Abstract
Area Deprivation Index (ADI) is a marker of neighborhood deprivation. This study investigates utilization of coronary bare-metal stent (BMS) and drug-eluting stent (DES) in Medicare patients across hospitals with varying ADI. Data were abstracted using Diagnosis-Related Group (DRG) codes 249 (BMS without major complications or comorbidities [MCC]), 246, and 247 (DES with and without MCC, respectively) from the 2011–2012 Medicare Provider Utilization and Payment Data Inpatient File, which was linked to American Hospital Association data (to determine bed size, location, ownership, teaching status), and ADI for each hospital zip code was obtained. Hospitals were divided into quintiles using ADI values: Quintile 1 (privileged) to Quintile 5 (deprived). Logistic regression was conducted to determine odds ratios (ORs) for DES utilization across ADI quintiles. There were 313,739 discharges with DRG codes 246 (52,839), 247 (203,928), and 249 (56,972). DES utilization was lower in the deprived quintile, irrespective of teaching status. It was lower in larger hospitals and hospitals with more annual stent discharges, urban locations and nongovernment not-for-profit institutes. Lower odds of DES utilization were found in Quintile 2 (OR–0.9, 95% confidence interval [CI] 0.87–0.93, P < 0.001), Quintile 3 (OR–0.89, 95% CI 0.86–0.92, P < 0.001), and Quintile 4 (OR–0.95, 95% CI 0.92–0.98, P = 0.001) versus Quintile 1 and there was no difference in utilization of DES in Quintile 5 (OR–1.01, 95% CI 0.98–1.04, P = 0.6) versus Quintile 1. Significant differences exist in DES utilization in a large, uniformly insured cohort based on neighborhood deprivation.
Introduction
C
As the role of DES use has accelerated, comparatively less attention has been given to the changes in relevant reimbursement and payment policies. 4 Factors that determine the decision to use a DES versus a BMS include physician choice, comorbidities, lesion anatomy, insurance status, 5 patient preference, race, 6 and income. 7 Socioeconomic disadvantage is a key determinant of health and is at least partially responsible for inequalities in health care delivery. 8,9
Area Deprivation Index (ADI) is a comprehensive marker of neighborhood socioeconomic disadvantage. 9 Composite area-based deprivation indices have been utilized to monitor and trend health care outcomes. 10 –12 Some measures of neighborhood disadvantage include: median family income, percentage of population below poverty level, level of education, and employment rates. Indices such as ADI provide a comprehensive net score using such multiple variables of neighborhood advantage and disadvantage for a particular zip code or geographic area. For example, disadvantaged neighborhoods based on ADI have been shown to have higher mortality and readmissions. 9,13
This study examines the impact of stent choice (DES versus BMS) in Medicare patients based on ADI.
Methods
Study cohort
The Centers for Medicare & Medicaid Services released information regarding hospitals receiving Medicare Inpatient Prospective Payment System payments for the 100 most frequently billed Medicare Severity Diagnosis-Related Group (MS-DRG) for fiscal year (FY) 2011 and 2012 (October 1, 2010 to September 30, 2012). Data were abstracted regarding MS-DRG codes: 246 (percutaneous cardiovascular procedure with DES with Major Complication or Comorbidity [MCC] or 4+ vessels/stents), 247 (percutaneous cardiovascular procedure with DES without MCC), and 249 (percutaneous cardiovascular procedure with non-DES without MCC). Data on MS-DRG 248 (percutaneous cardiovascular procedure with non-DES with MCC) were not released because it was not one of the 100 most frequently billed MS-DRGs. There were 168,023 and 153,979 discharges from 1436 and 1478 hospitals for FY 2011 and 2012, respectively. Data from MS-DRG codes 246 and 247 were pooled together to represent DES discharges.
