359 - Neonatal Bed Availability Before and After Private Equity Acquisition in the United States

Publication Number: 359.5888

Background: Private equity (PE) hospital ownership has increased over the past 20 years. PE acquisition may lead to expansion of profitable services lines including newborn and neonatal intensive care. How PE hospital acquisition impacts the neonatal bed supply has not been studied.

Objective: To assess newborn and neonatal intensive care unit (NICU) bed availability in hospitals acquired by PE compared to matched controls using causal methods.

Design/Methods: We identified PE acquired US acute care hospitals in American Hospital Association and Centers for Medicare and Medicaid Services (CMS) data from 1995 to 2021 using the PE Stakeholder Project PE Hospital Tracker, Pitchbook, the CMS change of hospital ownership database, and internet review. We matched PE hospitals to five non-acquired controls without replacement using characteristics the year prior to acquisition with exact matching on year, region, and bed-size category, and nearest neighbor matching on pediatric and neonatal bed number and pre-acquisition trends of these bed types. We imputed missing bed data using the prior survey-year. We then performed a stacked differences-in-differences (DiD) analysis to evaluate newborn and NICU beds in the 7 years before and after acquisition. We used longitudinal Poisson and logistic mix-effects models to examine differences in beds and unit closures respectively, with year fixed effects, grouping on matched set, and DiD estimate as the interaction term between time and PE-acquisition. Parallel trends were confirmed through the matching process.

Results: 219 PE hospitals were matched with 1095 controls. PE hospitals had 11.67 newborn beds and 4.26 NICU beds compared to 11.78 newborn beds and 4.14 NICU beds in matched controls during the year prior to acquisition (Table 1). Compared to all possible controls, there were no significant differences in matched characteristics between PE hospitals and matched controls with standardized differences well below 0.1 (Table 1). After PE acquisition, newborn beds decreased (adjusted IRR 0.93, 95% CI 0.90-0.95) and continued newborn services were less likely (adjusted OR 0.38, 95% CI 0.24-0.61) in PE-acquired hospitals compared to matched controls (Figures 1 and 2). There were no significant differences in NICU bed number or closures in PE-acquired hospitals compared to matched controls (Figure 2).

Conclusion(s): PE-acquisition decreased newborn bed availability and increased newborn service closures, while maintaining NICU services. Work studying how this discordance in neonatal care impacts perinatal outcomes is warranted.

Table 1. PE-acquired hospitals compared to all possible controls and matched controls in the United States, 1995-2021
* indicates p <0.05

We performed exact matches on year, region, and bed-size category in the year prior to PE acquisition and nearest neighbor matching on pediatric and neonatal bed number and pre-acquisition trends of these bed types over the previous 5 years. To assess match quality and trade-offs, we utilized standardized differences (weighted differences based on variability and scale) pre- and post-match. The pre-match control cohort includes all possible control hospitals. We utilized a stringent target standardized difference of <0.1, equivalent to a less than 10% difference in the factor.

Figure 1. Trend of newborn and neonatal intensive care unit (NICU) beds before and after private equity acquisition in private equity owned hospitals compared to matched controls
Private Equity hospitals were matched to 5 non-acquired controls using characteristics the year prior to acquisition with exact matching on year, region, and bed-size category, and nearest neighbor matching on pediatric and neonatal bed number and pre-acquisition trends of these bed types. Time was standardized to time of private equity acquisition. Due to missingness, these analyses include 215 PE hospitals and 1075 matched controls.

Figure 2. Difference-in-difference models examining the effect of private equity acquisition on newborn and neonatal intensive care unit (NICU) beds and continuation and services
Stacked difference-in-differences analyses were performed to evaluate newborn nursery and NICU beds and closures in the 7 years before and after NICU acquisition. For bed analyses, we used longitudinal Poisson mix-effects model with year fixed effects, grouping on matched set, and DiD estimate as the interaction term between time and PE-acquisition. For closures, we used a longitudinal logistic mixed effect model controlling for the same additional factors to predict the likelihood of continuation of the services lines of interest. Private Equity hospitals were matched to 5 non-acquired controls using characteristics the year prior to acquisition with exact matching on year, region, and bed-size category, and nearest neighbor matching on pediatric and neonatal bed number and pre-acquisition trends of these bed types. Due to missingness, these analyses include 215 PE hospitals and 1075 matched controls. Abbreviations: aIRR – adjusted incidence rate ratio, aOR – adjusted odds ratio