CoChat
Sign Up
Cited by 51Ethics and Social Impacts of AIArtificial Intelligence in Healthcare and EducationArtificial Intelligence in Healthcare

Artificial intelligence bias in the prediction and detection of cardiovascular disease

Open access
Mihan, Ariana, Pandey, Ambarish, Van Spall, Harriette G. C.·2024·npj Cardiovascular Health
DOI 10.1038/s44325-024-00031-9PMC PMC12912404
Open this paper in CoChat — free account

Abstract

AI algorithms can identify those at risk of cardiovascular disease (CVD), allowing for early intervention to change the trajectory of disease. However, AI bias can arise from any step in the development, validation, and evaluation of algorithms. Biased algorithms can perform poorly in historically marginalized groups, amplifying healthcare inequities on the basis of age, sex or gender, race or ethnicity, and socioeconomic status. In this perspective, we discuss the sources and consequences of AI bias in CVD prediction or detection. We present an AI health equity framework and review bias mitigation strategies that can be adopted during the AI lifecycle.

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.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s44325-024-00031-9

Cites

10 of 10 references verified0 citation errors · 0 not checkable

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.

A few free checks a day · papers already verified open instantly · results are public

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?

  • Even within the same geographic region, cardiovascular risk characteristics vary markedly across ethnic and healthcare subgroups.

    “Taken together, the evidence points to a central conclusion: even within the same geographic region, cardiovascular risk characteristics vary markedly across ethnic and healthcare subgroups.”

    From AI-driven cardiovascular risk prediction in patients with diabetes: bridging algorithmic innovation to equitable clinical application (2026) · cites this paper as [17]

    In “The role of AI in assessing risk, predicting, and detecting CVD” (page unavailable), the paper reports a disproportionate burden of CVD and cardiometabolic risk factors in some ethnic minorities and socioeconomically deprived groups. It does not establish that risk characteristics vary markedly within the same geographic region or across healthcare subgroups; its discussion of healthcare disparities does not make that specific comparison. (LLM, conf=0.82)

    Experimental — accuracy not yet measured

  • Assuming homogeneity across Asian populations or generalizing models developed from a single subgroup to the broader Asian population risks introducing systematic predictive bias and undermining equitable clinical utility.

    “Assuming homogeneity across Asian populations or generalizing models developed from a single subgroup to the broader Asian population risks introducing systematic predictive bias and undermining equitable clinical utility.”

    From AI-driven cardiovascular risk prediction in patients with diabetes: bridging algorithmic innovation to equitable clinical application (2026) · cites this paper as [17]

    The “Sources of AI bias” section (page unavailable) says that a training dataset that does not represent its target population can introduce bias, and that risk models may perform unequally across subpopulations. This supports the general warning about applying a model beyond the group in which it was developed, but the paper does not address homogeneity among Asian populations or generalization from one Asian subgroup to others. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • Treating 'Asian populations' as a homogeneous entity-or extrapolating models developed in one Asian subgroup to all Asian individuals-can introduce systematic prediction bias.

    “Treating “Asian populations” as a homogeneous entity—or extrapolating models developed in one Asian subgroup to all Asian individuals—can introduce systematic prediction bias.”

    From AI-driven cardiovascular risk prediction in patients with diabetes: bridging algorithmic innovation to equitable clinical application (2026) · cites this paper as [17]

    The “Sources of AI bias” section (page unavailable) says models can perform unequally across subpopulations when training data do not represent the target population, supporting the general risk of prediction bias from extrapolation. It does not discuss Asian populations, differences among Asian subgroups, or models developed in one Asian subgroup and applied to others. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • Bias can arise throughout problem formulation, dataset construction, model development, validation, and implementation in cardiovascular AI.

    “This is especially important because a recent scoping review found that fairness metrics are rarely reported in clinical risk prediction studies, while a recent cardiovascular AI perspective emphasized that bias can arise throughout problem formulation, dataset construction, model development, validation, and implementation (16, 17).”

    From AI-driven cardiovascular risk prediction in patients with diabetes: bridging algorithmic innovation to equitable clinical application (2026) · cites this paper as [17]

  • Substantial heterogeneity exists even within Asian populations.

    “Furthermore, substantial heterogeneity exists even within Asian populations, adding another layer of complexity to risk prediction.”

