Community demographics linked to hospital readmissions

Nearly 60 percent of the variation in hospital readmission rates appears to be associated with where the hospital is located rather than on the hospital's performance, finds a new study in Health Services Research.

"The biggest surprise was how much affect the county or the community had on readmission rates," said Jeph Herrin, Ph.D., senior statistician with Health Research and Educational Trust in Chicago and lead author on the study.

"Hospitals in the same area had similar readmission rates as others in the area," he said. This means that the hospital may not have as much control over its readmission rates as previously thought.

The federal Medicare program has linked hospital reimbursements to risk-standardized rates of readmission within 30 days of discharge for patients with certain conditions, such as acute myocardial infarction, heart failure, or pneumonia. High readmission rates trigger reductions in Medicare reimbursement.

Herrin and his colleagues matched data on hospital readmissions collected by the Centers for Medicare and Medicaid Services (CMS) with information about the hospitals and the counties where they are located. Data was collected for 4,073 hospitals in 2,254 counties and covered the period from July 2007 to June 2010. Information about hospitals included ownership type, teaching status, number of beds, and percentage of patients covered by Medicaid. Demographic information about counties included population, the number of Medicare beneficiaries, the number of general practitioners and specialists, and data on poverty and education levels.

Analysis showed that 58 percent of the variation in 30-day readmission rates was at the county level, before any information about the type of hospital or county was taken into account. In other words, individual hospital performance may account for only 42 percent of the variation in readmission rates. When demographic information was analyzed, having more general practitioners and more nursing homes per capita in a county was associated with lower readmission rates. When all the information was analyzed together, characteristics of the location of the hospital account for 47.5 percent of the variation in readmission rates between hospitals.

One conclusion of the study is that hospital readmission rates may need to be addressed with community based programs in addition to improvements within the hospital, said Herrin.

The findings reaffirm what is already known, said Georges C. Benjamin, M.D., executive director of the American Public Health Association in Washington. "We know clearly that place matters in relation to our health and that there is a strong correlation to where you live."

Penalizing hospitals for high readmission rates is a policy intended to improve hospital performance and bring health care costs down. "The penalty incentive is a good idea," said Benjamin. "It forces accountability."