Extending the Aberdeen Back Pain Scale to include the whole spine: a set of outcome measures for the neck, upper and lower back.

Many therapists treat the spine as a 'functional unit', but suitable condition-specific outcome measures for the whole spine are not available. One of the most rigorously tested measures for back pain related health status is the Aberdeen Back Pain Scale, but it is only suitable for the lower back. The Aberdeen Back Pain Scale was extended to create a set of interlocking outcome measures for the neck, upper and lower back. Questions in these extended outcome measures had to fulfil a series of psychometric criteria before being accepted into the final questionnaires. This involved response frequency, item-total correlation and principal component analysis. The final questionnaires were tested for reliability, criterion and construct validity, responsiveness and acceptability, on patients attending a primary care musculoskeletal clinic. One question was discarded from all three sets of questionnaires and a further question was removed from the neck questionnaire. Baseline scores approximated to normal distributions. Although not completely reproducible, they were internally consistent, so showed evidence of reliability. They were highly correlated with the SF-12 and their mean scores changed according to whether their health status, measured by a transition question, improved, stayed the same or worsened. Modified standardised response means showed large changes when health status improved and moderate-sized changes when health status worsened. Patients made few adverse comments about the questionnaires and found them acceptable. The Extended Aberdeen Spine Pain Scales for neck, upper and lower back pain, showed evidence of reliability, validity, responsiveness and acceptability. They can be used for single regions of the spine or combined as clinically necessary. They are particularly recommended for primary care patients.