Predicting recovery using continuous low back pain outcome measures

Abstract

Background context: There is a lack of research evaluating multiple follow-up visits, specifically when using continuous outcome measures. Continuous outcome measures with several follow-up assessments would allow us to evaluate rate of recovery. Purpose: To predict low back pain outcomes based on the quantification of initial conditions. Study design/setting This was a prospective study where patients were enrolled within the first month of low back pain symptoms and evaluated for 3 months. Patients were recruited from several primary care facilities. Patient sample Thirty-two patients with local low back pain symptoms were recruited for the study. Outcome measures There were four major outcome measures, including functional performance probability, symptom intensity, impairment of activities of daily living, and a summary outcome measure. Methods Regression models were constructed using the initial conditions, including psychological, psychosocial, physical workplace, and personal factors, to predict the rate of recovery for each outcome measure. Results Twenty-eight patients completed the study. The r 2 value for the rate of recovery regression models were 0.77 symptom intensity prediction, 0.85 activities of daily living prediction, 0.87 functional performance probability prediction, and 0.96 summary outcome measure prediction. Two functional performance patterns of recovery were found, including a steady improvement and a large jump in improvement. A discriminant function model identified the pattern of recovery in 91% of cases given initial conditions. Conclusions Continuous outcome measures can be accurately predicted given the initial conditions.