Skip to main content
Download PDF

Use of hospitals in the New York City Metropolitan Region, by race: how separate? How equal in resources and quality?

BMC Health Services Research volume 22, Article number: 1021 (2022) Cite this article

Abstract

Background

Although racial and ethnic minorities disproportionately use some hospitals, hospital-based racial and ethnic composition relative to geographic region and its association with quality indicators has not been systematically analyzed.

Methods

We used four race and ethnicity categories: non-Hispanic white (NHW), non-Hispanic black (NHB), Hispanic, and Asian/Pacific Islander/Alaskan Native/American Indian (API/AIAN), as well as a combined non-NHW category, from the 2010 (latest year publicly available) Medicare Institutional Provider & Beneficiary Summary public use file for 84 hospitals in the New York City region. We assessed the relative distribution of race and ethnicity across hospitals grouped at different geographic levels (region, county, hospital referral region [HRR], or hospital service areas [HSA]) using the dissimilarity index. Hospital characteristics included quality star ratings, essential professional services and diagnostic/treatment equipment, bed size, total expenses, and patients with dual Medicare and Medicaid enrollment. We assessed Spearman’s rank correlation between hospital-based racial and ethnic composition and quality/structural measures.

Results

Dissimilarity Index decreases from region (range 30.3–40.1%) to county (range 13.7–23.5%), HRR (range 10.5–27.5%), and HSA (range 12.0–16.9%) levels. Hospitals with larger non-NHW patients tended to have lower hospital ratings and higher proportions of dually-enrolled patients. They were also more likely to be safety net hospitals and non-federal governmental hospitals.

Conclusions

In the NYC metropolitan region, there is considerable hospital-based racial and ethnic segregation of Medicare patients among non-NHW populations, extending previous research limited to NHB. Availability of data on racial and ethnic composition of hospitals should be made publicly available for researchers and consumers.

Peer Review reports

Introduction

Studies have shown that Black and Hispanic patients are disproportionately hospitalized in some hospitals and, in some cases, are highly concentrated in a smaller number of hospitals—thereby, creating de facto separate hospital care by race and ethnicity, or segregation [1,2,3,4,5,6,7,8,9,10,11,12,13]. Other work, primarily for Black patients, has shown that hospitals with a higher concentration of minority patients may have poorer performance on quality measures [14,15,16]. These hospitals may have systematic differences in organizational structure, revenue, staffing, services, and other resources.

Multiple interrelated factors may contribute to differences in patient volume and health outcomes across racial and ethnic groups. Patients may select a hospital due to proximity which, in turn, could be linked to residential segregation. Choice of hospitals may depend on the type and extent of health insurance coverage and that, in turn, could be linked to employment and income opportunities that may be influenced by discrimination and segregation. Patients and family may voluntarily choose a particular hospital due to language, culture, or religious affinity; [17, 18] however, even there, these hospitals may have been founded in response to discrimination and exclusion of patients or staff from other facilities [19]. In this sense, “segregation” may be an important contributor to the observation of “separate care.” For this reason, we will use the term “segregation” to describe “separate care.” This usage is consistent with how the term segregation is used to describe the similar phenomenon in housing and education [20,21,22,23,24,25,26,27,28].

Despite growing recognition of and interest in racial and ethnic segregation in US healthcare, the problem is not well studied due to numerous methodological challenges. First, race and ethnicity are inconsistently, unreliably, and variably recorded in administrative data, particularly for Hispanics [29,30,31]. These issues, however, could be potentially mitigated using different algorithms [32,33,34]. For example, an algorithm developed by the Research Triangle Institute (RTI) improves the accuracy of coding for Hispanic and Asian or Pacific Islander Medicare beneficiaries by 159% and 45%, respectively [29].

In addition, there exists a wide range of approaches to quantitatively describe racial and ethnic segregation with no consensus on what is the best measure and whether a single measure can capture its extent [20,21,22,23]. For example, “minority-serving hospitals,” a commonly used term in studying hospital health disparities, is an example of an absolute measure of segregation, as it is based on a threshold (e.g., 90th or 75th percentile) of the direct proportion of minority patients in a given hospital. The dissimilarity index (DI), a measure of segregation or imbalance or unevenness used in demography, housing, and education, is less often used in healthcare disparity research [24]. Moreover, in estimating the extent of segregation in healthcare, the measure of spatial resolution is not inconsequential. For example, variation of racial and ethnic composition across counties may differ from that measured across census tracts. This issue is known as a modifiable area unit problem, as the measures are influenced by the shape and scale of the aggregation unit [35]. Finally, there is a wide range of outcome variables for evaluating the impact of racial and ethnic segregation, ranging from health service utilization, mortality, and hospital quality, further complicating the approach to evaluation [12, 36].

In this analysis, we examine the extent of racial and ethnic segregation of Medicare patients seen in hospitals across various geographic levels in the New York City (NYC) metropolitan region. In addition, we examine the extent to which the racial and ethnic composition of a hospital is associated with hospital structural characteristics (e.g., bed size, access to essential services and technologies, and hospital expenditure) and a hospital’s publicly reported measures of quality. We focus on the NYC metropolitan region—the most populous and one of the more diverse and residentially segregated metropolitan regions in the Northeastern United States. By concentrating on one region, we can more accurately evaluate the impact of community size, the organization of the healthcare market, and effects of closures and consolidations (which may lead to 2 very different facilities reporting under one combined hospital identity) on hospital segregation.

