• Health economics;
  • instrumental variables;
  • health care organizations and systems


To determine whether the rate of rehospitalization is lower among patients discharged to skilled nursing facilities (SNFs) with which a hospital has a strong linkage.

Data Sources/Collection

We used national Medicare enrollment, claims, and the Minimum Data Set to examine 2.8 million newly discharged patients to 15,063 SNFs from 2,477 general hospitals between 2004 and 2006.

Study Design

We examined the relationship between the proportion of discharges from a hospital and alternative SNFs on the rehospitalization of patients treated by that hospital–SNF pair using an instrumental variable approach. We used distances to alternative SNFs from residence of the patients of the originating hospital as the instrument.

Principal Findings

Our estimates suggest that if the proportion of a hospital's discharges to an SNF was to increase by 10 percentage points, the likelihood of patients treated by that hospital–SNF pair to be rehospitalized within 30 days would decline by 1.2 percentage points, largely driven by fewer rehospitalizations within a week of hospital discharge.


Stronger hospital–SNF linkages, independent of hospital ownership, were found to reduce rehospitalization rates. As hospitals are held accountable for patients' outcomes postdischarge under the Affordable Care Act, hospitals may steer their patients preferentially to fewer SNFs.