Gaps and Disparities in Primary Prevention Statin Prescription During Outpatient Care

Abstract

The 2018 ACC/AHA Guideline on the Management of Blood Cholesterol recommends statin therapy for eligible patients to reduce the risk of atherosclerotic cardiovascular disease (ASCVD). We extracted electronic health record (EHR) data for patients with at least one primary care or cardiology visit between October 2018 and January 2020 at an urban, academic medical center in New York City. Clinical and demographic data were used to identify patients eligible for primary prevention statin therapy. Statin prescription status was extracted from the EHR, and multivariate logistic regression was used to assess the association between statin prescription and age, sex, race, ethnicity, and other clinical and demographic covariables. Among 7,550 patients eligible for primary prevention statin therapy, 3,994 (52.9%) were prescribed statins on at least one visit. Statin prescription was highest among patients with diabetes mellitus (73.6%) and with 10-year ASCVD risk ≥ 20% (60.6%), and was lowest for those with 10-year ASCVD risk between 5% and 7.5% (18.7%). Compared to those never prescribed statins, patients prescribed statins were less likely to be women, mainly driven by lower statin prescription rates for women with diabetes. In a fully adjusted model, women remained less likely to be prescribed statin therapy (adusted odds ratios 0.79, 95% confidence interval, 0.71 to 0.88). In conclusion, primary prevention statin therapy remains underutilized.

Introduction

Cardiovascular disease (CVD) affects approximately 26.1 million Americans, contributing to 868,662 deaths in 2018.¹ Statin therapy is one of the most well-established approaches for primary prevention of atherosclerotic cardiovascular disease (ASCVD) and has been shown to reduce cardiovascular events and death.^1–3 Accordingly, the 2018 American Heart Association / American College of Cardiology Guideline on the Management of Blood Cholesterol recommends risk stratification of primary prevention patients to identify those with high risk of ASCVD events and who would most benefit from statin therapy, using both estimated 10-year ASCVD risk and the presence of additional risk enhancers.⁴ Despite clear guideline recommendations, studies have shown that 40% to 50% of patients eligible for primary prevention statin therapy are not prescribed statins.^5,6 Moreover, women, racial and ethnic minorities, and patients with socioeconomic barriers to healthcare are less likely to receive guideline-recommended statin therapy.^7,8,9 We undertook the present analysis to evaluate gaps and disparities in primary prevention statin therapy during the time period immediately following the release of the 2018 ACC/AHA guidelines, using data from a socioecomically diverse group of patients treated in primary care and cardiology clinics at an academic medical center.

Methods

Using electronic health record (EHR) data (Allscripts Sunrise Clinical Manager™, Chicago, IL, 2018), we identified all primary care or cardiology clinic visits between October 2018 and January 2020 at the Ambulatory Care Network affiliated with Columbia University Irving Medical Center, a clinic network that serves a diverse patient population in upper Manhattan. For each visit, the following data were extracted from the EHR: age, sex, race and ethnicity (self-reported but required for registration), medical comorbidities (based on International Statistical Classification of Diseases and Related Health Problems, 10^th revision [ICD-10] codes), allergies, lipid levels (total cholesterol, triglycerides, LDL-C, high-density lipoprotein cholesterol [HDL-C]), estimated glomerular filtration rate (eGFR), smoking status, hypertension treatment status, aspirin therapy status, statin prescription status, and statin dosing intensity at the time of visit.

