Researchers have extensively investigated provider- and market-level variations in Medicare spending and utilization patterns, but patterns of variation among the commercially insured are less well understood. In this analyses, we assess the hospital, market, provider and population characteristics that may affect variation in prices paid by a commercial health plan for inpatient hospital stays across Texas hospitals. We use stay-level Blue Cross Blue Shield of Texas (BCBSTX) inpatient claims data, American Hospital Association Survey data, Texas Health Care Information Collection (THCIC) inpatient discharge data and publicly available data from the American Community Survey (ACS), U.S. News and World Report (USNWR) hospital rankings, and hospital quality scores from the Centers for Medicare and Medicaid Services (CMS). We first use allowed amounts and diagnosis-related groups (DRGs) in 2013 BCBSTX inpatient claims data to create a hospital-level casemix-adjusted price index (“hospital price”) for each of 383 hospitals in Texas with at least 10 admissions. We then measure the level of variation in hospital prices, and use regressions to examine the association of hospital prices with hospital and market characteristics, where markets are defined as hospital service areas (HSAs). For each HSA, we measure the socioeconomic characteristics of the resident population based on the ACS, BCBSTX market share (the share of the population under age 65 enrolled in a commercial BCBSTX plan). Also, for each hospital we calculated a Herfindahl-Hirschman Index (HHI) using a two-step process—in the first step we calculate HHIs based on patient residence and, in the second step, we calculate hospital HHIs based on the residence of their patient flow. We find a very high level of variation in commercial hospital prices (coefficient of variation of 0.33). In the regression analysis, hospitals with more beds, teaching status, and located in areas with higher wage index tend to receive higher prices; while hospitals with higher BCBSTX market share, and serving higher poverty areas had lower prices. Hospitals in less-competitive markets, i.e. those higher HHIs, tended to receive significantly lower prices. Specialty hospitals like rehabilitation hospitals and psychiatric hospitals were associated with lower hospital prices. Surprisingly, quality and capacity measures—including CMS quality scores, USNWR regional or national rankings, casemix index, having level I trauma center—were not associated with hospital prices Our results are consistent with previous studies that have found that commercial hospitals prices for inpatient stays tend to be higher for large teaching hospitals, and hospitals in high-wage, high-income areas. We find no evidence, however, for quality scores being important drivers of commercial prices. The HHI results suggest that hospital negotiating leverage does not depend on whether it is physically located in a geographically isolated area.
