The Effect of Hospital-Physician Integration On Hospital Costs

This thesis evaluates whether hospitals that are integrated with physician practices have lower or higher costs than hospitals that are not integrated, using a large sample of U.S. hospitals from 2000-‐2013. Some economic theories predict that vertical integration lowers costs and other theories predict higher costs. I therefore conduct an analysis to test these predictions in the context of hospital-‐ physician integration. I use a variety of econometric methods, including regression analysis of cost functions using both Cobb-‐Douglas and translog specifications. I estimate fixed effects regressions to control for unobserved time-‐constant factors specific to individual hospitals. I also conduct matching analyses to account for potential endogeneity. The results demonstrate that hospital-‐physician integration is not associated with lower hospital costs. Instead, depending on the specification, the results show that integrated hospitals have costs that are higher than or equivalent to the costs of non-‐ integrated hospitals. Analyses that include other types of vertical relationships also do not show any consistent effect of these organizational forms on hospital costs. In addition, I find no consistent effects of vertical integration or other organizational forms on the quality of hospital care. These findings suggest that any increases in costs that may stem from hospital-‐physician integration are not due to higher quality care. The findings also have potential implications for public policy. Although vertical integration is sometimes viewed as a way to stem rising healthcare costs, the evidence suggests that this may not be the case. 1. Introduction Healthcare spending accounted for 18 percent of U.S. GDP in 2016 (U.S. Centers for Medicare and Medicaid, 2018). In addition to comprising a large share of GDP, healthcare spending is continuing to rise (U.S. Centers for Medicare and Medicaid, 2018). Factors that affect healthcare spending thus merit detailed investigation. In 2016, hospital care accounted for 32 percent of U.S. healthcare spending, and physician and clinical services accounted for 20 percent of healthcare spending (U.S. Centers for Medicare and Medicaid, 2018). Although hospitals and physician practices have generally been separate entities in the past, today many hospitals directly employ their physicians. This could affect spending on healthcare if it leads to changes in hospital costs or revenues. In this thesis, I examine whether hospital-‐physician integration affects hospital costs. The results may have implications for public policy related to healthcare. The integration of hospitals with physician practices has increased substantially in the past 15 to 20 years. Between 2010 and 2016, the percentage of primary care physicians working in an organization owned by a hospital or hospital system rose by 57 percent (Fulton, 2017). From 1999-‐2003, the number of physicians and dentists employed by community hospitals was roughly stable, but this began to trend upward in 2004, increasing by 56 percent from 2003 to 2014 (American Hospital Association, 2016). As of 2015, an analysis by Avalere found that 38 percent of physicians were employed by hospitals (Physicians Advocacy Institute, 2016). This trend appears to be a response in part to pressure on hospitals and other healthcare providers to contain rising