Patients with documented ASCVD (defined using ICD-10 codes, Supplemental Table) at the first visit during the study period or with documented allergy to statins at any time point were excluded from the study. Patients were defined as eligible for primary prevention statin therapy if at any primary care or cardiology visit during the study period they belonged to one of the following, mutually exclusive primary prevention statin eligible groups, defined based on the 2018 AHA/ACC Guideline on the Management of Blood Cholesterol: (1) age between 20 to 75 years old, with LDL-C ≥190 mg/dL; (2) age between 40 to 75, with diabetes mellitus (DM) and with LDL-C between 70 to 189 mg/dL; (3) age between 40 to 75, with LDL-C between 70 to 189 mg/dL and with estimated 10-year ASCVD risk ≥20% based on the Pooled Cohort Equations; (4) age between 40 to 75, with LDL-C between 70 to 189 mg/dL and with 10-year ASCVD risk ≥7.5% but <20%; and (5) age between 40 to 75, with LDL-C between 70 to 189 mg/dL, with 10-year ASCVD risk ≥5% but <7.5%, and with an ASCVD risk enhancer present. The following risk enhancers were extracted from the EHR: family history of premature ASCVD (using ICD-10 codes), primary hypercholesterolemia (most recent LDL > 160mg/dL), hypertriglyceridemia (triglycerides ≥ 175 mg/dL based on three determinations in the preceding five years), chronic kidney disease (CKD) based on eGFR < 60, inflammatory conditions (including autoimmune conditions and human immunodeficiency virus infection, as specified using ICD-10 codes provided in Supplemental Table), and high-sensitivity C-reactive protein (hs-CRP) ≥ 2 mg/L. Other risk enhancers were not included either due to very low numbers present in the EHR extract (high-risk race/ethnicity, premature menopause / pregnancy-associated conditions, elevated lipoprotein (a) (Lp(a)), and elevated apolipoprotein B (apoB)), or due to being unable to capture accurately using structured EHR data (metabolic syndrome [due to absence of waist circumference information] and ABI <0.9). Patients with age younger than 20 years or older than 75 years were not included in the study.

Statin prescription is defined as a patient having a statin listed in the active medication field (which capture medications that patients are actively prescribed at the time of the visit) during at least one visit where the patient met one of the above eligibility criteria. The type of statin therapy and dosage was also extracted to determine if the prescribe statin was high-intensity, moderate-intensity, or low-intensity, in accordance with the 2018 AHA/ACC Guideline.⁴

Descriptive statistics were used to assess baseline characteristics of patients eligible for primary prevention statin therapy, as well as prevalence of statin prescription during the study period for the entire sample and for subgroups based on sex and eligibility categories described above. Differences in continuous variables were analyzed using the Wilcoxon rank-sum test, and differences in categorical variables were evaluated using the Pearson’s chi-squared test.

Univariable and multivariable logistic regression models were used to estimate the association of patient demographic and clinical characteristics with statin prescription status. For each statin eligible group described above, sex and race/ethnicity were the main predictors of interest; multivariable analysis further adjusted for covariates including age, history of hypertension, smoking status, presence of risk enhancers, and statin eligibility criteria categories. All statistical analyses were performed using R software (version 4.0.2; R Core Team, 2020).

Results

Among 19,169 unique patients with primary care or cardiology visits during the study period, 7,624 were eligible for primary prevention statin therapy. After excluding those with documented statin allergy, 7,550 patients were included in the final analysis (Figure 1). Baseline characteristics are presented in Table 1. Statin therapy was prescribed in 3,994 (52.9%) of patients. In an unadjusted analysis, patients prescribed statin therapy were older, were more likely to be Hispanic, have a history of diabetes mellitus, have a history of hypertension, and were more likely to be smokers. Women were less likely to be prescribed statins. Among the risk enhancers included in this analysis, statin therapy was more likely to be prescribed in patients with hypertriglyceridemia and CKD, and less likely to be prescribed in patients with a family history of premature ASCVD (Table 1).

Figure 1.

Study Flow Diagram.

Table 1.

Baseline characteristics

Variable	All Patients (n = 7,550)	Statin Prescribed (n = 3,994)	Statin Not Prescribed (n = 3,556)	P value
Age in years, median (IQR)	63.0 (57.0 – 66.0)	65.0 (60.0 – 70.0)	60.0 (52.0 – 66.0)	<0.001
Women	4956	2563 (51.7%)	2393 (48.3%)	0.005
Men	2594	1431 (55.2%)	1163 (44.8%)
White	1004	523 (52.1%)	481 (47.9%)	<0.001
Black	957	442 (46.2%)	515 (53.8%)
Asian	198	111 (56.1%)	87 (43.9%)
Other/unknown	5391	2918 (54.1%)	2473 (45.9%)
Hispanic	4916	2724 (55.4%)	2192 (44.6%)	<0.001
Not Hispanic	2634	1270 (48.2%)	1364 (51.8%)
ASCVD risk score, median (IQR)	16.8% (8.4 – 25.3)	21.9% (12.8 – 31.1)	12.2% (6.8 – 17.7)	<0.001
Diabetes mellitus	2795	2060 (73.7%)	735 (26.3%)	<0.001
Hypertension	3778	2277 (60.3%)	1501 (39.7%)	<0.001
Smoker	1827	1019 (55.8%)	808 (44.2%)	0.005
Family history of premature ASCVD	427	190 (44.5%)	237 (55.5%)	<0.001
Primary hypercholesterolemia	722	360 (49.9%)	362 (50.1%)	0.09
†Hypertriglyceridemia	986	676 (68.6%)	310 (31.4%)	<0.001
‡Chronic Kidney Disease	729	475 (65.2%)	254 (34.8%)	<0.001
Inflammatory conditions	210	101 (48.1%)	109 (51.9%)	0.18
High-sensitivity C-reactive protein ≥ 2	1307	715 (54.7%)	592 (45.3%)	0.16
ASCVD risk enhancers
0	4225	2124 (50.3%)	2101 (49.7%)	<0.001
1+	3325	1870 (56.2%)	1455 (43.8%)