Methods

Data sources

Our analyses used data from multiple public use files (PUFs) from the Centers for Medicare and Medicaid Services (CMS). Hospital-level information on patient race and ethnicity was obtained from the 2010 (latest year publicly available) Institutional Provider and Beneficiary Summary (IPBS) PUF, which included 100% claims for the Medicare fee-for-service (FFS) population from hospitals and other institution types (n = 49,413) [37]. We also obtained the 2021 overall hospital quality star ratings from the Hospital General Information file provided by CMS (n = 5382) [38]. We used data from 2021 to address concerns over the previous methodology in use in 2010 and due to inaccessibility of the earlier data [39,40,41]. Information related to the hospital’s hospital referral regions (HRRs) and hospital service areas (HSAs) was obtained from the Dartmouth Atlas Project for the year 2010 (n = 4893) [42]. HSAs are essentially a group of ZIP codes that represent local markets for hospital care, and many HSAs contain only one 1 or a few number of hospitals. HRRs represent regional markets for tertiary care and contain at least one hospital performing major cardiovascular or neurosurgical procedures [43]. In addition, hospital structural characteristics (e.g., bed size, access to essential services and technologies, and hospital expenditure) were obtained from the American Hospital Association (AHA) annual survey for the year 2010 (n = 5915) [44]. We identified hospitals existed in all sources (n = 4893), while excluding hospitals located in the US territories (n = 79). We further limited analyses to 84 hospitals within the NYC metropolitan region that covers 9 counties, 6 HRRs, and 37 HSAs (Fig. 1). These included partial HRRs and HSAs as our study area of interest was the 9 counties within the state of New York (Supplemental Fig. S1).

Fig. 1
figure 1

Locations of hospitals (dots) in New York City metropolitan area (shaded area) with different geographic boundaries: Counties, Health Referral Regions (HRRs), and Health Service Areas (HSAs)

Full size image

Measures

Hospital racial/ethnic composition

The IPBS PUF summarized the number of beneficiaries who utilized services (e.g., stays, admissions, visits, episodes) of the institutional provider, and the numbers of beneficiaries for the following five groups using the enhanced race and ethnicity designation according to the RTI race code. We report on 4 race and ethnic categories both separately and combined: non-Hispanic White (NHW), non-Hispanic Black (NHB), Hispanic, Asian/Pacific Islander, Alaskan Native/American Indian (API/AIAN), and a combined category designated as minorities which includes the NHB, Hispanic, API/AIAN and the small (< 1% of all beneficiaries on average) categorized as “other” in the RTI race code and who would otherwise be excluded from the analysis due to sample size issues; this category, in effect, includes all groups that are not NHW (non-NHW). We calculated the proportion of each group by dividing the number of beneficiaries in each subgroup by the total number of beneficiaries. Subgroups with beneficiary numbers less than 11 are suppressed in the public data release to protect patient privacy. A hospital could have one or more subgroups with missing beneficiary counts. We calculated the difference between the total beneficiary counts and the sum of all available race and ethnicity subgroups, and divided this difference by the total number of subgroups with missing data to replace the missing number beneficiary counts.

Dissimilarity index

We used DI to assess the evenness of the distribution of racial and ethnic group across hospitals relative to the overall distribution of that racial and ethnic group in the hospital’s geography, similar to the method used for studying the US public school system [24]. For a particular racial/ethnic group x in a geographic area of G, \({DI}_{x}\) is derived according to the following equation: \({DI}_{x}=\sum_{i=1}^{n}\frac{{p}_{i}|{m}_{i}-M|}{2PM(1-M)}\), where \({p}_{i}\) is the number of patients in hospital i, P is the number of patients in area G, \({m}_{i}\) is the percentage of patients of racial/ethnic group x in hospital i, M is the percentage of patients of racial/ethnic x in area G, and area G has n hospitals (i = 1, 2,…n). We calculated DI for each racial/ethnic group (i.e., x = (NHW, NHB, Hispanic, API/AIAN, or Minorities)) at four different geographic levels (i.e., G = (NYC metropolitan area, counties, HRRs, or HSAs)).

