Table 2. Changes in Hospital Income and Use Measures After Private Equity Acquisition.
| Measure | Hospitals | Differential change | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Acquired hospitals (n = 204) | Control hospitals (n = 532) | Unadjusted, No.a | Adjusted, No. (%) [95% CI]b | P value | Corrected P valuec | |||||
| Before private equity | After private equity | Change | Before private equity | After private equity | Change | |||||
| Net income per y, $ | 8 527 119 | 12 861 680 | 4 334 561 | 7 655 125 | 10 092 820 | 2 437 695 | 1 896 866 | 2 302 391 (27.0) [956 660 to 3 648 123] | .001 | .009 |
| Total charge per inpatient day, $ | 5789 | 7766 | 1978 | 5583 | 6928 | 1345 | 633 | 407 (7.0) [296 to 518] | <.001 | <.001 |
| Emergency charge to cost ratio | 3.81 | 5.52 | 1.71 | 4.00 | 5.03 | 1.02 | 0.69 | 0.61 (16.0) [0.48 to 0.73] | <.001 | <.001 |
| Total charge to cost ratio | 4.17 | 5.02 | 0.85 | 3.90 | 4.38 | 0.48 | 0.37 | 0.31 (7.4) [0.26 to 0.37] | <.001 | <.001 |
| Case mix index | 1.42 | 1.47 | 0.05 | 1.36 | 1.41 | 0.05 | 0.00 | 0.02 (1.4) [0.01 to 0.02] | .001 | .007 |
| Medicare’s share of discharges, % | 40.3 | 36.8 | −3.5 | 39.1 | 37.1 | −2.0 | −1.56 | −0.96 (−2.4) [−1.45 to −0.46] | <.001 | .002 |
| Medicaid’s share of discharges, % | 13.2 | 12.2 | −1.0 | 15.2 | 14.3 | −0.9 | −0.07 | −0.16 (−1.2) [−0.86 to 0.53] | .64 | >.99 |
| Total discharges per y, No. | 8948 | 9181 | 233 | 8504 | 8353 | −151 | 384 | 98 (1.1) [−54 to 250] | .21 | >.99 |
Private equity–acquired hospitals were matched to controls at event time 0 (time of acquisition). R Package MatchIt was used to generate at most 8 controls per acquired hospital. We used nearest neighbor matching on total beds and exact matching on year, ownership, region (Northeast, Midwest, South, and West), and teaching hospital status. The unadjusted model refers to the mixed-effects model, which included a random intercept for the matched group and for the provider group, with no covariates. Values indicate means before and after private equity acquisition of acquired and control hospitals and were calculated using the unadjusted model.
The adjusted model included a random intercept term for the matched group and for the provider group and adjusted for calendar year, case mix index, and total hospital beds. Case mix index was not included as a covariate when evaluated as an outcome. Percentage differential change was calculated by dividing the adjusted differential change by the preacquisition mean among private equity–acquired hospitals.
Bonferroni correction used for multiple comparisons testing.