Abstract
Background: Patients at risk of atherosclerotic cardiovascular disease (ASCVD) have low statin use. Clinician perceptions of the ASCVD risk estimates that guide statin prescribing may contribute to poor uptake. At the time of the study, the only equations used to predict ASCVD risk (the Pooled Cohort Equations; PCE) provided race-specific estimates, a controversial practice and a potential reason why clinicians may scrutinize these estimates. We sought to examine how clinicians perceived ASCVD estimates, in relation to their perceptions of race and, also, more broadly.
Methods: We conducted an interpretive description study using ten 45-minute semistructured interviews with primary care physicians in North Carolina between March and April 2022. Interviews focused on the PCE ASCVD risk calculator and perspectives of race as it relates to ASCVD. Responses were analyzed using both deductive and inductive approaches to identify primary topics.
Results: 5 men and 5 women participated. Of these, 6 identified as White, 2 as Black, and 2 as Asian. Three main topics emerged. First, participants felt conflicted about the role of race in ASCVD risk. Second, they had several concerns with the calculator that went beyond race, including its emphasis on statin use and lack of social determinants of health. Finally, participants universally valued the PCE ASCVD calculator as a tool to educate patients and inspire statin initiation and behavioral change.
Conclusions: The PCE ASCVD risk calculator was seen as most useful in facilitating discussions regarding behavior and lifestyle changes, suggesting the potential benefit of incorporating variables related to patients' health behaviors in a revised model. The new PREVENT equations provide a helpful first step by removing race and including social determinants. The next step may be to add health behaviors and visual images to facilitate patient counseling and comprehension.
- Atherosclerosis
- Cardiology
- Cardiovascular Diseases
- Cardiovascular Risk Factors
- Clinical Decision-Making
- Counseling
- Health Behavior
- Health Disparities
- Lifestyle
- North Carolina
- Primary Care Physicians
- Primary Health Care
- Race Factors
- Risk Score
- Social Determinants of Health
- Statins
Introduction
The pooled cohort equations (PCE) are a set of race and sex-specific formulas that estimate an individual’s 10-year risk of developing a myocardial infarction or stroke. Current guidelines1 recommend that clinicians use these equations for patients without a history of ASCVD to identify those at greater risk and take preventive action via statin initiation. However, a significant gap remains between the calculator-guided recommendations for statin use and patients uptake of statin medications. In Clough and colleagues’ 2017 survey of primary care clinicians, 78% reported frequently using the tool.2 However, among patients of these clinicians, only 22% of persons eligible to receive a statin based on their ASCVD score received treatment.2 Other studies on statin prescribing have found that eligible patients are often not treated.3⇓–5 While patient preferences naturally impact statin prescribing patterns, it is possible that clinician perceptions of the value of the ASCVD estimates themselves play a role. To our knowledge, no studies have evaluated these perceptions.
One aspect of the PCE that may impact clinician use is its incorporation of race in risk estimation. The equations categorize racial groups as Black and non-Black (the White and “Other” categories use the same equations). To critics, using race to determine patient care is a controversial practice as it can promote the notion that Black people are biologically different from other racial groups, contributing to racial essentialism6 and the potential mistreatment of Black people. Concerns regarding essentialism7 may impact how clinicians use and incorporate PCE estimates. Although an alternative race-neutral ASCVD risk estimator (named PREVENT8) has recently been developed, understanding the perspectives impacting clinicians use of the PCE remains important as it is still commonly used; race remains employed in other medical algorithms; and there may be other aspects of the PCE that concern clinician that are also applicable to PREVENT.
To address these issues, we sought to examine how clinicians perceived the PCE (the only ASCVD risk tool at the time of the study) particularly in relationship to their understanding of what race represents, but also more broadly. We used in-depth semistructured interviews to explore how clinicians employed the PCE, their concerns about the tool, and their understanding of race as it relates to ASCVD risk.
