1. Data and Method
1) Data and Measures
This study employs quality of care and patient satisfaction as organizational performance. To measure each hospital’s overall quality of care, we adopt the Overall Quality Star Rating provided by the Center for Medicare & Medicaid Services (CMS) each year. The score is calculated as a weighted average for five sub-measure groups: mortality, safety, readmission, patient experience, and timely and effective care. The rating spans from 1 star to 5 stars. The level of patient satisfaction with a hospital is measured by an overall star rating from the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey, which is also provided by the CMS every year. This survey offers scores of 29 items to assess the patient experience of each hospital. We measure patient satisfaction with an overall rating of 1 to 5 stars.
Patient engagement is measured as an indicator variable, taking a value of one if a hospital has an established patient and family advisory council (PFAC) and taking a value of zero otherwise. The data comes from the American Hospital Association (AHA) Annual Hospital Survey. Hospitals’ physical and financial resources are measured as the logged values of the total fixed assets and the logged values of the net patient revenues, respectively (Barney, 1991; Short et al., 2002). Data used for both measures are from the RAND Hospital Data. Human resources are measured as the number of full-time equivalent (FTE) physicians, dentists, and FTE registered nurses (Su et al., 2009; Lee et al., 2023). Both variables are normalized using the total number of beds in each hospital, and the data are from the AHA survey. To measure the social and institutional resources, we adopted a percentage of the non-white population in each county. By adopting this measure, we can capture the level of social input from the minority population to the hospitals.
Our analyses also use multiple hospital- and county-level variables to control heterogeneity. Table 1 shows the measures of each variable and the data sources. Hospital size is measured as the number of beds in each hospital. Whether a hospital is a major or minor teaching affiliation organization, the shares of patients covered by Medicaid or Medicare programs and the Case Mix Index (CMI) are controlled. Regarding county-level differences, we measure the community demand for financial assistance with healthcare services as the logged value of the unreimbursed and uncompensated care costs. We also include the total population, percentage of individuals older than 65, income level, and uninsured rate in each county. The population and income data are from the U.S. Census and the U.S. Bureau of Economic Analysis, respectively.
Table 1. Variables and Measure