kashou, A. H., López-Jiménez, F. A., Attia, Z. I., Kapa, S., Friedman, P. A., & Noseworthy, P. A. (2020) · European Heart Journal Acute Cardiovascular Care · doi:10.1093/ehjacc/zuaa021 · Cited by 31 · see citing works
Self-citationBrito, B. O. D. F., Attia, Z. I., Martins, L. N. A., Perel, P., Nunes, M. C. P., Sabino, É. C., Cardoso, C. S., Ferreira, A. M., Gomes, P. R., Ribeiro, A. L. P., et al. (2021) · PLoS neglected tropical diseases · doi:10.1371/journal.pntd.0009974 · Cited by 26 · see citing works
Self-citationKlein, C. J., Ӧzcan, I., Attia, Z. I., Cohen, M. S., Lerman, A., Medina‐Inojosa, J. R., López-Jiménez, F. A., Friedman, P. A., Milone, M., & Shelly, S. (2022) · Mayo Clinic Proceedings Innovations Quality & Outcomes · doi:10.1016/j.mayocpiqo.2022.08.003 · Cited by 11 · see citing works
Self-citation
Flagged first, then by citations
Cited by
Sorted by citations
Artificial Intelligence and ECG: A New Frontier in Cardiac Diagnostics and Prevention
Bartusik‐Aebisher, D., Rogóż, K., & Aebisher, D. (2025) · Biomedicines · doi:10.3390/biomedicines13071685 · Cited by 40 · see citing works
Varzideh, F., Pande, S., Jankauskas, S. S., Mone, P., Kansakar, U., & Santulli, G. (2026) · Frontiers in Neuroscience · doi:10.3389/fnins.2026.1788653 · Cited by 4 · see citing works
Khalifa, A., Abdulaziz, M., Parvez, S. S., Salem, A. M., Panhwer, H. S., Saleem, U., Eldeeb, M., Weld-Ali, T., & Mosaad, M. (2025) · Cureus · doi:10.7759/cureus.95059 · Cited by 1 · see citing works
Tang, Q., Deng, K., Cui, Y., Liu, H., & Qian, B. (2026) · International Journal of General Medicine · doi:10.2147/ijgm.s610298
“13,14,37,38 For safety-net use, model performance should therefore be judged by the balance among sensitivity, positive predictive value, referral burden, and follow-up completion, rather than by discrimination alone.”
InfluentialCroon, P. M., Dhingra, L. S., Pedroso, A. F., & Khera, R. (2026) · Current Heart Failure Reports · doi:10.1007/s11897-026-00764-x · Cited by 1 · see citing works
Ruan, H., Lee, J.-K., Xiong, J., Zou, L.-J., & Li, S.-S. (2026) · Digital Health · doi:10.1177/20552076261481028
- Authors
- Lopez-Jimenez, F., Alger, H. M., Attia, Z. I., Barry, B., Chatterjee, R., Dolor, R., Friedman, P. A., Greene, S. J., Greenwood, J., Gundurao, V., Hackett, S., Jain, P., Kinaszczuk, A., Mehta, K., O'Grady, J., Pandey, A., Pullins, C., Puranik, A. R., Ranganathan, M. K., Rushlow, D., Stampehl, M., Subramanian, V., Vassor, K., Zhu, X., & Awasthi, S.
- Affiliations
- Nference (India) (IN)
- Mayo Clinic (US)
- Duke University (US)
- Mayo Clinic in Florida (US)
- The University of Texas Southwestern Medical Center (US)
- Novartis (United States) (US)
- Corresponding author
- Alger, Heather M.
- ORCID
- Venue
- American Heart Journal Plus: Cardiology Research and Practice
- Publisher
- Elsevier BV
- Published
- 2025-06
- Volume
- 54
- Pages
- 100528
- ISSN
- 2666-6022 (ISSN-L)
- PMC
- PMC12017965
- OpenAlex
- W4408726039
- Cited by
- 6
- References
- 19
- Citations by year
- 2026: 4 · 2025: 2
- FWCI
- 1.44
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- American Heart Journal Plus Cardiology Research and Practice (publishedVersion, cc-by-nc-nd)
- PubMed Central (publishedVersion, cc-by-nc-nd)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Health Sciences › Health Professions › Health Information Management › Electronic Health Records Systems
- Health Sciences › Medicine › Cardiology and Cardiovascular Medicine › ECG Monitoring and Analysis
- Keywords
- clinical decision support systems, left ventricular ejection fraction, electronic health record integration, echocardiography, cardiovascular risk stratification
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