Methods
Study Population
This was an interpretative description study,9 guided by a constructivist interpretive framework that acknowledges the role of the researcher in cocreating meaning.10 Ten actively-practicing North Carolina family medicine and internal medicine physicians were recruited between February and March of 2022 to participate in this study. The narrowness of our topic area justified the use of our small sample size.11 Participants were identified using the North Carolina Network Consortium (NCNC)12 and the University of North Carolina (UNC) Family Medicine department listserv. While the NCNC is administered by UNC, participating clinics include nonacademic clinics in North Carolina. Using purposive sampling,13 with attention to gender, ethnic, and racial diversity, clinicians belonging to the NCNC were individually invited to participate in Zoom interviews focused on cardiovascular disease risk counseling. As race was a core aspect of this study, participants were also asked to forward the invitation to racially minoritized colleagues whom they believed would be appropriate for the study. Seven participants were recruited in this manner. Finally, to reach our recruitment goals, the remaining 3 participants were solicited using the UNC Family Medicine Department Listserv, with respondents selected with attention to race, gender, and ethnicity. Before the interview, participants were asked via e-mail to provide an open-ended description of their race, ethnicity and gender, which were verbally confirmed before the interview. Each participant received a $80 incentive. This study was approved by the Institutional Review Board of the University of North Carolina at Chapel Hill.
In-Depth Interviews
Our semistructured interviews focused on physicians’ perspectives, experiences, and challenges encountered in their use of the PCE (termed “ASCVD risk calculator” in interviews as the PCE was the only recommended risk equation at the time of the study); how they conceptualize and counsel patients regarding their ASCVD risk; and their perspectives on race as ASCVD risk factors. The interview guide (Appendix 1) was developed with a senior qualitative researcher. Interview questions moved from broad to narrow, and due to the sensitivity of the topic, participants perspectives on race were sought at the end of the interview. Interviews lasted approximately 45 minutes, were audio recorded, and transcribed verbatim using a transcription service.
Reflexivity
Reflexivity was achieved through writing and discussing memos detailing the primary coders’ (EO, OA) relationship to the project. Both coders identify as Black, and one (EO) was also a family medicine physician. The objective was not to eliminate subjectivity, but to enhance the research's validity14 by examining and reducing potential biases and their impact on data analysis and interpretation.
Analysis
Data analysis involved deductive and inductive approaches to develop topics. Deductive codes were derived from the central topics of the interview questions. Inductive codes were nuanced topics identified as subcodes of the more encompassing deductive codes or unanticipated topics derived from information that was not directly solicited. Intercoder triangulation was achieved by using 2 coders to code each transcript. Initial transcripts were reviewed by EO, who developed a preliminary set of codes. This coding framework was subsequently refined through collaboration with a senior qualitative researcher (PM) and other team members. To ensure rigor, EO and OA independently coded the first 4 transcripts and resolved discrepancies via discussions with each other and the team. For the remaining 6 transcripts, a similar approach was followed, with each author coding 3 transcripts that were then reviewed by the other coder, with disagreements resolved through consensus meetings. Transcripts were reviewed to determine if new, relevant ideas had surfaced requiring new codes. This iterative process provided an avenue to determine whether sufficient conceptual depth was reached.15 Given that no new codes were discernable in the last 6 transcripts—other than one participant’s uncertainty in answering a particular question—we determined that we had reached sufficient conceptual depth of topics across the interviews. Codes were organized into primary topics, while continuous revision of both codes and topics took place throughout the entire coding process. We explored relationships between codes for contextualization but did not explicitly code these relationships. The study team convened twice a month over a 3-month period to solidify topics, adapt the codebook, and refine the study’s findings.
Results
Five men and 5 women participated. Of these, 6 identified as non-Hispanic White, 2 as non-Hispanic Asian American (here, subcategories of Asian American were provided by respondents but recategorized by researchers), and 2 as non-Hispanic Black American. Because of the small sample, responses were not contextualized by participants’ identities. Three main topics emerged. Participants felt conflicted about the role of race in ASCVD risk, they had several concerns with the calculator that went beyond race, and they universally valued the calculator as a tool to educate patients and inspire behavioral change (see Appendix 2).
Race and the ASCVD Risk Estimator
Clinicians held varied perspectives on the nature of race, describing it as a primarily biological, sociobiological, or social trait. Some indicated an evolution in their understanding of race over their medical training and career, while others noted their perceptions about race remained unchanged. Some clinicians viewed the ASCVD calculator as treating race as a biological variable while others were uncertain or believed race was used to represent socially derived risk. Clinicians who believed that race was used as a biological variable in the model cited the lack of other social factors and the time period in which the model was developed as justification for their perspective.
I assume that they're putting it in there as a biological trait. … I don't think people, up until more recently, [were] really sensitive about the social factors that might be inherent in any purported increased risk of hypertension. (Participant 9)
Given that the ASCVD risk calculator is clear that it uses race in its risk estimation, clinicians were asked to describe how they addressed race-based ASCVD risk with their patients. Several mentioned that they did not discuss the relationship between race and ASCVD, and a subset of these clinicians stated that because race was a nonmodifiable factor in the model, it was not viewed as relevant when counseling patients. Some clinicians advised patients that their racial background might be associated with an increased risk of ASCVD.
