Lack of a substantive effect of insurance and the national US payment system on the relative distribution of surgical cases among hospitals in the State of Iowa: A retrospective, observational, cohort study

Abstract

Study objectiveOur aim was to quantify the extent to which the distribution of patients among payers and changes to the payers' policies has influenced the market of surgery among hospitals in a relatively rural state. DesignRetrospective cohort study. SettingIowa Hospital Association data analyzed were from 2007 through 2016 for the N = 121 hospitals with at least one case performed that included a major therapeutic procedure. MeasurementsWe used five categories of payer (e.g., Medicare), five categories of patient age (e.g., 18 to 64 years), and three categories of patient residence location (e.g., neither from the county of the hospital nor from a county contiguous to the county of the hospital). Main resultsSorting hospitals in descending sequence of numbers of surgical cases, depending on year, the top 10% of hospitals performed 58.4% to 59.2% of the cases. Increases in numbers of cases among patients with commercial insurance increased the heterogeneity among hospitals in numbers of surgical cases (P < 0.0001). However, the magnitude of the effect was very small, with an estimated relative marginal effect on the overall Gini index of only 0.9% ± 0.2% (SE). Increases in numbers of cases of patients with Medicare insurance reduced the heterogeneity in numbers of cases among hospitals (P < 0.0001), but also with very small magnitude (−0.9% ± 0.2%). In contrast, factors encouraging patient travel contributed to larger hospitals becoming larger, and smaller hospitals becoming smaller (3.9% ± 0.7%, P < 0.0001). ConclusionsWe found the absence of a substantive effect of insurance and national US payment systems on the relative distribution of surgical cases among hospitals. Anesthesia groups should focus on payer and payment reform in terms of their effects on payment rates (e.g., average payment per relative value guide unit), not on their potential effects on hospital caseloads.