Predicting timely recovery and recurrence following multidisciplinary rehabilitation in patients with compensated low back pain.

STUDY
DESIGN: Historical cohort study.
OBJECTIVES: We investigated factors predictive of timely and sustained recovery following multidisciplinary rehabilitation in Workers' Compensation claimants with low back pain. SUMMARY OF BACKGROUND DATA: It is still unknown which factors predict better outcomes among back pain patients enrolled in intensive rehabilitation programs. Previously, few consistent predictors have been reported.
METHODS: We created and tested predictive models using data from clinical and administrative databases of the Alberta Workers' Compensation Board. Predictive models were built on a cohort of subjects admitted for multidisciplinary rehabilitation in 1999 and tested on subjects admitted in 2000. Cox regression was used to evaluate days to time-loss benefit suspension and days to claim closure following admission for rehabilitation. Logistic regression was used to evaluate risk of future recurrence as judged through time-loss benefit resumption, claim reopening, or new back-related claims filing.
RESULTS: Prediction models were variable between exploratory and confirmatory stages, and few variables were found to predict consistently. The number of preadmission healthcare visits was the most robust predictor of all recovery outcomes. Recurrence rates were 18% in 1999 and 22% in 2000. A higher number of preadmission healthcare visits and more previous back-related claims were associated with higher risk of recurrence.
CONCLUSIONS: The number of preadmission healthcare visits was the most robust prognostic indicator with more healthcare visits related to delayed recovery and higher risk of recurrence. Recurrence rates following successful functional restoration were consistent with the episodic and recurrent nature of low back pain.