I wish actually the calculator would plop out both numbers without race versus with race being identified. I don't think I'd bring it up. (Participant 1)
Participants attributed Black patients’ elevated risk of ASCVD to a combination of social factors, including structural racism, and cultural factors, such as dietary habits. No clinician expressed that biological differences explained racial differences in ASCVD, though some felt that race had biological components. Statements regarding the relationship between race, cultural practices, and social factors were captured in statements such as:
I'm predominantly vegetarian, I can't tell my African American patient from Charlotte, North Carolina, to eat vegetarian. (Participant 2)
For ASCVD, my sense is that this is one where race might be more related to observations, related to social factors, structural racism, other things like that rather than a biologic difference. (Participant 5)
Clinician responses reflected two conflicting perspectives. Clinicians recognized that Black patients faced an increased risk of ASCVD due to social and cultural factors. However, they expressed concern about using race to calculate this increased risk. Participants who viewed the inclusion of race in the ASCVD risk calculator as a limitation commonly contextualized this perspective within broader statements about medicine’s transition away from using race to determine disease risk.
I think it should take into account race, sadly, because I do think it's important and I think it has an impact. (Participant 10)
I have somebody who's doing all the right things and just [because] they're African American, their risk is higher? (Participant 6)
Clinicians suggested the following approaches to address the use of race within the ASCVD calculator: provide race-neutral and race-specific estimates; offer solely race-neutral estimates; and continue using race for risk estimation while clarifying that race represents socially derived risk factors.
What I don't like about it is it reinforces the idea that there's [a] genetic basis … It would be nice if it reminded people that it's not. (Participant 10)
Beyond Race-Based Risk: Limitations of the ASCVD Risk Estimator
In addition to the use of race in the ASCVD calculator, clinicians described the following shortcomings of the tool: the excessive focus on statin treatment; absence of social determinants of health (SDOH) and health behaviors; the potential confusion resulting from different thresholds for initiating aspirin and statins; the tool's limited age range; the overemphasis of age in the risk equations; and doubts about its accuracy, particularly when used to calculate the risk of diverse patient groups.
Clinicians expressed concern that risk estimates were exclusively tied to statin treatment. Concerns regarding the impact of age in the risk equations and the emphasis on statin prescribing were ideas that were sometimes expressed together, with some clinicians questioning the value of calculating ASCVD risk for healthy older adults, who would inevitably be offered statins.
I guess that's the bias of the calculator is that it really gets you focused on statin medication but there … are diet and exercise and therapeutic lifestyle changes that can be part of an approach to thinking about preventing ASCVD in an individual. (Participant 1)
I use it rarely now … and the reason is that everybody just needed to be on a statin. … Why am I spending the time doing this when everybody's on a statin according to the calculator? (Participant 3)
SDOH and health behaviors were viewed as a critical component of ASCVD risk and the fact that these were missing from the ASCVD calculator was seen as limiting its applicability to high-risk patient populations.
The calculator is projecting the patient to be healthier than they actually are because of what I know about the patient's life. So, I take it with a grain of salt. So maybe like 20 or 30% off. … Part of why is because they don't factor [in] any social determinants. And then the second reason why is because we also don't know what else [patients are] doing. You know, stress, substances, lifestyle changes. (Participant 2)
At the same time, clinicians recognized the difficulty in creating risk models that included factors that are hard to measure. In contrast to the variables used in the ASCVD risk calculator that can be easily ascertained from the medical record, social risks are hard to capture.
The biggest risk factors … [are] lack of education, lack of understanding, and apathy. Those are the biggest risk factors, but we can't quantify that. (Participant 7)
Importance of the ASCVD Risk Estimator to Patient Education
Many clinicians used the ASCVD calculator to understand their patients’ ASCVD risk, particularly with patients for whom clinicians had uncertainty about their risk of ASCVD. However, all participants felt that the risk calculator was useful when assessing patients’ readiness to change, facilitating changes in patients’ attitudes and behaviors, and fostering shared decision making regarding the initiation of statin medications. Even clinicians who infrequently employed the calculator, regarded it as a tool that improved patients’ understanding of their ASCVD risk. The perspectives of a low-utilizer clinician was captured in the following statement:
If I'm having a hard time getting people to buy in to some of the things they need to do to change, I'll pull it out and show them what their risk is and chances in the next 10 years of them having an event. And that's helpful in that situation sometimes. (Participant 4)
However, clinicians also acknowledged that informing patients about their ASCVD risk was not always helpful, particularly for patients who had difficulty understanding their risk scores or whom clinicians felt were not particularly concerned about experiencing a heart attack or stroke.