Hospital characteristics

We included the following eight hospital characteristics. The CMS 5-start hospital overall rating is a composite measure of hospital quality based on individual indicators from five domains: mortality, safety of care, readmission, patient experience, and timely and effective care [38]. The rating is on a 0–5 scale with 5 being the highest rating, and was calculated according to the updated algorithm. From the AHA data on hospital services, we surveyed 5 physicians to identify services that would be most essential in the care of an adult medical/surgical patient in a community hospital (excluding services that could be appropriately performed elsewhere after discharge or if the patient could be reasonably transferred to a tertiary hospital). From their consensus ratings, we defined “essential professional services” as the total number of the following 7 services: Emergency Department, Adult cardiology services, Neurological services, Oncology services, Orthopedic services, Psychiatric child/adolescent services, and Palliative Care Program. We also defined “essential diagnostic, treatment and imaging” as the total number of the following 9 items: Adult diagnostic catheterization, Adult interventional cardiac catheterization, Adult cardiac electrophysiology, Hemodialysis, Optical Colonoscopy, Endoscopic ultrasound, Computed-tomography (CT) scanner, Magnetic resonance imaging (MRI), and Ultrasound. In addition, we calculated the relative proportion of ICU beds to total general beds, the total expense per hospital bed, and ownership (non-federal government (n = 16) or non-profit (n = 68)) [44]. We also included the proportion of patients with dual Medicare and Medicaid enrollment from the IPBS PUF as an indicator of socioeconomic disadvantage. Finally, according to the Agency for Healthcare Research and Quality (AHRQ) definition, we identified safety net hospital (n = 32) based on top quartile of hospitals ranked by the percentage of total discharges that are composed of uninsured patients or those with Medicaid [45].

Statistical analysis

We summarized patient racial and ethnic composition using box-plots. We reported the median and interquartile range (IQR) of DI for each race and ethnicity group at four geographic levels. We also reported the Spearman’s rank correlation coefficients between specific hospital-level racial/ethnic compositions (e.g., percent NHW) and hospital characteristics. We also compared hospital race/ethnicity composition and characteristics by ownership and safety net status using Wilcoxon rank-sum tests. The analyses were conducted using R (v4.0.2) with RStudio (v1.3.1073).

Results

For the entire NYC metropolitan area, the proportion of patients who were NHW, NHB, Hispanic, AAPI/AIAN, and minorities was 66.6%, 15.1%, 13.0%, 3.7%, and 33.4%, respectively. Racial and ethnic compositions at county-, HRR-, and HSA-level are shown in Fig. S2 in the Supplementary Information, and hospital characteristics at county-level are shown in Table 1. For the entire NYC metropolitan area, the DIs for NHW, NHB, Hispanic, and API/AIAN were 40.1%, 36.1%, 37.6%, and 30.3%, respectively (Table 2). A regional DI of 40.1% can be interpreted as the share of NHW patients (or combined minority patients) that would need to change hospitals to create a perfectly integrated region-wide hospital system. The range of median DI diminished to between 13.7% and 23.5% at the county level, between 10.5% to 27.5% at the HRR level, and between 12% and 16.9% at the HSA level (Table 2).

Table 1 Summary descriptive statistics of hospital characteristics at county level
Full size table
Table 2 Distribution of the Dissimilarity Index (median (interquartile range, IQR)) by racial/ethnic group by geographic level, New York City metropolitan area, 2010
Full size table

Across 84 hospitals in the NYC metropolitan area, the median percentage of patients who were NHW, NHB, Hispanic, API/AIAN, and racial/ethnic minorities was 66.1% (IQR 33.5%-83.5%), 12.6% (IQR 7.3%-28.9%), 9.1% (IQR 5.1%-28.9%), and 2.2% (IQR 1.5%-3.9%), respectively (Fig. 2).

Fig. 2
figure 2

Distributions of patient racial and ethnic compositions among 84 hospitals in the New York City metropolitan area. Note: NHW = non-Hispanic White, NHB = non-Hispanic Black, API/AIAN = Asian/Pacific Islander, Alaskan Native/American Indian, non-NHW = a combined category designated as minorities which includes the NHB, Hispanic, API/AIAN and the small (< 1% of all beneficiaries on average) categorized as “other” in the RTI race code

Full size image

The proportion of NHB was closely correlated with the proportion of Hispanic (Spearman’s correlation coefficient (ρ) = 0.73, p < 0.0001), while not significantly correlated with the proportion of API/AIAN (ρ = 0.15, p = 0.18). The proportion of NHW patients was significantly inversely correlated with other groups (Table 3). The racial and ethnic distribution of patients varied at HRR, county, and HSA levels. For example, the percentage of minority patients ranged from 7.4 to 26.1% in Suffolk county, 8.2 to 41.2% in Nassau county, 56.2 to 95.5% in the Bronx, 23.6 to 94.5% in Brooklyn, 38.3 to 83.8% in Queens, 15.1 to 95.4 in Manhattan, and 7.6 to 68.9% in Westchester county (see Table S1). Two hospitals a short distance apart could have marked differences in the percentage of NHBs (e.g., 23.3 vs 7.0% in the same part of Manhattan) or Hispanics (e.g., 36.1 vs. 8.3% in the same part of Manhattan) (Table S1).

Table 3 Correlations of race specific percentages at level of individual hospitals (n = 84) in New York City metropolitan area
Full size table

The proportion of NHB, Hispanic, API/AIAN, and combined racial/ethnic minority groups were all positively correlated with the proportion of patients with dual enrollment in Medicare and Medicaid (ρ ranged from 0.27 to 0.91, all p < 0.0001, Table 4), while the opposite was found for NHW (ρ = -0.91, p < 0.0001). Indicators of hospital structure/resources were in general not correlated with patient racial and ethnic composition (Table 4).