ADI
ADI was determined for each hospital based on zip code. 14 Hospital zip codes were individually verified with zip code tabulation area codes. ADI is composed of 17 coefficients based on the 2000 census data. 9 Lower ADI values indicate more privileged status and higher ADI values indicate greater neighborhood deprivation.
There were 1350 and 1391 hospitals with unique zip code tabulation areas in FY 2011 and 2012, respectively. ADI information was not available for 34 hospitals in FY 2011 and 41 hospitals in FY 2012. Hospitals in the database were stratified in quintiles based on ADI, with Quintile 1 being the most privileged and Quintile 5 being the most deprived. FY 2011: Quintile 1 (Range: −92.7–81.7), Quintile 2 (Range: 81.8–94), Quintile 3 (Range: 94.1–101.4), Quintile 4 (Range: 101.5–107.1) and Quintile 5 (Range: 107.2–131.3). FY 2012: Quintile 1 (Range: −92.7- +82), Quintile 2 (Range: 82.1–94), Quintile 3 (Range: 94.1–101.1), Quintile 4 (Range: 101.2–107) and Quintile 5 (Range: 107.1–131.3).
Hospital characteristics
Hospital bed size, ownership status, teaching status, and location (urban/rural; state; census region [northeast, midwest, south, west]) were abstracted from the American Hospital Association database. Hospitals were divided into quartiles based on bed size (FY 2011: Quartile 1 ≤ 190, Quartile 2 191–289, Quartile 3 290–428, Quartile 4 ≥ 429. FY 2012: Quartile 1 ≤ 185, Quartile 2 186–284, Quartile 3 285–425, Quartile 4 > 425) and annual institutional coronary stent discharges (FY 2011: Quartile 1 ≤ 41, Quartile 2 42–84, Quartile 3 85–145, Quartile 4 ≥ 145. FY 2012: Quartile 1 ≤ 38, Quartile 2 39–75, Quartile 3 76–131, Quartile 4 > 131).
Statistical analysis
Coronary stent discharges (DES and BMS) were compared across ADI quintiles and hospital characteristics by chi-square (categorical variables). Categorical variables were presented as percentages. P ≤ 0.05 was considered statistically significant. Univariate logistic regression was conducted. Significant covariates were entered in a stepwise fashion in a multivariable logistic regression model, with receipt of DES being the dependent variable. Covariates in the final multivariate logistic regression model included: hospital location (urban/rural and census region), teaching affiliation, hospital ownership, bed size, and annual institutional coronary stent discharges. Covariates were displayed as odds ratios with 95% confidence intervals. Statistical analyses were performed using STATA 11 (StataCorp LP, College Station, TX).
Results
There were a total of 313,739 patients with MS-DRG codes 246 (52,839 discharges), 247 (203,928 discharges) and 249 (56,972 discharges). Of these, 256,767 (81.84%) discharges were DES and 56,972 (18.16%) discharges were BMS placement. There were 260,900 discharges with stent placement (DES/BMS) without MCC; 78.16% received DES and 21.84% received BMS.
Hospitals in the most deprived neighborhoods (higher ADI) used fewer DES compared to hospitals in less deprived neighborhoods (lower ADI) (Quintile 1–5: 83%, 81.4%, 80.7%, 81.7% and 82.2%, P < 0.001) (Table 1). This trend was consistent regardless of teaching affiliation of hospitals, urban location, varying population densities across geographic census regions (north, midwest, west, and south), and in hospitals owned by nongovernment not-for-profit entities and government-owned nonfederal institutions. Hospital characteristics across ADI quintiles and utilization of DES and BMS are shown in Table 2.
BMS, bare-metal stent; DES, drug-eluting stent; MCC, major complications or comorbidities.
Total institutional discharges: Quartile 1 (≤190 for 2011 and ≤185 for 2012 discharges); Quartile 2 (191–289 for 2011 and 186–284 for 2012 discharges); Quartile 3 (290–428 for 2011 and 285–425 for 2012 discharges); Quartile 4 (≥429 for 2011 and >425 for 2012 discharges).