    From AI-driven cardiovascular risk prediction in patients with diabetes: bridging algorithmic innovation to equitable clinical application (2026) · cites this paper as [17]

    The ‘Sources of AI bias’ and ‘How AI bias may limit CVD prediction and detection’ sections discuss differences in risk prediction across demographic groups, but do not address heterogeneity within Asian populations. Searches for ‘Asian’ and ‘heterogeneity within Asian subgroups’ found no passage supporting that specific claim; page numbers were not available in our copy. (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

Flagged first, then by citations

Research works printed but not linked: 34

  1. [1]Mihan, A. & Van Spall, H. G. C. Interventions to enhance digital health equity in cardiovascular care. Nat. Med. 30, 628–630 (2024). 38355972 · DOI
  2. [2]Muse, E. D. & Topol, E. J. Transforming the cardiometabolic disease landscape: multimodal AI-powered approaches in prevention and management. Cell Metab. 36, 670–683 (2024). 38428435 PMC10990799 · DOI · Paper page
  3. [3]Averbuch, T. et al. Applications of artificial intelligence and machine learning in heart failure. Eur. Heart J. Digit. Health3, 311–322 (2022). PMC9707916 36713018
  4. [5]Atehortua, A. et al. Cardiometabolic risk estimation using exposome data and machine learning. Int. J. Med. Inf. 179, 105209 (2023). 37729839 · DOI
  5. [6]Fagherazzi, G. et al. Towards precision cardiometabolic prevention: results from a machine learning, semi-supervised clustering approach in the nationwide population-based ORISCAV-LUX 2 study. Sci. Rep. 11, 16056 (2021). 34362963 PMC8346462 · DOI
  6. [7]Ibrahim, H. et al. Health data poverty: an assailable barrier to equitable digital health care. Lancet Digit. Health 3, e260–e265 (2021). 33678589 · DOI · Paper page
  7. [8]Vervoort, D. et al. Addressing the global burden of cardiovascular disease in women: JACC state-of-the-art review. JACC83, 2690–2707 (2024). 38897679
  8. [9]Filbey, L. et al. Improving representativeness in trials: a call to action from the global cardiovascular clinical trialists forum. Eur. Heart J. 44, 921–930 (2023). PMC10226751 36702610
  9. [11]Kontopantelis, E. et al. Excess years of life lost to COVID-19 and other causes of death by sex, neighbourhood deprivation, and region in England and Wales during 2020: A registry-based study. PLoS Med. 19, e1003904 (2022). PMC8846534 35167587
  10. [16]Mihan, A., Pandey, A. & Van Spall, H. G. C. Mitigating the risk of artificial intelligence bias in cardiovascular care. Lancet Digit. Health10.1016/S2589-7500(24)00155-9 (2024). 39214762
  11. [17]Arnett, D. K. et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation 140, e596–e646 (2019). 30879355 PMC7734661
  12. [18]Navar, A. M. et al. Earlier treatment in adults with high lifetime risk of cardiovascular diseases: what prevention trials are feasible and could change clinical practice? Report of a National Heart, Lung, and Blood Institute (NHLBI) Workshop. Am. J. Prev. Cardiol. 12, 100430 (2022). 36439649 PMC9691440 · DOI
  13. [19]Adedinsewo, D. A. et al. Cardiovascular disease screening in women: leveraging artificial intelligence and digital tools. Circ. Res. 130, 673–690 (2022). 35175849 PMC8889564 · DOI
  14. [20]Segar, M. W. et al. Development and validation of machine learning–based race-specific models to predict 10-year risk of heart failure: a multicohort analysis. Circulation 143, 2370–2383 (2021). 33845593 PMC9976274 · DOI
  15. [21]Segar, M. W. et al. Incorporation of natriuretic peptides with clinical risk scores to predict heart failure among individuals with dysglycaemia. Eur. J. Heart Fail. 10.1002/ejhf.2375 (2022). PMC10535364 34730265 · DOI
  16. [22]Daniel Tavares, L. et al. Prediction of metabolic syndrome: A machine learning approach to help primary prevention. Diabetes Res. Clin. Pract.10.1016/j.diabres.2022.110047 (2022). 36029889 · DOI
  17. [23]Pastika, L. et al. Artificial intelligence-enhanced electrocardiography derived body mass index as a predictor of future cardiometabolic disease. NPJ Digit. Med. 7, 167 (2024). 38918595 PMC11199586 · DOI
  18. [24]Nadarajah, R. et al. Machine learning to identify community-dwelling individuals at higher risk of incident cardio-renal-metabolic diseases and death. Future Health. J. 11, 100109 (2024). · DOI
