Hospitalization risk following admission to an academic nursing home

OBJECTIVES: Nursing home residents are at risk for experiencing acute illness and hospitalization early in nursing home stay. The objectives of this study were (1) to measure the probability of hospitalization over time from nursing home admission, and (2) to determine resident risk factors for early hospitalization. DESIGN: Prospective observational study. SETTING: 409-bed urban, not-for-profit academic nursing home. PARTICIPANTS: 204 consecutive newly admitted residents who did not previously reside in the nursing home. MEASUREMENTS: Demographic profile, chronic comorbidity, physical function, recent medical events, admission condition and anticipated length of stay were ascertained on admission to the nursing home. Hospitalization and acute illness were recorded each week after nursing home admission. Data were collected primarily by chart review except for acute illness, which was identified by nurse report and chart review according to previously established criteria. RESULTS: Subjects were 69% female and 62% white. The mean age was 83. The percentage of residents experiencing their first hospitalization peaked 2 and 7 weeks after nursing home admission and after 14 weeks plateaued at a low rate. The percentage with acute illness (commonly gastroenteritis, urinary tract infection, and pneumonia) paralleled that of hospitalization. In multivariate proportional hazards analysis, resident factors independently associated with early first hospitalization were atrial fibrillation (21% of subjects, 2.6 risk ratio), history of stroke (26%, 2.5), history of depression (25%, 1.9), absence of dementia (62%, 2.3), and receiving antibiotics on admission (23%, 3.3). Residents with 0, 1, 2,> or = 3 of these characteristics had 9%, 19%, 34% and 62% probability, respectively, of being hospitalized in the first 8 weeks of nursing home stay. CONCLUSIONS: In a large, urban, academic nursing home, residents' risk of hospitalization varied over time from admission and among different subgroups of residents. These results can be used to develop interventions to prevent acute illness and hospitalization early in nursing home stay and to enhance the communication of prognostic risk to nursing home residents, their caregivers and medical providers.