Revising the atherosclerotic cardiovascular disease calculator without race

Abstract

Momentum has increased to eliminate race correction from clinical algorithms, but efforts have focused primarily on examples in which adjusting by race might worsen an existing racial inequity, such as in the revised vaginal birth after cesarean section calculator and equations for estimated glomerular filtration rate.1Vyas DA Eisenstein LG Jones DS Hidden in plain sight—reconsidering the use of race correction in clinical algorithms.N Engl J Med. 2020; 383: 874-882Crossref PubMed Scopus (396) Google Scholar Other examples, such as the atherosclerotic cardiovascular disease (ASCVD) Risk Estimator, raise the question of whether race correction is acceptable if used in algorithms that could be considered somewhat protective towards racial or ethnic groups that have been historically marginalised.2Poppas A American College of CardiologyAmerican College of Cardiology response to congressional letter on race and health inequity.https://waysandmeans.house.gov/sites/democrats.waysandmeans.house.gov/files/documents/FINAL%20-%20ACC%20Response%20to%20Congressional%20Letter%20on%20Race%20and%20Health%20Inequity.pdfDate: Sept 25, 2020Date accessed: October 15, 2021Google Scholar The pooled cohort equations (PCE) used in the ASCVD tool tend to produce higher 10-year cardiovascular risk estimates for patients identified as Black than for their white counterparts, potentially increasing the likelihood that Black patients will receive medical interventions, such as statin therapy. In The Lancet Digital Health, Ramachandran S Vasan and Edwin van den Heuvel characterise the magnitude of the differential estimates in Black versus White individuals pRoduced by the ASCVD calculator.3Vasan RS 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. 2022; 4: e55-e63Summary Full Text Full Text PDF PubMed Scopus (2) Google Scholar The authors demonstrate that race correction in the ASCVD tool creates sizable differences between Black versus White individuals in predicted 10-year risk of cardiovascular event that are “substantial and seem biologically implausible”. They reported differences in absolute cardiovascular disease risk in Black versus White individuals to be as large as 22·8% (median 6·25%, range 0·15–22·8) with a median relative risk of 2·40 (range 1·02–12·6). Their study makes an important contribution to the literature by quantifying the potential effect of including race in ASCVD risk stratification. Since most differences between Black versus White individuals produced higher risk estimates for Black patients, the authors hypothesise that Black patients might be more likely to be prescribed medications like statins. Nonetheless, extant research suggests that despite this potential for over-treatment in Black populations, racial inequities in statin prescribing persist, emphasising the need for further research to investigate how differential risk estimates correlate to clinical consequences, including the effect on medication initiation, side-effect burden, and cardiovascular risk reduction in Black patients.4Nanna MG Navar AM Zakroysky P et al.Association of patient perceptions of cardiovascular risk and beliefs on statin drugs with racial differences in statin use: insights from the patient and provider assessment of lipid management registry.JAMA Cardiol. 2018; 3: 739-748Crossref PubMed Scopus (53) Google Scholar In addition to demonstrating the potential clinical consequences, the authors highlight that the continued use of race in ASCVD risk reifies the concept of race as biologically meaningful, noting that, “by using race in the PCE, we might be normalising and legitimising a social construct as a medically valid classifier (Black versus White), leading the uninitiated to equate race-related differences in estimated cardiovascular disease risk with actual biological differences in disease susceptibility.” Our understanding of race is as a social and political construct rather than a proxy for biological difference, but a myriad of studies show that clinicians believe true biological difference exists between races, at times resulting in differential medical treatment.5Amutah C Greenidge K Mante A et al.Misrepresenting race—the role of medical schools in propagating physician bias.N Engl J Med. 2021; 384: 872-878Crossref PubMed Scopus (73) Google Scholar Tools like ASCVD that implement race alongside variables such as blood pressure and serum cholesterol levels reinforce this false and harmful notion of race as an equivalent biological category. The limited racial categories offered by the ASCVD tool exclude large groups of the global population and do not accurately represent the complexity of patients’ racial and ethnic identities.6National Heart, Lung, and Blood InstituteAssessing cardiovascular risk: systematic evidence review from the Risk Assessment Work Group, 2013.https://www.nhlbi.nih.gov/sites/default/files/media/docs/risk-assessment.pdfDate: 2013Date accessed: October 15, 2021Google Scholar The dramatic Black versus White differences highlighted by Vasan and van den Heuvel also echo previous calls to reconsider calibration of the PCE, which have been shown to overestimate or misestimate cardiovascular risk in several racial and ethnic groups.7Yadlowsky S Hayward RA Sussman JB McClelland RL