ASCVD = atherosclerotic cardiovascular disease; IQR = interquartile range.

Risk enhancers not included due to insufficient data: metabolic syndrome, conditions specific to women, high-risk race/ethnicity, Lp(a) ≥50, apoB ≥130, and ABI <0.9.

^*Low-density lipoprotein >160.

^†Triglycerides TG ≥175 on 3 determinations.

Among the ACC/AHA eligibility groups, statin therapy was most likely to be prescribed in patients with diabetes mellitus (73.6%) and least likely to be prescribed in patients with ASCVD risk score ≥5% but <7.5% with a risk enhancer present (18.7%). Diabetic women were significantly less likely to be prescribed a statin compared to diabetic men (71.6% vs. 77.2%, respectively; p=0.001). There were no statistically significant gender differences in statin prescription in other eligibility groups. Of note, high-intensity statin therapy was prescribed in only 58% of patients with LDL-C ≥190 mg/dL (Table 2).

Table 2.

Statin prescription based on ACC/AHA eligibility groups for primary prevention

Variable	Eligible Patients	Statin Prescribed	High* Intensity Statin	Moderate† Intensity Statin	Women Prescribed Statin	Men Prescribed Statin	P value
LDL ≥190 mg/dL‡	194	112 (57.7%)	65 (58.0%)	44 (39.2%)	92/152 (60.5%)	20/42 (47.6%)	0.14
Diabetes mellitus†	2751	2024 (73.6%)	1070 (52.9%)	871 (43.0%)	1273/1778 (71.6%)	751/973 (77.2%)	0.001
ASCVD risk ≥ 20%§	1270	769 (60.6%)	359 (46.7%)	360 (46.8%)	440/720 (61.1%)	329/550 (59.8%)	0.64
ASCVD risk ≥ 7.5% & < 20%§	2891	1006 (34.8%)	383 (38.1%)	565 (56.2%)	688/1954 (35.2%)	318/937 (33.9%)	0.50
Without RE∥	1899	600 (31.6%)	226 (37.7%)	335 (55.8%)	393/1249 (31.5%)	207/650 (31.8%)
With RE∥	992	406 (40.9%)	157 (38.7%)	232 (57.1%)	295/705 (41.8%)	111/287 (38.7%)
ASCVD risk ≥ 5% & < 7.5%‡ with RE∥	444	83 (18.7%)	27 (32.5%)	47 (56.6%)	70/352 (19.9%)	13/92 (14.1%)	0.21
Total	7550	3994	1904	1887	2563	1431

ACC = American College of Cardiology; AHA = American Heart Association; LDL = low-density lipoprotein; RE = risk enhancer; all other abbreviations as in Table 1.

^*High intensity statin: atorvastatin 40 to 80 mg, rosuvastatin 20 to 40 mg.

^†Moderate intensity statin: atorvastatin 10 to 20 mg, rosuvastatin 5 to 10 mg, simvastatin 20 to 40 mg, pravastatin 40 to 80 mg, lovastatin 40 mg, fluvastatin 80 or 40 mg BID, pitavastatin 3 to 4 mg.

^‡Inclusion: age 20 to 75. Inclusion: age 40 to 75, LDL ≥70 and <190 mg/100 ml, no DM.

