Hospital Readmissions and Hospital Location

Regular readers know we are no fans of CMS’ hospital readmission reduction program in particular, and have concerns about the fairness of all such efforts, given current shortfalls in understanding causes and methods that actually work for reductions.   (HSR Article)   New research published in Health Services Research does nothing to ease our qualms.  The authors examined relationships between county-level measures of community characteristics and hospital readmission rates.  They used CMS’ readmission data and hospital and county specific data from a number of sources.  Three primary sets of factors were examined.  One was the level of certain patient characteristics, like living alone, employment status and educational level.  Another  was measures of access to care, particularly primary care.  The final one was the number and quality of nursing homes in the county.  4,073 hospitals in 2,254 counties were included in the analysis.  Even before accounting for hospital or county-specific factors, the researchers found that 58% of the variation in readmission rates was at the county level.  The factors that explained the greatest amount of variation among counties, included the number of primary care physicians per capita, rural versus urban status and the average percent of high-risk, long stay nursing home residents with pressure sores.  Specific county characteristics  that were associated with higher readmission rates were proportion of the population never married, number of Medicare beneficiaries per capita and low education status, along with higher numbers of specialists per capita and hospital beds per capita. Lower readmission rates appeared related to rural status for the country or it being a retirement area, higher numbers of primary care physicians per capita and more nursing homes per capita.  Altogether the county characteristics explained about 47.5% of the variation in readmissions rates across counties.  So the lesson is that individual hospital factors and performance probably only account for half of variation in performance, but the CMS program does not take this into account.  No wonder many hospitals are on the verge of just giving up on improving their readmission rates.