Table 4 Spearman’s correlation coefficients between proportion of race and ethnicity groups and measures of structure and quality of the hospital
Full size table

The proportion of NHB and Hispanic was negatively correlated with the overall hospital rating (ρ = -0.57 and -0.40, respectively, both p < 0.0001, Table 4), while the proportion of NHW patients was significantly positively correlated with the overall hospital rating (ρ = 0.55, p < 0.0001). The correlation between hospital rating and the proportion of API/AIAN was not significant (ρ = 0.01, p = 0.96). Safety net hospitals had a higher proportion of racial/ethnic minority patients or dual eligible patients, and lower overall hospital rating (Figure S3). Similar trends were seen in non-federal governmental hospitals (Figure S4).

Discussion

We found considerable racial and ethnic segregation among Medicare patients in hospitals in the NYC region. The segregation was present within county, HRR, and HSA geographic levels, although the unevenness of racial and ethnic distribution became progressively (and expectedly) less at each level. This segregation at the geographic level was accompanied by examples of more extreme disparities in racial and ethnic composition between some hospitals. Our analyses indicate that the uneven racial and ethnic composition at the hospital level was weakly associated with structural measures of healthcare service capacity but more strongly associated with publicly reported hospital quality rankings, safety net hospital status, and hospital ownership type.

These findings are consistent with previous work that has shown that there is racial segregation in where patients receive hospital care in the United States [7,8,9,10,11,12,13, 46]. The majority of that research has focused on segregation of the Black population. Moreover, these findings similarly find segregation associated with lower hospital quality [1,2,3,4,5,6, 47]. The current analysis adds to this literature by quantifying the extent of racial and ethnic segregation using a validated algorithm to classify race from administrative data [29]. These findings demonstrate that the association of these disparate racial distributions with measures of hospital quality extend beyond Black populations to other racial/ethnic minority group populations. We also build upon previous research by demonstrating the extent of racial and ethnic segregation in the most populous and one of the most ethnically and racially diverse metropolitan regions in the United States. Ongoing questions about the existence of structural racism throughout American medicine including within its most high profile publications underlie the importance of further quantifying and illustrating the extent of hospital segregation [48].

The dissimilarity we report is slightly lower than the levels of dissimilarity reported for residential communities, but similar to school segregation in the US [24, 25]. Nationally, residential segregation in metropolitan areas has decreased from DIs of 72.7 in 1980 to 55.2 in 2020 for the Black-White DI, from 50.2 to 45.3 for the Hispanic-White DI, and from 40.4 to 40.0 for the Asian-White DI [26]. In the New York-Jersey-City-White Plains metropolitan areas, higher residential segregation was found ranging from DIs of 79.1 for Black-White, 63.1 for Hispanic-White, and 49.5 for Asian-White in 2010 [25]. Public school segregation for Black and Hispanic students was about 32% nationally and 55% in metropolitan areas between 1995 and 2015 [24]. While similar to trends in public school segregation, the DI values we report decreased with smaller geographic units, disparities remain even at the finest geographic levels. Smaller geographic divisions may be more homogenous in race and ethnicity either due to segregationist policies or preferences (e.g., a predominantly ethnic community), and examining dissimilarity at different geographic levels is important to understanding its context. It is not surprising that healthcare is highly associated with where people reside and suggests that many of the underlying causes of residential, school, and hospital segregation are pervasive and shared.

Some of the underlying causes of segregation in healthcare that are not encountered in housing and school segregation are worth noting. In particular, the United States maintains a system of health insurance with eligibility conditioned on employer, age, income, service in the armed forces, and other factors. The design of insurance products (with payments of varying generosity, preferred networks, and physician participation or non-participation) provides further opportunity for sorting by income and race [49, 50]. For example, unlike hospitals, a small percentage of physicians choose not to participate in Medicare and a larger number place limits on new patients due to payment levels or other conditions; many more physicians limit Medicaid participation [51, 52]. Existing patient-physician relationships and physician-hospital affiliations will heavily influence where patients will be referred for hospitalization. Some patients will have no choice but to rely on safety net hospitals [53, 54]. Indeed, we found that AHRQ-definition safety-net hospitals had a higher median percentage of racial/ethnic minority Medicare patients [45, 55]. Longstanding provider selections made under limited coverage options may continue even after more choices become available through Medicare eligibility. Apart from features of the insurance market, patients may also intentionally choose healthcare settings due to geographic proximity, language, cultural or religious considerations [56, 57]. However, it is important to note that some of these healthcare organizations were created in response to exclusion and discrimination against specific patients and/or hospital staff such as physicians of specific races, ethnicities, or religions. As a result, the organization of healthcare and the structure of health insurance may lead to systemic segregation, which may affect where receive care, what resources will be available to treat them and, in turn, affect healthcare quality and patient outcomes. This healthcare landscape would suggest that insurance coverage of the population, including all-payor hospital rate regulation approaches such as that in Maryland, focused efforts to improve quality at low-performing hospitals, and public reporting of race and ethnicity by hospitals and health plans may be important and needed tools in a multifaceted approach to reduce and correct health facility segregation [14, 58, 59].

