RESUMEN
OBJECTIVES: To quantify geographic variation in home health expenditures per Medicare home health beneficiary and investigate factors associated with this variation. STUDY DESIGN: Retrospective study design analyzing US counties in which at least 1 home health agency served 11 or more beneficiaries in 2016. Several sources of 2016 national public data were used. METHODS: The key variable is county-level Medicare home health expenditures per home health beneficiary. Counties were grouped into quintiles based on per-beneficiary expenditures. Analyses included calculation of coefficients of variation, computation of the ratio of 90th percentile to 10th percentile in expenditures, and linear regression predicting expenditure. The control variables included characteristics of patients, agencies, and communities. RESULTS: Significant variation in home health expenditures was identified across county quintiles, with a 90th-to-10th-percentile expenditure ratio of 2.5. The percentage of for-profit agencies in the lowest quintile was 15.7 compared with 81.7 in the highest quintile of spending. Unadjusted spending differed by $3864 (95% CI, $3793-$3936), compared with $3611 (95% CI, $3514-$3708) in the adjusted model, between counties in spending quintiles 1 and 5. Although state fixed effects explained nearly 20% of the variation in home health expenditures, 42% of the variation remained unexplained. CONCLUSIONS: Home health care exhibits considerable unwarranted variation in per-patient expenditures across counties, signifying inefficiency and waste. Given the expected growth in home health demand, strategies to reduce unwarranted geographic variation are needed.