The feasibility of developing a pain scale for detecting psychological disturbance in patients with low back pain was investigated in two studies. In the first study, stepwise discriminant analysis was used to identify the best combination of pain terms that would discriminate the functional group from the organic group. Thirteen pain variables were identified which correctly identified 93.6% of the cases of low back pain. These pain terms were used to categorize 159 new cases in the second. Cross-validation shrinkage was only 10.6%. One hundred and thirty-two out of 159 cases in the new sample were correctly classified. The advantages of this pain scale over existing methods are discussed, along with implications for clinical orthopedic practice. © 1979.
