Predictors of outcome following hip fracture rehabilitation.

Objective: To determine the potential predictors of functional outcome after hip fracture rehabilitation in a large acute inpatient rehabilitation facility.

Design: Retrospective chart review study.

Setting: Large acute inpatient rehabilitation hospital.

Participants: Patients admitted with a primary admitting diagnosis of hip fracture (n = 753). Patients were excluded if their hip fracture was the result of high-velocity trauma and if their stay was shorter than 48 hours (196 patients).

Methods: Independent variables included body mass index, gender, ethnicity, smoking history, alcohol consumption, past living situation, past ambulatory status, medical history, prealbumin level, medications that increase the risk of falling, and evidence of prior osteoporosis workup and treatment. These data were entered into a password-encrypted database. Univariate analyses were carried out to evaluate the relationship between independent variables and main outcomes, and multivariate analyses were performed to assess the impact of medical history of diabetes adjusting for other covariates.

Main Outcome Measurements: Discharge location; length of rehabilitation hospital stay (LOS); Functional Independence Measure (FIM) gain, which is calculated as the FIM(discharge) - FIM(admission); and length of stay efficiency (LOSE), which is calculated as the FIM gain divided by the LOS, and measures the rate of FIM change.

Results: Patients with diabetes had a worse LOSE (P = .0008). Multiple linear regression analysis revealed that patients who have a medical history of diabetes have a 0.33 reduction of LOSE compared with other patients. Predictors of better LOSE included younger age (P < .001), fewer medications that predispose to falls (P < .0001), and independent ambulation before fracture (P = .0003).

Conclusion: We have found several significant patient characteristics that portend a better functional outcome after hip fracture. These include younger age, female gender, absence of diabetes mellitus, independent prefracture ambulation, not living alone before fracture, and being prescribed fewer medications that predispose to falling during rehabilitation. On the contrary, one of our most interesting findings is that patients with diabetes made slower gains, had a longer LOS and were less likely to be discharged directly home from the acute rehabilitation facility. There are numerous factors that may contribute to this, and suggestions are made for future research.