Predicting Short-Term Response and Non-Response to Neck Strengthening Exercise for Chronic Neck Pain

Abstract

Summary of Background Data. People with chronic neck pain respond variably to exercise therapy. The likely success of exercise for chronic neck pain cannot be predicted.Study Design. This prospective study tested the relationship between patient characteristics assessed prior to intervention and response to an exercise program for chronic neck pain.Objectives. To investigate whether responders differed from non-responders with respect to presenting characteristics.Method. Data were collected on 336 consecutive patients receiving physiotherapy for chronic neck pain. Subject age, sex, duration of symptoms, compensation status, Neck Disability Index (NDI) item and total scores, cervical spine range of movement, and isometric strength were assessed prior to treatment. Response to therapy was defined as a change of 14/100 or more NDI points. Subjects received a progressive, concentric/eccentric, strengthening program using computerized equipment designed for cervical spine assessment and rehabilitation. Logistic regression formulae for predicting outcome at the completion of the program were developed on one sample (n = 122) data and tested on another (n = 214).Results. In the first sample, NDI scores and item responses predicted response or non-response with approximately 70% accuracy. When tested on the second sample, a prediction model using NDI item scores predicted responders (positive predictive value 56%) and non-responders (negative predictive value 74%) with considerable loss of responder prediction accuracy. Participants with low initial NDI scores (higher functional ability) appeared to benefit less from the exercise program.Conclusions. Functionally able patients may benefit less from exercise for chronic neck pain than those with greater functional limitations.