^§Inclusion: age 40 to 75, LDL< 190 mg/100 ml.

^¶Risk enhancers included: family history of premature ASCVD, LDL ≥160, chronic kidney disease, inflammatory diseases, persistently elevated triglycerides ≥175, high sensitivity C-reactive protein ≥2.

In a multivariable model, after adjustment for patient characteristics, statin therapy was most likely to be prescribed in patients ≥65 years (OR 2.26, 95% CI 2.01–2.53), Hispanic patients (OR 1.25, 95% CI 1.12–1.39), patients with hypertension (OR 1.36, 95% CI 1.23–1.51), and smokers (OR 1.19, 95% CI 1.06–1.35). Women were significantly less likely to be prescribed statin therapy (OR 0.79, 95% CI 0.71–0.88). Black patients were less likely to be prescribed statin therapy, though this difference did not reach statistical significance (OR 0.85, 95% CI 0.69–1.04) (Table 3).

Table 3.

Patient Characteristics Associated with Statin Prescription for Primary Prevention.

Variable	% Prescribed Statin	Unadjusted OR (95% CI)	Adjusted OR (95% CI)*	P value
Age (years)
<65	1822/4267 (42.7%)	1 [Reference]	1 [Reference]	NA
65–75	2172/3283 (66.2%)	2.62 (2.38 – 2.88)	2.26 (2.01 – 2.53)	<0.001
Men	1431/2594 (55.2%)	1 [Reference]	1 [Reference]	NA
Women	2563/4956 (51.7%)	0.87 (0.79 – 0.96)	0.79 (0.71 – 0.88)	<0.001
White	481/1004 (47.9%)	1 [Reference]	1 [Reference]	NA
Black	442/957 (46.2%)	0.93 (0.78 – 1.11)	0.85 (0.69 – 1.04)	0.10
Asian	111/198 (56.1%)	1.39 (1.02 – 1.89)	1.09 (0.71 – 1.53)	0.62
Other/unknown	2918/5391 (54.1%)	1.28 (1.12 – 1.47)	1.10 (0.95 – 1.28)	0.21
Not Hispanic	1270/2634 (48.2%)	1 [Reference]	1 [Reference]	NA
Hispanic	2724/4916 (55.4%)	1.33 (1.21 – 1.47)	1.25 (1.12 – 1.39)	<0.001
Hypertension
No	1710/3746 (45.6%)	1 [Reference]	1 [Reference]	NA
Yes	2277/3778 (60.3%)	1.81 (1.65 – 1.98)	1.36 (1.23 – 1.51)	<0.001
Smoker
No	2975/5723 (52.0%)	1 [Reference]	1 [Reference]	NA
Yes	1019/1827 (55.8%)	1.16 (1.05 – 1.30)	1.19 (1.06 – 1.35)	0.004
Risk Enhancers
0	2124/4225 (50.3%)	1 [Reference]	1 [Reference]	NA
1+	1870/3325 (56.2%)	1.27 (1.16 – 1.39)	1.29 (1.12 – 1.49)	<0.001
Eligibility Criteria
ASCVD 5% – 7.5% with RE	83/444 (18.7%)	1 [Reference]	1 [Reference]	NA
ASCVD 7.5% – 20% without RE	600/1899 (31.6%)	2.01 (1.56 – 2.61)	1.78 (1.34 – 2.39)	<0.001
ASCVD 7.5% – 20% with RE	406/992 (40.9%)	3.01 (2.31 – 3.96)	2.06 (1.57 – 2.73)	<0.001
ASCVD ≥ 20%	769/1270 (60.6%)	6.66 (5.14 – 8.72)	3.42 (2.56 – 4.61)	<0.001
Diabetes mellitus	2024/2751 (73.6%)	12.1 (9.42 – 15.65)	8.76 (6.74 – 11.52)	<0.001
Low-density lipoprotein ≥ 190 mg/dL	112/194 (57.7%)	5.91 (4.10 – 8.62)	4.44 (3.04 – 6.52)	<0.001

Abbreviations: CI = confidence interval; NA = not applicable; OR = odds ratio; all other abbreviations as in Table 1 and 2.

^*Adjusted odds ratios were obtained from a multivariable logistic regression with statin prescription as the outcome variable and the above baseline characteristics as independent variables.

