Significant Clinical Improvement Was Predicted in a Cohort of Patients With Low Back Pain Early in the Care Process

The purpose of this study was to determine the proportion of patients with low back pain who achieved clinical improvement in disability within 3 or 6 physical therapy visits, identify factors that predicted improvement, and predict the probability of improvement by the third and sixth visits.

This retrospective, observational study looked at patients (N = 6523) who completed a numeric pain scale and Modified Low Back Disability Questionnaire (MDQ) at every visit. Four prediction models were developed: 30% improvement by visit 3 and by visit 6 and 50% improvement by visit 3 and by visit 6. A logistic regression model was fit to predict patients' improvement in disability using the MDQ. Predictive models used age, disability scores, sex, symptom duration, and payer type as factors. Receiver operating characteristic curves and area under the curve were computed for the models. Nomograms illustrate the relative impacts of the predictor variables.

Disability improved 30% in 42.7% of patients by visit 3 and 49% by visit 6. Disability improved 50% in 26% of patients by visit 3 and 32.9% by visit 6. First visit score (MDQ1) was strongest factor to predict 30% improvement by visit 3. The visit 3 score (MDQ3) was strongest factor to predict a 30% or 50% improvement by visit 6. The combination of MDQ1 and MDQ3 scores was strongest overall predictive factor for visit 6. The area under the curve values for models using only the MDQ1 and MDQ3 scores to predict 30% or 50% improvement by the sixth visit were 0.84 and 0.85, respectively, representing excellent overall diagnostic accuracy of the prediction models.

Excellent discrimination to predict patients' significant clinical improvement by visit 6 using 2 outcome scores was demonstrated. Gathering outcomes routinely enhances assessment of prognosis and clinical decision making.

Understanding prognosis of clinical improvement supports physical therapists' contribution to value-based care.