Falcone D, Broyles R.
Access to long-term care: race as a barrier.
J Health Polit Policy Law 1994;19(3):583-95.

This study assessed whether hospital discharges of nonwhite patients in North Carolina were delayed to a significant extent because of difficulty finding long-term care placement (generally in a nursing home). Subjects included patients awaiting discharge in May of 1991 from 76 acute care hospitals (two-thirds of all such hospitals in the state).

In multivariate analyses that assessed factors associated with experiencing delay in discharge, having a financial problem in arranging a discharge had the largest effect (OR=31.53, p<0.001). Other important variables included nonwhite race (OR=1.31, p=0.003), having a family problem in arranging a discharge, and being discharged to a rest home versus to a nursing home (beds are only scarce in nursing homes). In multivariate analysis that assessed factors associated with longer delays, the same factors were important, as were the patient’s need for special conditions (patient need of a ventilator, dialysis, oxygen; the patient is intubated, tracheotomied, obese, comatose; or the patient has decubitus ulcer, cognitive impairment, or other conditions), the patients' dependence on private funding 3 months after discharge, and age.

The authors note two previously offered explanations for these findings. Since this study only assessed those patients awaiting nursing home placement, the findings could not be attributed to patient preferences against nursing home care because the decision for nursing home placement had already been made. Others have speculated that the longer delays of nonwhite patients have been due to the fact that nursing homes had to "match" patients on race; that is, the homes had to defer to the racial preferences of the patients in the roommate selection.

The authors conclude, however, that "the only explanation for the longer delays of nonwhites is the preference of nursing home owners or operators for white patients."

Interpreting the results of this study is limited by lack of detailed information about the sample selection process, the data collection process, and a description of the variables included in the multivariate analyses. Other factors, such as severity of illness, type of illness on admission, and length of hospital stay prior to release orders, might be considered in future analyses as possible confounders of the race effect. However it is unlikely that adding such variables would eliminate the large and consistent race effect reported in this study.