  19. [25]Myers, K. D. et al. Precision screening for familial hypercholesterolaemia: a machine learning study applied to electronic health encounter data. Lancet Digit. Health10.1016/S2589-7500(19)30150-5 (2019). PMC8086528 33323221
  20. [27]Reddy, H. et al. A critical review of global digital divide and the role of technology in healthcare. Cureus 14, e29739 (2022). 36340549 PMC9621721
  21. [29]Arora, A. et al. The value of standards for health datasets in artificial intelligence-based applications. Nat. Med. 29, 2929–2938 (2023). 37884627 PMC10667100 · DOI
  22. [30]Panch, T., Mattie, H. & Atun, R. Artificial intelligence and algorithmic bias: implications for health systems. J. Glob. Health 9, 10318 (2019). PMC6875681 31788229 · DOI · Paper page
  23. [34]Vasan, R. S. & van den Heuvel, E. Differences in estimates for 10-year risk of cardiovascular disease in Black versus White individuals with identical risk factor profiles using pooled cohort equations: an in silico cohort study. Lancet Digit. Health 4, e55–e63 (2022). 34952676 PMC8715354 · DOI
  24. [35]Varga, T. V. Algorithmic fairness in cardiovascular disease risk prediction: overcoming inequalities. Open Heart10.1136/openhrt-2023-002395 (2023). PMC10649900 37963683
  25. [36]Tat, E., Bhatt, D. L. & Rabbat, M. G. Addressing bias: artificial intelligence in cardiovascular medicine. Lancet Digit. Health10.1016/S2589-7500(20)30249-1 (2020). 33328028
  26. [37]Mulvagh, S. L. et al. The Canadian Women’s Heart Health Alliance ATLAS on the epidemiology, diagnosis, and management of cardiovascular disease in women — Chapter 9: Summary of current Status, challenges, opportunities, and recommendations. CJC Open 6, 258–278 (2024). 38487064 PMC10935707 · DOI
  27. [38]Elias, P. et al. Deep learning electrocardiographic analysis for detection of left-sided valvular heart disease. J. Am. Coll. Cardiol. 80, 613–626 (2022). 35926935 · DOI · Paper page
  28. [40]Li, F. et al. Evaluating and mitigating bias in machine learning models for cardiovascular disease prediction. J. Biomed. Inform.10.1016/j.jbi.2023.104294 (2023). PMC11104322 36706849 · DOI
  29. [41]Hong, C. et al. Predictive accuracy of stroke risk prediction models across Black and White race, sex, and age groups. JAMA10.1001/jama.2022.24683 (2023). PMC10408266 36692561
  30. [48]Dupulthys, S. et al. Single-lead electrocardiogram Artificial Intelligence model with risk factors detects atrial fibrillation during sinus rhythm. Europace10.1093/europace/euad354 (2024). PMC10872711 38079535
  31. [49]Segar, M. W. et al. Machine learning-based models incorporating social determinants of health vs traditional models for predicting in-hospital mortality in patients with heart failure. JAMA Cardiol. 10.1001/jamacardio.2022.1900 (2022). PMC9260645 35793094 · DOI
  32. [50]Meng, J. & Xing, R. Inside the black box: embedding clinical knowledge in data-driven machine learning for heart disease diagnosis. Cardiovasc. Digit. Health J. 3, 276–288 (2022). 36589311 PMC9795264 · DOI
  33. [51]Ho, V. et al. Physician- and patient-elicited barriers and facilitators to implementation of a machine learning–based screening tool for peripheral arterial disease: preimplementation study with physician and patient stakeholders. JMIR Cardio10.2196/44732 (2023). PMC10660241 37930755
  34. [52]Hutson, M. How AI is being used to accelerate clinical trials. Nature10.1038/d41586-024-00753-x (2024). 38480968

Cited by

Sorted by citations

Authors
Mihan, A., Pandey, A., & Van Spall, H. G. C.
Affiliations
  • McMaster University (CA)
  • The University of Texas Southwestern Medical Center (US)
ORCID
Venue
npj Cardiovascular Health
Publisher
Springer Science and Business Media LLC
Published
2024-11-21
Volume
1
Issue
1
ISSN
2948-2836 (ISSN-L)
DOI
10.1038/s44325-024-00031-9
PMC
PMC12912404
OpenAlex
W4404600835
Cited by
51
References
10
Citations by year
2026: 30 · 2025: 21
FWCI
3.88
Type
journal article
Language
en
Open access
Licence
Locations
  • npj Cardiovascular Health (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
Topics
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
algorithmic bias, cardiovascular disease prediction, cardiovascular disease detection, algorithmic fairness, health equity, algorithmic discrimination, bias mitigation, AI lifecycle, racial and ethnic disparities, socioeconomic status, early 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