Our analysis has some limitations. We focused on segregation that might be observed between hospitals, and segregation that might occur within a facility is beyond the scope of our data. The race and ethnicity variable was obtained from Medicare fee-for-service data from 2010, and the quality outcome variables were obtained from the 2021 CMS Hospital General Information file.–We used the 2010 Medicare data because it was the latest hospital-level public use file available with the RTI classification of patient race and ethnicity. Ideally, we would have used quality ratings more proximate in time to 2010, but the methods of the hospital overall quality rating have been updated over time to overcome criticism of older methods. We used the more recent quality ratings because we found that the relative rating of the hospitals (n = 58 available for all 6 years) to be relatively stable in analyses of quality ratings from 2015–2020 (see Fig S5). This approach was carried out to provide the research team with the opportunity to test the use and limitations of these data to justify the cost and privacy safeguards necessary to perform the research with more recent all-payor datasets. Nevertheless, our analysis involves Medicare – the largest single payor in the United States. Although more recent data would be useful, recent data on the experience of New York area public and private hospitals during the SARS-CoV-2 pandemic would suggest that the racial and ethnic distributions of these hospitals have not substantively changed [60].

Furthermore, our analysis was limited to the NYC metropolitan area, thereby potentially limiting its generalizability. Our approach sought to better understand the data on a more granular level than would have been possible with a national focus. By focusing on the regional level, we were able to understand why no data (due to hospital consolidation) was available for some hospitals that were known to exist, why only limited data existed for some hospitals (due to closures), and why selected data elements were not available for some hospitals (due to exemption from reporting). Our nuanced, local understanding gave the research team greater confidence in the data, as well as an understanding of its limitations. Future research should consider how the consolidation of hospitals may potentially reduce and obscure differences in the racial and ethnic distribution of the unmerged entities.

Finally, our research relied on the RTI classification of race and ethnicity using data that is not consistently self-reported, may misclassify some individuals, and may aggregate some categories (e.g., among Hispanics or the API/AIAN category) [61]. Nonetheless, we believe this analysis advances previous research that may have used race and ethnicity variables that are often inconsistently and unreliably recorded in administrative data. Understanding the experience of smaller racial and ethnic subgroups is critical and may require other types of data and research approaches.

Conclusions

We report that segregation of Medicare patients in NYC metropolitan areas hospitals is pervasive, and associated with disparities in measures of hospital quality. Segregation was observed across NHB, Hispanic, and API/AIAN groups. The organization of healthcare and of health insurance may lead to systemic segregation, which may affect where patients receive care, which hospitals are used, which doctors are seen, what resources will be available to treat their conditions, and the outcomes of care. Additional research is needed with updated data, inclusive of other payors, and with a national focus to further our understanding of how care is segregated and how best to advance equity in health services for the American people.

Availability of data and materials

The study used several data sets from the public domain, including 1) the 2010 (latest year publicly available) Institutional Provider and Beneficiary Summary (IPBS) public use files PUF from the Centers for Medicare and Medicaid Services (CMS), available from https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/BSAPUFS/IPBS_PUF; 2) the 2021 overall hospital quality star ratings from the Hospital General Information file provided by CMS, available from https://data.cms.gov/provider-data/topics/hospitals/overall-hospital-quality-star-rating/; 3) the Dartmouth Atlas Project for the year 2010, available from https://data.dartmouthatlas.org/supplemental/#hospital, for information related to the hospital’s HRRs and HSAs; and 4) the American Hospital Association annual survey for the year 2010, available from https://www.ahadata.com/aha-annual-survey-database, for the following hospital-level information about hospital structural characteristics (e.g., bed size, access to essential services and technologies, and hospital expenditure). The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.

References

  1. Peterson K, Anderson J, Boundy E, Ferguson L, McCleery E, Waldrip K. Mortality Disparities in Racial/Ethnic Minority Groups in the Veterans Health Administration: An Evidence Review and Map. Am J Public Health. 2018;108:e1-11.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Zavala VA, Bracci PM, Carethers JM, Carvajal-Carmona L, Coggins NB, Cruz-Correa MR, et al. Cancer health disparities in racial/ethnic minorities in the United States. Br J Cancer. 2021;124:315–32.

    Article  PubMed  Google Scholar 

  3. Graham G. Disparities in cardiovascular disease risk in the United States. Curr Cardiol Rev. 2015;11:238–45.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Bock F, Stewart TG, Robinson-Cohen C, Morse J, Kabagambe EK, Cavanaugh KL, et al. Racial disparities in end-stage renal disease in a high-risk population: the Southern Community Cohort Study. BMC Nephrol. 2019;20:308.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  5. Mayeda ER, Glymour MM, Quesenberry CP, Whitmer RA. Inequalities in dementia incidence between six racial and ethnic groups over 14 years. Alzheimer’s Dementia. 2016;12:216–24.

    Article  PubMed  Google Scholar 

  6. Asch DA, Islam MN, Sheils NE, Chen Y, Doshi JA, Buresh J, et al. Patient and hospital factors associated with differences in mortality rates among black and white US medicare beneficiaries hospitalized with COVID-19 infection. JAMA Netw Open. 2021;4:e2112842.