Bed size: Quartile 1 (≤41 for 2011 and ≤38 for 2012 discharges); Quartile 2 (42–84 for 2011 and 39–75 for 2012 discharges); Quartile 3 (85–145 for 2011 and 76–131 for 2012 discharges); Quartile 4 (>145 for 2011 and >131 for 2012 discharges).
BMS, bare-metal stent; DES, drug-eluting stent; Govt, government.
Hospitals located in the most deprived neighborhoods (Quintile 5) had a higher proportion of complicated DES discharges (MS-DRG code 246 [DES with MCC or 4+ vessels/stents]) compared to less deprived neighborhoods (Quintile 1–4): 21.6% compared to 20.1%, 19.8%, 20.4%, and 20.5% in Quintiles 1–4 respectively (P < 0.001) (Table 1).
Univariate logistic regression showed lower odds of DES utilization in Quintiles 2–5 compared to Quintile 1 (Table 3). The adjusted multivariate logistic regression model demonstrated lower odds for utilization of DES in Quintiles 2–4 compared to Quintile 1.
Adjusted models for hospital location (urban/rural and census region), teaching status, bed size, hospital ownership, and stent volume.
Boldface denotes statistical significance.
ADI, Area Deprivation Index.
Discussion
This study investigates coronary stent discharges in a large, 2-year Medicare cohort across areas of varying neighborhood deprivation. As neighborhood deprivation worsens, there is lower use of DES (Table 1). This trend persists in Quintiles 2–4 compared to Quintile 1 after adjusting for hospital characteristics.
Neighborhood deprivation has been associated with increased incidence of coronary artery disease, 15 cancer, 16 and acquired immunodeficiency syndrome. 17 Despite the impact of neighborhood-level influences on individual health and health outcomes, little research focuses on health care resource utilization in deprived and privileged neighborhoods. Neighborhood deprivation has a differential effect on members of the community. 18 Additional factors of family median income, extent of comorbid health conditions, level of education, racial composition, sociocultural beliefs, and health care awareness impact individual health and neighborhood deprivation. Additionally, neighborhoods differ in access to health care facilities, 19 availability of recreational spaces, 20 and availability, cost, and consumption of healthy foods. 21
This study demonstrates a higher proportion of complicated DES discharges (MS-DRG code 246 [DES with MCC or 4+ vessels/stents]) in the most deprived neighborhoods (Quintile 5) compared to less deprived neighborhoods (Quintiles 1–4): 21.6% compared to 20.1%, 19.8%, 20.4% and 20.5% in Quintiles 1–4, respectively (P < 0.001) (Table 1). This is suggestive of a higher prevalence of comorbid health conditions and/or multi-vessel coronary interventions in hospitals located in deprived neighborhoods in the study cohort. A higher incidence of disease (diabetes mellitus, coronary heart disease, stroke, colorectal cancer and/or depression), mortality, and increased health care costs have been demonstrated in areas of deprivation. 22 Surprisingly, the present study demonstrates a trend of lower utilization of DES in areas of deprivation (statistically significant in Quintiles 2–4 compared to Quintile 1 using multivariate logistic regression) despite a higher proportion of complex MS-DRG code 246 discharges (indicative of a population with multiple comorbid conditions) in a “like insurance” national cohort. Following is a discussion of possible physician and patient factors that may be responsible for this paradoxical relationship of lower DES usage despite higher disease complexity.
