Benefits of Home Health Care After Inpatient Rehabilitation for Hip Fracture: Health Service Use by Medicare Beneficiaries, 1987-1992.

Benefits of Home Health Care After Inpatient Rehabilitation for Hip Fracture: Health Service Use by Medicare Beneficiaries, 1987-1992

Intrator O, Berg K (Center for Gerontology and Health Care Research, Brown University, Providence, RI), Arch Phys Med Rehabil. 1998;79:1195-1199.

This study sought to determine if the addition of home health care to inpatient rehabilitation helped reduce subsequent hospitalizations and nursing home admissions. The researchers conducted a 1% random sampling of the Medicare claims information of elderly beneficiaries who sustained hip fractures in 1986. The inclusion criteria were a primary diagnosis of hip fracture, a minimum age of 70 years, admission to inpatient rehabilitation after hospital discharge, and discharge to the community with or without the provision of additional home health care services within 21 days of the discharge. Claims were excluded if the subjects had received other inpatient or home health care services during the previous year. The sample consisted of 324 elderly people who were generally healthy and dwelled in the community. The outcomes were defined as any rehospitalization or any admission to a non-skilled nursing facility (non-SNF) occurring within 12 months of discharge from the hospital.

Because only patients who were home-bound--who may have had lower levels of function and mobility--qualified for home health care services, the authors tried to minimize the effects of this inherent selection bias by calculating a propensity score for each patient. This score used patient characteristics to calculate the probability of receiving home health care. The authors then stratified the cohort into quartiles based on these propensity scores and, within each quartile, made comparisons between those subjects who received home care and those who did not. To test for an overall effect of home health care on subsequent hospitalization or admission to a non-SNF, the authors used Cox proportional hazards analysis, stratified by propensity quartiles, and obtained adjusted odds ratios and confidence intervals.

One hundred ninety-five patients received home health care, whereas 129 were discharged without additional services provided. The group that received home care had more comorbidities and tended to be less healthy overall. This group also had a longer length of rehabilitation (20 days vs 17 days). The average length of home health care services was 40 days.

The raw data indicated an overall trend toward higher rehospitalization rates in the home care group. The researchers expected this trend based on this group's lower general health status. The results of the Cox proportional hazards analysis, however, demonstrated that the addition of home care to rehabilitation actually reduced the risks for hospitalization and nursing home admission in the subsequent year. The authors offered several possible reasons for this benefit, including specificity of training. Specificity of training was suggested because home care involves training in the same environment in which the patient functions. They also hypothesized that the longer length of postacute treatment, rather than the nature of the home care, may contribute to these benefits.

In conclusion, the authors suggested that increased investment in postacute care may help avoid the use of additional services within 1 year after a hip fracture, resulting in lower overall costs, not just the costs for individual services or for a particular episode of care.

Andi Beth Mincer, PT
Armstrong Atlantic State University
Savannah, Ga