    Article  PubMed  Google Scholar 

  7. Haider AH, Hashmi ZG, Zafar SN, Hui X, Schneider EB, Efron DT, et al. Minority trauma patients tend to cluster at trauma centers with worse-than-expected mortality: can this phenomenon help explain racial disparities in trauma outcomes? Ann Surg. 2013;258:572–81.

    Article  PubMed  Google Scholar 

  8. Sloane PD, Yearby R, Konetzka RT, Li Y, Espinoza R, Zimmerman S. Addressing systemic racism in nursing homes: a time for action. J Am Med Dir Assoc. 2021;22:886–92.

    Article  PubMed  Google Scholar 

  9. Rush B, Danziger J, Walley KR, Kumar A, Celi LA. Treatment in disproportionately minority hospitals is associated with increased risk of mortality in sepsis: a national analysis. Crit Care Med. 2020;48:962–7.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Howell EA, Egorova NN, Balbierz A, Zeitlin J, Hebert PL. Site of delivery contribution to black-white severe maternal morbidity disparity. Am J Obstet Gynecol. 2016;215:143–52.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Manuel JI. Racial/ethnic and gender disparities in health care use and access. Health Serv Res. 2018;53:1407–29.

    Article  PubMed  Google Scholar 

  12. Cole AP, Nguyen D-D, Meirkhanov A, Golshan M, Melnitchouk N, Lipsitz SR, et al. Association of care at minority-serving vs non–minority-serving hospitals with use of palliative care among racial/ethnic minorities with metastatic cancer in the United States. JAMA Netw Open. 2019;2:e187633.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Danziger J, de la ÁngelArmengolHoz M, Li W, Komorowski M, Deliberato RO, Rush BNM, et al. Temporal trends in critical care outcomes in us minority-serving hospitals. Am J Respir Crit Care Med. 2020;201:681–7.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Gaskin D, Zare H, Haider AH, LaVeist T. The quality of surgical and pneumonia care in minority-serving and racially integrated hospitals. Health Serv Res. 2016;51:910–36.

    Article  PubMed  Google Scholar 

  15. Jha AK, Orav EJ, Epstein AM. Low-quality, high-cost hospitals, mainly In South, care for sharply higher shares of elderly black, Hispanic, and medicaid patients. Health Affairs. 2011;10:1904–11.

    Article  Google Scholar 

  16. Jha AK, Orav EJ, Li Z, Epstein AM. Concentration and quality of hospitals that care for elderly black patients. Arch Intern Med. 2007;167:1177–82.

    Article  PubMed  Google Scholar 

  17. Halperin EC. The rise and fall of the American Jewish Hospital. Acad Med. 2012;87:610–4.

    Article  PubMed  Google Scholar 

  18. The development of a community-based integrated health care system for the San Francisco Chinese community.. The Free Library. 2007 [cited 2022 Feb 28]. Available from: https://www.thefreelibrary.com/The development of a community-based integrated health care system for...-a0161127951

  19. Baron J. The Mount Sinai Hospital–a brief history. Mt Sinai J Med. 2000;67:3–5.

    CAS  PubMed  Google Scholar 

  20. Acevedo-Garcia D, Lochner KA, Osypuk TL, Subramanian SV. Future directions in residential segregation and health research: a multilevel approach. Am J Public Health. 2003;93:215–21.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Massey DS, White MJ, Voon-Chin P. The dimensions of segregation revisited. Sociol Methods Res. 1996;25:172–206.

    Article  Google Scholar 

  22. White MJ. The measurement of spatial segregation. Am J Sociol. 1983;88:1008–18.

    Article  Google Scholar 

  23. Duncan OD, Duncan B. A methodological analysis of segregation indexes. Am Sociol Rev. 1955;20:210–7.

    Article  Google Scholar 

  24. When Is a School Segregated? Making Sense of Segregation 65 Years after Brown v. Board of Education. 2019 [cited 28 Feb 2022]. Available from: https://www.urban.org/research/publication/when-school-segregated-making-sense-segregation-65-years-after-brown-v-board-education

  25. Logan JR, Stults BJ. The Persistence of Segregation in the Metropolis: New Findings from the 2010 Census. 2011 [cited 28 Feb 2022]. Available from: https://s4.ad.brown.edu/Projects/Diversity/data/report/report2.pdf

  26. Logan JR, Stults BJ. Metropolitan Segregation: No Breakthrough in Sight. 2021 [cited 28 Feb 2022]. Available from: https://s4.ad.brown.edu/Projects/Diversity/Data/Report/report08122021.pdf

  27. Diez Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci. 2010;1186:125–45.

    Article  PubMed  Google Scholar 

  28. Kramer MR, Hogue CR. Is segregation bad for your health? Epidemiol Rev. 2009;31:178–94.

    Article  PubMed  Google Scholar 

  29. Eicheldinger C, Bonito A. More accurate racial and ethnic codes for Medicare administrative data. Health Care Financ Rev. 2008 Spring;29(3):27–42.

  30. Grafova IB, Jarrín OF. Beyond black and white: mapping misclassification of medicare beneficiaries race and ethnicity. Med Care Res Rev. 2021;78:616–26.