Physician factors
The study team hypothesizes that the following perceptions on the physicians' side may affect their decision to choose BMS over DES: (1) implicit physician bias based on sex, race, and economic status in deprived areas; (2) perceptions about patients' adherence to dual antiplatelet therapy (DAPT); and (3) patients' lack of access to primary care physicians in deprived areas. In relation to choosing DES, this bias may stem from the fact that nonadherence to DAPT predisposes the patient to stent thrombosis (ST). ST (acute and long-term) is the main concern with DES usage. 3 The PREMIER study revealed higher risk of mortality even in the first 6 months after DES related to ST. 23 Prior studies have shown that increased prevalence of mental health disorders, 24 and drug and alcohol abuse 25 in deprived areas could impact clinical decision making in view of adherence to DAPT. The study team notes that data for the present study were obtained a few years after 2008, by which time concerns about ST in the general population were settled. By this period it was realized that ST was mainly a risk to those patients who were not adherent to DAPT.
Studies have shown that the socioeconomic status of the patient affects physicians' clinical decision making. 26 Lower-income patients with ST elevation myocardial infarction are less likely to receive an angiogram within 24 hours as compared to their higher-income counterparts. 27 Other studies have demonstrated longer wait times for primary care visits, short encounter times, and poor post-visit patient satisfaction scores in deprived areas, reflecting potential physician hesitance to care fully for deprived patients. 28
Patient factors
In the new age of shared decision making, a patient from a more deprived neighborhood may not be able to meaningfully participate because of a lower literacy level. These patients may not be able to fully understand the pros and cons of DES versus BMS, use DAPT as prescribed, value the importance of close post-procedure and long-term follow-up or have the resources for procuring medications or follow-up. Self-pay patients may not be able to visit consultants in addition to their primary care physician. This can be addressed by increasing provider awareness of care of the stented patient.
There was no statistically significant difference between DES discharges in Quintile 5 compared to Quintile 1 after accounting for confounders using multivariate logistic regression. Increased prevalence of comorbid conditions such as diabetes, higher representation of African Americans and Hispanics, increasing complexity of coronary artery disease, and higher incidence of acute ST elevation myocardial infarction in deprived areas could explain a slight increase in utilization of DES in Quintile 5 compared to Quintiles 2–4 (still lower than Quintile 1), resulting in statistical nonsignificance in multivariate regression. Other studies have demonstrated disparities in care with no commensurate increase in rate of carotid endarterectomy despite a higher rate of symptomatic carotid disease in patients from deprived socioeconomic groups. 29
Future direction
Heart disease is the leading cause of death in the United States. Socioeconomic inequality in cardiovascular care is a bane to American society. With the implementation of the Affordable Care Act, lower socioeconomic patients are increasingly encountered. To ensure equal care, physicians should avoid bias based on deprivation and indulge in the shared decision-making process with the patient. Increasing patient information avenues such as free outreach clinics and support groups will help cardiac patients understand their disease and avail themselves of the newest cardiovascular advances.
Limitations
This large, nationwide, single-payer insurance cohort study had limitations. The Medicare data set consisted of patients classified as having an inpatient admission, hence trends in outpatient stent utilization could not be delineated. This study utilizes administrative claims data and has limitations encountered by such data as previously noted. 30 Information was not available on individual coronary anatomy, lesion length, individual patient demographics, comorbid conditions, and concomitant noninvasive testing. Additionally, the Medicare claims used for this study were derived solely from fee-for-service patients older than 65 years of age; thus, the results may not apply to younger patients, or to patients covered by other types of insurance.
Deprivation indices assess neighborhood deprivation and this study could not account for individual-level deprivation. ADI is an indirect marker of individual deprivation. Despite this, the present study observed a significant decrease in the utilization of DES in more deprived areas after accounting for hospital-level characteristics. Despite attenuation of neighborhood disadvantage and access to health care after accounting for individual-level characteristics, a significant residual association remains. 19
Conclusion
DES utilization was lower in deprived areas after accounting for hospital level characteristics: teaching affiliation, bed size, not-for-profit status, population density, and location. Further studies are required to understand individual deprivation and its impact on health care outcomes.
Footnotes
Author Disclosure Statement
The authors declared no conflicts of interest with respect to the research, authorship, and/or publication of this article. The authors received no financial support for the research, authorship, and/or publication of this article.