Posthospital Care for Viulnerable Elderly: Are Discharge Patterns Equitable?
Publication Date: January 1993
Publisher(s): Pardee Rand Graduate School
Author(s): Andrea Steiner
Since implementation of the U.S. Medicare Prospective Payment System (PPS), patients' need for post-acute care and the importance of delivering it appropriately have increased. Yet there is disturbing evidence of variation in posthospital treatment across regions of the country and among sub-groups of elderly. This dissertation examines the factors associated with use of Medicare Part A posthospital services (skilled nursing facility, home health care, and inpatient rehabilitation) and compares both use and outcomes (readmission or death within 90 days of discharge) for patients who are aged 85-plus, female, non-white, or low-income with other Medicare patients. The data source is a 1987/88 linked episode file that combines a 20-percent national random sample of Medicare hospital and posthospital claims with provider-market, and state-level data (n=1.9 million). The primary analytic technique was logistic regression involving 60 predictor variables. Central findings are as follows: (1) Post-PPS, large variations in use of posthospital services as a function of age, gender, race, and income persist, and are associated with significant variations in outcomes; (2) The elderly poor are significantly less likely than other beneficiaries to receive inpatient rehabilitation; (3) For home health care and rehabilitation, the greatest differences between users' and non-users' outcomes occurred for the oldest old (e.g., 90-day mortality rates for stroke patients were 25 percent for non-users versus 12 percent for users of rehabilitation care); (4) Although non-whites are disproportionately high users of home health care--and disproportionately low users of skilled nursing facility (SNF) care--white users appeared to benefit most from home health care treatment (e.g., for orthopedic diagnoses, 90-day mortality rates for non-whites were the same for users and non-users; for whites, users' death rates were 10 percentage points lower than non-users'); (5) Similarly, the mortality difference between users and non-users of home health care was greater for women than men. Findings point to inequities in selection, with the possibility that the presence of informal support may pave the way for clinically inappropriate discharge decisions.