    Article  PubMed  Google Scholar 

  31. Jarrín OF, Nyandege AN, Grafova IB, Dong X, Lin H. Validity of race and ethnicity codes in medicare administrative data compared with gold-standard self-reported race collected during routine home health care visits. Med Care. 2020;58:e1-8.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Elliott MN, Morrison PA, Fremont A, McCaffrey DF, Pantoja P, Lurie N. Using the Census Bureau’s surname list to improve estimates of race/ethnicity and associated disparities. Health Serv Outcomes Res Method. 2009;9:69–83.

    Article  Google Scholar 

  33. Filice CE, Joynt KE. Examining race and ethnicity information in Medicare administrative data. Med Care. 2017;55:e170–6.

    Article  PubMed  Google Scholar 

  34. Adjaye-Gbewonyo D, Bednarczyk RA, Davis RL, Omer SB. Using the Bayesian Improved Surname Geocoding Method (BISG) to Create a Working Classification of Race and Ethnicity in a Diverse Managed Care Population: A Validation Study. Health Serv Res. 2014;49:268–83.

    Article  PubMed  Google Scholar 

  35. Openshaw S, Taylor P. The modifiable areal unit problem. Quantitative Geography: A British View. Eds N Wrigley and RJ Bennett. London: Routledge; 1981.

    Google Scholar 

  36. Kaye DR, Norton EC, Ellimoottil C, Ye Z, Dupree JM, Herrel LA, et al. Understanding the relationship between the centers for Medicare and Medicaid services’ hospital compare star rating, surgical case volume, and short-term outcomes after major cancer surgery: CMS star rating, surgical volume, and outcomes. Cancer. 2017;123:4259–67.

    Article  PubMed  Google Scholar 

  37. The Centers for Medicare & Medicaid Services. Institutional Provider & Beneficiary Summary PUF. 2013 [cited 28 Feb 2022]. Available from: https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/BSAPUFS/IPBS_PUF

  38. The Centers for Medicare & Medicaid Services. Overall hospital quality star rating. 2021 [cited 2022 28 Feb Feb]. Available from: https://data.cms.gov/provider-data/topics/hospitals/overall-hospital-quality-star-rating/

  39. Austin JM, Jha AK, Romano PS, Singer SJ, Vogus TJ, Wachter RM, et al. National hospital ratings systems share few common scores and may generate confusion instead of clarity. Health Aff. 2015;34:423–30.

    Article  Google Scholar 

  40. Bilimoria KY, Barnard C. An evolving hospital quality star rating system from CMS: aligning the stars. JAMA. 2021;325:2151.

    Article  PubMed  Google Scholar 

  41. Bilimoria KY, Barnard C. The new CMS hospital quality star ratings: the stars are not aligned. JAMA. 2016;316:1761.

    Article  PubMed  Google Scholar 

  42. Dartmouth Atlas Data. 2010 [cited 28 Feb 2022]. Available from: https://data.dartmouthatlas.org/supplemental/#hospital

  43. Appendix on the Geography of Health Care in the United States [Internet]. Dartmouth Atlas. [cited 13 June 2022]. Available from: https://data.dartmouthatlas.org/downloads/methods/geogappdx.pdf

  44. American Hospital Association Annual Survey Database. 2010 [cited 28 Feb 2022]. Available from: https://www.ahadata.com/aha-annual-survey-database

  45. Unintended Consequences: How New York State Patient and Safety-Net Hospitals are Shortchanged. Community Service Society, New York State Health Foundation; 2018 Jan. Available from: https://nyshealthfoundation.org/wp-content/uploads/2018/01/new-york-state-patients-safety-net-hospitals-jan-2018.pdf

  46. Largent EA. Public health, racism, and the lasting impact of hospital segregation. Public Health Rep. 2018;133:715–20.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Gaskin DJ, Spencer CS, Richard P, Anderson G, Powe NR, LaVeist TA. Do minority patients use lower quality hospitals? Inquiry. 2011;48:209–20.

    Article  PubMed  Google Scholar 

  48. Rivara FP, Bradley SM, Catenacci DV, Desai AN, Ganguli I, Haneuse SJPA, et al. Structural racism and JAMA Network open. JAMA Netw Open. 2021;4:e2120269.

    Article  PubMed  Google Scholar 

  49. Allen H, Gordon SH, Lee D, Bhanja A, Sommers BD. Comparison of utilization, costs, and quality of Medicaid vs subsidized private health insurance for low-income adults. JAMA Netw Open. 2021;4:e2032669.

    Article  PubMed  Google Scholar 

  50. Kaiser Family Foundation. How Much More Than Medicare Do Private Insurers Pay? A Review of the Literature. 2020 [cited 28 Feb 2022]. Available from: https://www.kff.org/medicare/issue-brief/how-much-more-than-medicare-do-private-insurers-pay-a-review-of-the-literature/

  51. Barnes H, Richards MR, Martsolf GR, Nikpay SS, McHugh MD. Association between physician practice Medicaid acceptance and employing nurse practitioners and physician assistants: a longitudinal analysis. Health Care Manage Rev. 2022;47:21–7.