I have a good number of patients that don't care what their risk is. I can plug and chug, hit the calculator, spit out a number and show it to them and say, “All right this is saying that in ten years, you have a 30% chance of a cardiovascular or other vascular event happening.” And sometimes I just don't get the level of … people being impressed that I would expect. (Participant 1)
Perceptions of the validity of the ASCVD calculator did not determine its use with patients. Physicians used the calculator to counsel patients despite their ambivalence regarding its accuracy. This could be explained by the lack of other tools to estimate ASCVD risk.
I remember when ASCVD first came out, there [were] actually a lot of questions about the validity of it. … Those questions just went away after a few years and we had no other alternative so that's just simply what we used was that calculator. (Participant 1)
While clinicians used the calculator to identify patients at greater risk of ASCVD, several were clear that they did not need the calculator to do so. Therefore, the ASCVD calculator was perceived as a tool that, for some patients, proved valuable in facilitating beneficial changes aimed at lowering their risk of ASCVD. This was the case regardless of clinicians views regarding its accuracy and value in estimating such risk.
Discussion
Participants expressed nuanced views regarding using race to assess risk, had varied perceptions of the PCE ASCVD risk calculator’s limitations, and unanimously agreed that this ASCVD risk calculator was useful for patient counseling. Race was commonly viewed as an inadequate substitute for social determinants of health—which some participants believed warranted its replacement—and the use of race to estimate risk was sometimes perceived as contributing to racial essentialism. Clinician enthusiasm for the PCE ASCVD calculator was most notable for its usefulness in facilitating patient understanding and behavior change, which suggests that incorporating health behaviors in the ASCVD estimates may improve its value.
Our study was small, but the finding that some clinicians saw race as partly biological, aligns with prior research. Futterman et. al16 found that that most medical educators viewed race as having biological components and many educators tied race to disease risk, particularly hypertension. Nevertheless, our study participants did not attribute racial differences in ASCVD to biological differences between racial groups, agreeing with findings from Okah and colleagues17 showing that, among family medicine physicians, social factors were seen as a greater contributor to racial health inequity than genetic differences. Thus, although clinicians may see race as biologically significant, they still place importance on the social factors that contribute to health inequities.
In response to calls to remove race from ASCVD risk estimation18,19 the American Health Association developed the PREVENT Online Calculator,8,20 which excludes race and includes SDOH, thereby addressing concerns expressed by some participants regarding racial essentialism and the lack of social factors in the PCE. Notably, Ghosh and colleagues did not find a meaningful difference in predictive accuracy when comparing a race-neutral and SDOH-inclusive ASCVD risk model to a race-inclusive and SDOH-neutral ASCVD model.21 Work by Colantonio et al. demonstrating a small improvement in predictive accuracy with the inclusion of SDOH supports these findings.22 Both studies suggest that although racial groups may differ in their socioeconomic profiles, the increased risk of ASCVD among Black people is likely accounted for by variables already incorporated in the PCE, as explicitly noted by Ghosh et al.21
However, we also found that perceptions of validity were not relevant to clinicians’ use of the PCE ASCVD risk calculator and the unanimous belief that the risk estimator was helpful in counseling patients and encouraging behavior change, particularly for patients who were able to comprehend their risk scores. Greater attention is needed regarding how the PCE, or the newer PREVENT, can better facilitate lifestyle change, especially since other studies show that concerns about statin side effects are a significant barrier to patient uptake.5,23 Identifying the health behaviors that clinicians find most useful to include in the calculator would be a helpful next step. In addition to estimating the impact of smoking cessation (currently available in the PCE), clinicians may want to show their patients what happens to their ASCVD risk when they make changes to their physical activity level or dietary behaviors, which cannot be done with either the PCE or PREVENT. Therefore, developing a model that incorporates health behaviors that go beyond medication initiation, in a manner that facilitates patient comprehension, may be more suitable for clinicians and patients.
