Responsiveness of Common Outcome Measures for Patients With Low Back Pain.

Taylor JS, Taylor AE, Foy AM, Fog ABJ (Orthopaedic Department, Princess Margaret Hospital; Physiotherapy Division, Biomedical Sciences, Kings College, London, United Kingdom), Spine. 1999;24:1805-1812.

The use of patient self-assessments as a measure of patient satisfaction and overall improvement during a regimen of physical therapy is becoming increasingly common. There are a number of disease-specific and generic health questionnaires whose validity and reliability have been established. The responsiveness of these instruments, however, has not been routinely reported. The authors defined responsiveness as the ability of a measurement tool to detect a clinically significant change over time as a result of treatment.

In this study the investigators examined the responsiveness of the 8 scales and 2 summary scales of the 36-Item Short Form Health Survey (SF-36) questionnaire with the responsiveness of the Oswestry Disability Index and Low Back Outcome Score questionnaires. To quantify responsiveness, the authors used the standardized effect size, a standardized measure of change that allows a direct comparison between different instruments that use different units. Using one common interpretation the authors considered an effect size of 0.2 to be small, 0.5 to be moderate, and 0.8 or greater to be large.

The authors examined 318 data sets of patients with the preliminary diagnosis of low back pain or sciatica. Patients underwent either surgery (120) or facet injection (136) or received physical therapy (62). Patients were given the questionnaires prior to and at varying times after intervention. The patients receiving injections or physical therapy received the questionnaires 2 months after treatment, whereas the patients receiving surgery received the questionnaires 6, 12, and 24 months after treatment. The follow-up questionnaire also contained a 7-point global effect rating scale ranging from "much better" (1) to "much worse" (7).

The Wilcoxon signed ranked test was used to determine whether a significant difference existed between pretreatment and posttreatment results. A Spearman rank correlation coefficient was used to examine whether there was a significant correlation between the change scores for each instrument and the patient global effect rating scale.

The results indicated that, in all scales, there were some inconsistencies between the patient's globally rated self-assessment and the change scores. However, the majority of scales of the SF-36 and disease-specific questionnaires used in the study were responsive to change in health status for patients with low back pain and sciatica. The greatest overall responsiveness was shown by the disease-specific questionnaires, but the SF-36 physical function subscale was more sensitive to change than either of the other questionnaires for patients whose status improved or remained unchanged.

Tiziano Marovino, PT Preferred Medicine Spine and Sports Allen Park, Mich