Mortality and locomotion 6 months after hospitalization for hip fracture: risk factors and risk-adjusted hospital outcomes

CONTEXT: Hip fracture is a common clinical problem that leads to considerable mortality and disability. A need exists for a practical means to monitor and improve outcomes, including function, for patients with hip fracture. OBJECTIVES: To identify and compare the importance of significant prefracture predictors of functional status and mortality at 6 months for patients hospitalized with hip fracture and to compare risk-adjusted outcomes for hospitals providing initial care. DESIGN: Prospective study with data obtained from medical records and through structured interviews with patients and proxies. SETTING AND PARTICIPANTS: A total of 571 adults aged 50 years or older with hip fracture who were admitted to 4 New York, NY, metropolitan hospitals between August 1997 and August 1998. MAIN OUTCOME MEASURES: In-hospital and 6-month mortality; locomotion at 6 months; and adverse outcomes at 6 months, defined as death or needing assistance to ambulate, compared by hospital, adjusting for patient risk factors. RESULTS: The in-hospital mortality rate was 1.6%. At 6 months, the mortality rate was 13.5%, and another 12.8% needed total assistance to ambulate. Laboratory values were strong predictors of mortality but were not significantly associated with locomotion. Age and prefracture residence at a nursing home were significant predictors of locomotion (P =.02 for both) but were not significantly associated with mortality. Adjustment for baseline characteristics either substantially augmented or diminished interhospital differences in outcomes. Two hospitals had 1 outcome (functional status or mortality) that was significantly worse than the overall mean while the other outcome was nonsignificantly better than average. CONCLUSIONS: Mortality and functional status ideally should be considered both together and individually to distinguish effects limited to one or the other outcome. Hospital performance for these 2 measures may differ substantially after adjustment, probably because different processes of care are important to each outcome.