There is substantial hospital-level variation in health care quality in the United States.1 While there are many tools that provide information to patients about hospital quality and performance, the degree of alignment between hospital performance and market share remains unclear.2 The Centers for Medicare and Medicaid Services developed the Hospital Compare website, now referred to as Hospital Care Compare, to enable beneficiaries to make more informed decisions about where to seek care.3 This study assessed the association between publicly reported hospital performance scores in Hospital Care Compare and hospital market share.
This cross-sectional ecological study used hospital-level, risk-adjusted performance measures and patient volume for the following 3 conditions or procedures available in the Centers for Medicare and Medicaid Services’ Hospital Care Compare database: acute myocardial infarction (AMI), coronary artery bypass graft (CABG), and hip and/or knee replacement. They were selected to exemplify nonelective, semielective, and mostly elective conditions or procedures, respectively. Performance measures included 30-day mortality rate and 30-day readmission rate for AMI and CABG and 90-day complication rate for hip and/or knee replacement for the 2019 report year, reflecting performance in July 1, 2015, through June 30, 2018. Each hospital’s condition- or procedure-specific volume was obtained from the 2020 report year, reflecting patient volumes in July 1, 2016, through June 30, 2019.4 The performance score period preceded the volume period by 1 year to account for how consumers have access only to past, as opposed to contemporary, quality measurements. Market share was calculated by dividing the hospital’s volume by the sum of the volume of all hospitals in the hospital referral region (HRR). Hospital and HRR characteristics were derived from the Hospital Care Compare database, the Dartmouth Atlas,5 the American Community Survey, and the US Decennial Census.6 This research did not meet the National Institutes of Health definition of human subject research as specified in 45 CFR part 46, and was therefore exempt from institutional review board approval. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Fractional regressions via generalized linear models with binomial family and logit links were used to model hospital market share for each condition or procedure while adjusting for hospital and HRR characteristics. All estimates reflected marginal effects along with their 95% CIs. Variables adjusted for included number of beds (in tercile), HRR concentration, hospital type (acute care vs critical access), hospital ownership, teaching status, region, HRR population size (in tercile), HRR percentage of the non-White population (in tercile), HRR percentage of the population in rural area (in tercile), HRR percentage of the population with less than a high school education (in tercile), and HRR percentage of the population below the federal poverty level (in tercile). The reference level for tercile variables was the first (lowest) tercile. Terciles were calculated by measure. Two-tailed P < .05 was considered statistically significant. Analysis was conducted using RStudio, version 1.2 (RStudio PBC). The eMethods in the Supplement describes additional methodologic considerations.
Of 4930 hospitals examined, 1993 (40.4%) were included for AMI readmission and 2160 (43.8%) for mortality; 962 hospitals (19.5%) were included for CABG readmission, 971 (19.7%) for mortality, and 2635 (53.4%) for hip and/or knee complication. These hospitals have diverse characteristics (Table 1). The eMethods in the Supplement describes exclusion criteria in detail.
For AMI readmission, each percentage point increase in a hospital’s risk-adjusted complication rate was associated with a 1.7 percentage point decrease in its market share (95% CI, −3.10 to −0.25 percentage points; standardized coefficient, −0.16; P = .02). There was no significant association between a hospital’s performance score and market share for AMI mortality (−0.85 percentage point; 95% CI, −2.07 to 0.38 percentage points; standardized coefficient, −0.09; P = .17), CABG readmission (−0.49 percentage point; 95% CI, −2.49 to 1.52 percentage points; standardized coefficient, −0.04; P = .63), and CABG mortality (−1.24 percentage points; 95% CI, −4.40 to 1.93 percentage points; standardized coefficient, −0.06; P = .44). For hip and/or knee replacement, each percentage point increase in a hospital’s risk-adjusted complication rate was associated with a 4.2 percentage point decrease in its market share (95% CI, −6.56 to −1.88 percentage points; standardized coefficient: −0.25; P < .001) (Table 2).
The findings of this cross-sectional ecological study suggest that better hospital performance score was associated with larger market share for hip and/or knee replacement and AMI procedures but not for CABG. The effect size was greater for the elective procedure (hip and/or knee replacement) than for the nonelective procedure (AMI), suggesting that patients may be more able to choose their facility for elective procedures.
For CABG, consumers may have less ability to choose their facility based on quality reports owing to the fewer number of facilities available with CABG capacity. While the observed association between hospital performance and market share for hip and/or knee replacement and AMI is encouraging, the effect sizes were modest, highlighting a continued need to better align where patients receive care and the quality of care they receive. This may be achieved by improving patients’ awareness of hospital quality ratings and their ability to choose, by increasing the capacity or geographic accessibility of high-performing centers, or by advancing hospital quality through value-based models of care.
Although we cannot draw causal inferences, these data provide useful initial information about the potential role of public reporting in channeling patients to higher-quality clinicians. Other study limitations include hospital-level analysis in lieu of patient-level analysis and the use of a limited set of performance measures for fee-for-service Medicare beneficiaries only. Further research using patient-level data, other performance measures, and younger patient populations would provide additional insights.
Accepted for Publication: August 17, 2021.
Published: October 21, 2021. doi:10.1001/jamanetworkopen.2021.30353
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Oddleifson A et al. JAMA Network Open.
Corresponding Author: August Oddleifson, BA, Yale University School of Medicine, 333 Cedar St, New Haven, CT 06510 (august.oddleifson@yale.edu).
Author Contributions: Mr Oddleifson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Oddleifson, Xu, Du, Desai.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Oddleifson, Desai.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Oddleifson, Du.
Obtained funding: Oddleifson.
Administrative, technical, or material support: Oddleifson, Desai.
Supervision: Du, Desai.
Conflict of Interest Disclosures: Dr Xu reported having a contract through Yale University from the Centers for Medicare and Medicaid Services to develop and maintain performance measures outside the scope of the current work. Dr Desai reported having a contract with the Centers for Medicare and Medicaid Services to develop and maintain performance measures used for public reporting and pay-for-performance programs and receiving research grants and consulting fee from Amgen, Astra Zeneca, Boehringer Ingelheim, Cytokinetics, MyoKardia, Novartis, Relypsa, and SCPharmaceuticals. No other disclosures were reported.
Funding/Support: This study was supported by grant T35AG049685 from the National Institute on Aging, which is funded by the National Institutes of Health (NIH).
Role of the Funder/Sponsor: The National Institute on Aging of the NIH had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.