Min Y Basu S Clinical implications of revised pooled cohort equations for estimating atherosclerotic cardiovascular disease risk.Ann Intern Med. 2018; 169: 20-29Crossref PubMed Scopus (78) Google Scholar The PCE were developed from population-based cohorts containing nearly 5 times as many White individuals as Black individuals.6National Heart, Lung, and Blood InstituteAssessing cardiovascular risk: systematic evidence review from the Risk Assessment Work Group, 2013.https://www.nhlbi.nih.gov/sites/default/files/media/docs/risk-assessment.pdfDate: 2013Date accessed: October 15, 2021Google Scholar Previous research by Yadlowsky and colleagues reported extreme and implausible ASCVD risk estimates for Black individuals, suggesting the original PCE model was overfitted owing to the small sample size of Black individuals included, ultimately leading to poor calibration for Black patients.7Yadlowsky S Hayward RA Sussman JB McClelland RL Min Y Basu S Clinical implications of revised pooled cohort equations for estimating atherosclerotic cardiovascular disease risk.Ann Intern Med. 2018; 169: 20-29Crossref PubMed Scopus (78) Google Scholar For these reasons, we agree with Vasan and van den Heuvel that the race variable in the ASCVD calculator should be reconsidered. In fact, we believe race should be removed altogether. The authors propose that race should be replaced with the causal factors that it represents and could even be substituted with a social deprivation index to account for social determinants: “If replaced by appropriate causal variables, race should no longer improve prediction in the risk algorithms.” However, adjusting predictive models for social determinants is also controversial and might simply recreate the pitfalls of race correction by disadvantaging individuals with relative social deprivation. The reflex to adjust away the data signal that appears for race must be revisited. Many predictive clinical tools ultimately include race because observational studies find a persistent and statistically significant correlation between race and a health outcome of interest.1Vyas DA Eisenstein LG Jones DS Hidden in plain sight—reconsidering the use of race correction in clinical algorithms.N Engl J Med. 2020; 383: 874-882Crossref PubMed Scopus (396) Google Scholar Despite adjusting for multiple social determinants. such as income, housing, and insurance status, tool developers might find that a signal for race still persists. This scenario should not be surprising given the relationship between structural racism and health. Individual level factors that are conventionally used to reflect social determinants do not capture the internalised and interpersonal experience of racism and its impact on health outcomes.8Jones CP Toward the science and practice of anti-racism: launching a national campaign against racism.Ethn Dis. 2018; 28: 231-234Crossref PubMed Scopus (93) Google Scholar For example, studies that adjust for individual income cannot adequately capture the ways that social inequality and spatial polarisation caused by racial economic inequality influence health. Efforts have been made in the social sciences to move beyond individual factors in our descriptive measures, such as Douglas Massey's introduction of the Index of Concentration at the Extremes, to account for concentrations of privilege and deprivation; Krieger and colleagues have since applied this framework to population health.9Krieger N Waterman PD Spasojevic J Li W Maduro G Van Wye G Public health monitoring of privilege and deprivation with the index of concentration at the extremes.AJPH. 2016; 106: 256-263Crossref PubMed Scopus (121) Google Scholar Seeking more nuanced descriptive tools that account for relative deprivation could equip researchers to better represent the complex effects of racism on health and to better understand the data signal that often persists between race and health outcomes. These measures should be used for population health monitoring to help to envision interventions that target the structural forces that shape and reinforce inequity. By contrast, including social determinants or markers of social deprivation in predictive modelling could recapitulate the dangers of race correction by replacing one measure of sociopolitical disadvantage (racism) with others. To capitalise on the important findings of Vasan and van den Heuvel, the American College of Cardiology and American Heart Association should expand on its recent advancement of race as a social construct by committing to formal reappraisal and revision of the ASCVD tool by removing race.10Churchwell K Elkind MSV Benjamin RM et al.Call to action: structural racism as a fundamental driver of health disparities: a presidential advisory from the American Heart Association.Circulation. 2020; 42: e454-e468Google Scholar Such a change would help to ensure that our stated ideals about racial justice align with our clinical standards in advancing health equity. We declare no competing interests. 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 studyThe PCE might generate substantially divergent cardiovascular disease risk estimates for Black versus White individuals with identical risk profiles, which could introduce race-related variations in clinical recommendations for cardiovascular disease prevention. Full-Text PDF Open Access