To our knowledge no previous studies have examined clinician perspectives on the PCE ASCVD risk calculator. Therefore, our findings cannot be directly compared with prior research. However, related publications support our finding that clinicians value visual tools that facilitate decision making. For instance, in a qualitative study conducted by Ahmed et al., clinicians and patients expressed that images and visual tools were valuable for shared decision making regarding statin initiation.24 In addition, the belief in the importance of using ASCVD risk estimates to facilitate conversations regarding behavior change was the justification provided by Mendez and colleagues in their development of a tool to visualize these estimates.25 Likewise, QRISK3,26 a tool used in the United Kingdom to predict the 10-year risk of heart attack or stroke, and Mayo Clinic’s Statin Choice Decision Aid27 include a visualization with the risk estimate, potentially for patient education.
There were notable strengths and limitations to this study. To our knowledge, this is the first study to evaluate clinician perceptions of a commonly used medical calculator. In addition, this study is timely, as efforts to revise race-inclusive algorithms are underway and could benefit from end-user feedback. Regarding limitations, this was a small qualitative study with only 10 participants. This is an inherent limitation of qualitative work as it does not seek a representative sample in its study design. Instead, qualitative research focuses on the depth of information provided by participants. Second, study participants were primary care clinicians from one state, and findings may not be applicable to clinicians practicing outside of North Carolina or to specialists, such as cardiologists, who may use the risk equations. Third, given the nature of the questions, participants were aware that the interviewers were seeking information regarding their thoughts on how race is related to ASCVD risk, and their comments on the appropriateness of race as a variable in the calculator may reflect awareness of that fact. Finally, race concordance was not used in this study, potentially affecting how participants responded to questions regarding race. Nevertheless, most of our questions were about the ASCVD calculator and ASCVD risk and not race itself.
Conclusion
The PCE ASCVD risk calculator was unanimously viewed as useful in facilitating conversations regarding behavior and lifestyle changes. Therefore, inclusion of variables regarding patients’ health behaviors could further support clinician in these efforts. For some clinicians, PREVENT provides a helpful first step by removing race and including a measure of social determinants. Nevertheless, the inclusion of additional health behaviors to facilitate patient counseling may facilitate clinician use of the calculator. Further, larger survey studies will be needed to determine how, or which, additional health behaviors should be included.
Appendix 1
Interview Guide
When you think about atherosclerotic cardiovascular disease what goes through your mind? From now on I’ll refer to it as ASCVD.
When you talk to other clinicians about ASCVD risk, what kinds of things do you think about?
Do you have an approach in preventing ASCVD in your patients?
How did you develop that approach?
i. Did you use literature?
ii. Did you receive training in medical school or residency?
Can you walk me through a recent time you met with a patient to discuss their ASCVD risk?
What do you consider to be the most important risk factors for developing ASCVD?
Can you tell me more about that?
Can you talk about why ____ is important?
Do you consider this factor to be related to behavior or genetics? How so?
How do you approach counseling patients about their ASCVD risk factors for which they have no control?
Risk related to their race?
Risk ASCVD risk related to their sex?
How do your patients respond to this counseling?
How do approach counseling patient about their ASCVD risk for factors for which they have control?
Risk related to their blood pressure?
Risk related to their cholesterol intake?
Risk related to their smoking?
How do your patients respond to this counseling?
When do you use the ASCVD risk calculator?
Are there clinical situations in which you are more likely to use it than others? If so, can you describe these situations?
Does anything concern you about the ASCVD risk calculator? For example, are there any variables in the calculator that you believe are inappropriate?
Walk me through how you use the calculator. Where do you get information for the calculator?
How well do you think the ASCVD risk calculator approximates risk of ASCVD?
What do you base that opinion on?
Do you have any thoughts on how you might relate the increased risk of ASCVD among men to an individual male patient’s risk of ASCVD?
Do you have any thoughts on how you might relate the increased risk of ASCVD in the Black community to an individual Black patient’s risk of ASCVD?
Do you think the calculator is treating race as a social or biological trait? Can you tell me more about that?
As a primary care physician, do you think race a social or biological trait? Can you tell me more about that?
Have your thoughts about race changed during your time as a physician? Can you tell me more about that?
Appendix 2
Themes, Codes, Subcodes, and Example Quotes Related to Clinicians’ Perceptions of Race-Based ASCVD Risk and the ASCVD Risk Estimates
Notes
This article was externally peer reviewed.
Funding: Ebiere Okah was supported by the National Institutes of Health’s National Center for Advancing Translational Sciences under Award Number UL1TR002494. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health’s National Center for Advancing Translational Sciences.
Conflict of interest: The authors have no conflicts of interest to declare.
To see this article online, please go to: http://jabfm.org/content/38/3/464.full.
- Received for publication August 29, 2024.
- Revision received December 5, 2024.
- Accepted for publication January 13, 2025.