Discussion

Our study provides insights into contemporary practice for primary prevention statin prescription for eligible patients, and highlights potential gaps and disparities for guideline-concordant therapy. In our study, the prevalence of statin prescription among statin-eligible patients after the 2018 AHA/ACC Update on the Guideline of Management of Blood Cholesterol was 52.9%, similar to that reported from studies prior to this guideline update.^5,6,9,10 We also demonstrate significant variability in statin prescription based on statin eligibility group: among the five statin eligible groups, patients with diabetes had the highest prevalence of statin prescription, whereas the new category of statin eligible patients with 10-year ASCVD risk ≥5% but <7.5% and with risk enhancers had the lowest prevalence of statin prescription. Women also had lower prevalence of statin prescription compared to men, mainly driven by the subgroup with diabetes. To the best of our knowledge, this is the amongst the first study to assess gaps in guideline-recommended primary prevention statin prescription since the released of 2018 AHA/ACC Guideline update, with the clear clinical implication that substantial opportunities remain to improve primary prevention statin therapy, especially in high-risk groups.

Our finding that women are less likely to be prescribed primary prevention statin therapy is consistent with prior studies.^10–13 For example, an NHANES study from 2002 to 2004 showed that in patients with diabetes, women with were 38% less likely than men to be prescribed statins.¹¹ More recently, an analysis of patients with diabetes enrolled in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study showed that statin use was 11% lower for white women, and 12.4% lower for black women, when compared with white men.¹⁰ Two additional studies from the Patient and Provider Assessment of Lipid Management (PALM) registry also found that women were less likely than men to receive guideline-recommended statin therapy, especially for patients with diabetes and those with 10-year ASCVD risk ≥ 7.5%.^9,12 It is possible that these sex-based disparities are in part driven by patient preference or lower estimated ASCVD risk in women.¹⁴ However, recent studies have also shown that women with myocardial infarction (MI) are less likely to received guideline-recommended treatment and have worse outcomes compared to men, highlighting the urgent need to reduced sex-based disparities across the continumm of cardiovascular care.^15,16

Prior studies also found underuse of statin therapy in Black and Hispanic patients.^7,12,17,18 In our study population, there was a lower rate of statin prescription in Black patients that did not reach statistical significance, while Hispanic patients were significantly more likely to be prescribed statins compared to non-Hispanic patients. Since our study included a predominantly Hispanic patient population with access to primary care at an academic medical center, we may be less likely to uncover race/ethnity-based disparities. Despite this, we found substantial opportunities to improve guideline-concordant statin prescription and dosing, including for high-risk patient populations such as those with LDL-C ≥190 mg/dL. While current guidelines call for shared decision-making for initiating primary prevention statin therapy, there remains a paucity of evidence for best approaches to address patient and provider-level barriers to implementation. Interventions such as EHR-based decision support models that utilize decision aids that remind providers to prescribe statins for eligible patients have had only limited impact,^19–21 highlighting a broader need to identify more effective, scalable, population-based approaches for ASCVD risk reduction.²²

Our study has several limitations. The use of EHR data limits our ability to incorporate all risk enhancers into the analysis, including those specific to women such as preeclampsia and early menopause. However, this would suggest that more women in the overall study population might have been eligible for primary prevention statin therapy. Similarly, EHR data can be prone to coding errors, particularly those relying on ICD-10 administrative coding such as family history of premature ASCVD, which may lead to both over- and under-estimates.²³ EHR data is also unlikely to capture patient-reported side effects such as myalgia, although recent studies show that many of these patients should still be rechallenged with statins, as muscle-related side effects may often be due to the “nocebo” phenomenon.^24,25 Separately, the Pooled Cohort Equations were not caliberated for US Hispanic patients who are often of heterogeneous heritage, and can over- or under-estimate risk.²⁶ Finally, our sampling is from an academic medical center affiliated ambulatory care network in New York City serving a predominantly Hispanic patient base, which may limit the external validity and generalizability of our findings.

Despite these limitations, we provide contemporary evidence that there are substantial gaps in guideline-concordant primary prevention statin therapy, with persistent sex-based disparities. Future studies will need to further elucidate patient, clinician, and health-system level barriers underlying such disparities and identify effective, scalable interventions for ASCVD risk reduction.