SEARCH

SEARCH BY CITATION

Keywords:

  • Hospital-based skilled nursing facility;
  • prospective payment policy;
  • instrumental variables;
  • rehospitalization

Objective

To examine the effect of reductions in hospital-based (HB) skilled nursing facility (SNF) bed supply on the rate of rehospitalization of patients discharged to any SNF from zip codes that lost HB beds.

Data Source

We used Medicare enrollment records, Medicare hospital and SNF claims, and nursing home Minimum Dataset assessments and characteristics (OSCAR) to examine nearly 10 million Medicare fee-for-service hospital discharges to SNFs between 1999 and 2006.

Study Design

We calculated the number of HB and freestanding (FS) SNF beds within a 22 km radius from the centroid of all zip codes in which Medicare beneficiaries reside in all years. We examined the relationship between HB and FS bed supply and the rehospitalization rates of the patients residing in corresponding zip codes in different years using zip code fixed effects and instrumental variable methods including extensive sensitivity analyses.

Principal Findings

Our estimated coefficients suggest that closure of 882 HB homes during our study period resulted in 12,000–18,000 extra rehospitalizations within 30 days of discharge. The effect was largely concentrated among the most acutely ill, high-need patients.

Conclusions

SNF patient-based prospective payment resulted in closure of higher cost HB facilities that had served most postacute patients. As other, less experienced SNFs replaced HB facilities, they were less able to manage high acuity patients without rehospitalizing them.