    Article  PubMed  Google Scholar 

  52. Kaiser Family Foundation. How Many Physicians Have Opted-Out of the Medicare Program?. 2020 [cited 28 Feb 2022]. Available from: https://www.kff.org/medicare/issue-brief/how-many-physicians-have-opted-out-of-the-medicare-program/

  53. Chatterjee P, Qi M, Werner RM. Association of Medicaid expansion with quality in safety-net hospitals. JAMA Intern Med. 2021;181:590–7.

    Article  PubMed  PubMed Central  Google Scholar 

  54. Sarkar RR, Courtney PT, Bachand K, et al. Quality of Care at Safety-Net Hospitals and the Impact on Pay-for-Performance Reimbursement. Cancer. 2020;126:4584–92.

    Article  PubMed  Google Scholar 

  55. Sutton JP, Fingar KR, Washington R, Elixhauser A. Healthcare Cost and Utilization Project (HCUP): Characteristics of Safety-Net Hospitals, 2014. 2016 [cited 28 Feb 2022]. Available from: https://www.hcup-us.ahrq.gov/reports/statbriefs/sb213-Safety-Net-Hospitals-2014.jsp

  56. Smith H, Currie C, Chaiwuttisak P, Kyprianou A. Patient choice modelling: how do patients choose their hospitals? Health Care Manag Sci. 2018;21:259–68.

    Article  PubMed  Google Scholar 

  57. de Groot IB, Otten W, Dijs-Elsinga J, Smeets HJ, Kievit J, de Marang-vanMheen PJ, et al. Choosing between hospitals: the influence of the experiences of other patients. Med Decis Making. 2012;32:764–78.

    Article  PubMed  Google Scholar 

  58. Japinga M, McClellan M. Uniquely Similar: New Results from Maryland’s All-Payer Model and Paths Forward for Value-Based Care. New York: Millbank Memorial Fund; 2020. https://www.milbank.org/publications/uniquely-similar-new-results-from-marylands-all-payer-model-and-pathsforward-for-value-based-care/.

  59. Blumberg LJ, Holahan J, Karpman M, Elmendorf C. Characteristics of the Remaining Uninsured: An Update. 2018. https://www.urban.org/sites/default/files/publication/98764/2001914-characteristics-of-the-remaining-uninsured-an-update_2.pdf.

  60. Marcello RK, Dolle J, Grami S, Adule R, Li Z, Tatem K, et al. Characteristics and outcomes of COVID-19 patients in New York City’s public hospital system. Liu SY, editor. PLoS One. 2020;15:1–16.

    Google Scholar 

  61. Flanagin A, Frey T, Christiansen SL, Bauchner H. The reporting of race and ethnicity in medical and science journals: comments invited. JAMA. 2021;325:1049.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

This work was supported by the National Institute on Aging grant number P30AG028741 and K24AG062785-01A1.

Funding

This work was supported by the National Institute on Aging grant number P30AG028741 and K24AG062785-01A1.

Author information

Authors and Affiliations

  1. Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, NY, USA

    Bian Liu & Emma K. T. Benn

  2. Brookdale Department of Geriatrics and Palliative Medicine, Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Place, Box 1640, New York, NY, 10029, USA

    Katherine A. Ornstein, Julia L. Frydman, Amy S. Kelley & Albert L. Siu

  3. Geriatric Research, Education, and Clinical Center, James J Peters Veterans Affairs Medical Center, Bronx, NY, USA

    Amy S. Kelley & Albert L. Siu

Authors
  1. Bian Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Katherine A. Ornstein
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Julia L. Frydman
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Amy S. Kelley
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Emma K. T. Benn
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Albert L. Siu
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Bian Liu. The first draft of the manuscript was written by Albert L Siu, Bian Liu, Julia L. Frydman, and Katherine A. Ornstein, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Albert L. Siu.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Fig. S1.

The locations of hospitals (dots) in New York City metropolitan area (shaded area) with different geographic boundaries: Counties (red outlines), Health Referral Regions (HRRs; green outlines), and Health Service Areas (HSAs; blue outlines). Fig. S2. Distributions of patient racial/ethnic compositions at different level of geography level in the New York City metropolitan area. CNTY=county, HRR=health referral region, HSA=health service area. Fig. S3. Distributions of patient racial/ethnic compositions and hospital characteristics by top safety net hospital status. Fig. S4. Distributions of patient racial/ethnic compositions and hospital characteristics by hospital ownership. Fig. S5. Distribution of hospital overall rating during the years 2015-2020. Table S1. Proportion of patient race/ethnicity across 84 hospitals in the New York City Metropolitan area. (Minorities are the sum of non-Hispanic Black, Hispanics, Asian/Pacific Islander, Alaskan Native/American Indian, and other race/ethnicity groups that are not non-Hispanic White).

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Liu, B., Ornstein, K.A., Frydman, J.L. et al. Use of hospitals in the New York City Metropolitan Region, by race: how separate? How equal in resources and quality?. BMC Health Serv Res 22, 1021 (2022). https://doi.org/10.1186/s12913-022-08414-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12913-022-08414-3

Keywords

BMC Health Services Research

ISSN: 1472-6963

Contact us