70
Patient Factors Associated with Variation in Same-Hospital Readmission
Initiatives introduced in the Affordable Care Act, such as the Centers for Medicare and Medicaid Services Readmissions Reduction program and the Centers for Medicare and Medicaid Initiatives Bundled Payments for Care Initiative, hold hospitals at risk for readmissions that occur at other hospitals. This study describes the extent of hospital readmissions at the original place of treatment (i.e., same-hospital readmissions) and patient characteristics that predict same-hospital readmissions.
Data
This study used data from16 U.S. states in the Healthcare Cost and Utilization project State Inpatient Databases for calendar years 2010 and 2011. These states had patient identifiers that could be used to track a patient across hospital stays within the state during the study period.
Methods
Patients were included if they were aged 18 or older and hospitalized with a CMS Hospital Readmissions Reduction Program target condition (acute myocardial infarction, heart failure, total hip or knee arthroplasty, and pneumonia) or a DRG that is part of the CMMI Bundled Payments for Care Improvement Initiative (major joint replacement of the lower extremity, revision of the hip or knee, spinal fusion, and urinary tract infection).
All patients with a readmission within 30 days of a qualifying index visit for the selected diagnoses and DRGs were identified. The hospital identifier of the index hospital was compared to the readmission hospital to determine whether the readmission occurred at the same hospital.
The analysis included three steps. First, the distribution of same-hospital readmissions across hospitals was examined. Second, the same-hospital readmission rate was calculated for selected patient characteristics: age, sex, expected primary payer, severity of illness, whether the readmission was through the emergency department, and whether the patient resided in a different county from the initial index hospital.
Finally, the statistical association between patient characteristics and likelihood of same-hospital readmission was calculated. A general linear model was estimated using general estimating equations at the patient and hospital level to address patient clustering within hospitals.
Results
Among the conditions studied, hospitals lose approximately one quarter of their readmissions to other hospitals. The distribution of same-hospital readmission rates for orthopedic admissions were notably different than for other conditions, with a trend of more hospitals capturing 100% of their readmissions. Several patient and hospitalization characteristics were associated with a same-hospital readmission: age, primary payer, severity of illness, and location of index hospital relative to patient residence. Across all diagnoses and DRGs, patients whose index hospital was in a different county and patients who were readmitted through the ED were more likely to switch hospitals. Among orthopedic conditions, younger in age, private insurance, and lower severity predicted same-hospital readmission. For AMI, heart failure, UTI, and pneumonia, older in age, female, and greater severity predicted same-hospital readmission.
Conclusions
Hospitals may want to consider methods to increase their same-hospital readmission rate to enhance continuity and so they have more control over resources given their potential financial risk. Policy makers should be aware that patient mix contributes to variation in same-hospital readmissions when crafting policies and programs.