Sensitivity Analysis , Cabg, And Pci - 1308 Words

Sensitivity Analysis-Colectomy, CABG, and PCI Multiple analyses were conducted to test the sensitivity of price regression results: 1) HHI measurement by different definitions and levels; 2) GLM regression using different distribution families; and 3) non-linearity of age, HHI, and HMO penetration. HHI Measured by Admissions vs Inpatient Days The main GLM regression models included a hospital admission HHI where the total number of admissions in a hospital is measured as a proxy of the hospital’s market share. Another way to measure a hospital’s market share is to calculate the total inpatient day share of the hospital. Table 9 shows the GLM equations with the admission HHI vs. the inpatient days HHI for all samples. This sensitivity analysis results were consistent with the results from the main GLM regression. HHI at Hospital, System, Network and System or Network levels HHI can also be measured at different levels: 1) hospital; 2) system; 3) network; and 4) system or network level. Table 10 shows the different specifications using HHI at different levels. Similarly, this sensitivity analysis supports the results from the main GLM regressions. Coefficients of different types of health systems changed slightly at the third decimal place when different HHI variables were included. It is worthwhile to pay attention to the systematic differences in HHI coefficients. The magnitude of coefficients became larger from model 1 to model 4 for colectomy, CABG and PCI samples.