Abstract
Background
We have developed a LBP (low back pain) repository with IPD (individual patient data) from therapist-delivered intervention RCT's from various institutions, nationally and internationally (19 trials, 9328 patients). We will perform IPD meta-analyses to identify and determine the most effective and cost-effective treatments for subgroups of patients. Disability due to LBP is a key outcome of interest. However, there was no common disability measurement that was used in all the RCT's. We report here our assessment as to whether different disability measures are sufficiently comparable to allow pooling across trials.
Method
We used data from seven trials (n = 5015 patients) that had at least two of the following measurements: RMDQ (Roland Morris Disability Questionnaire), MVK (Modified Von Korff disability score), physical component scale of SF-12/36, Patient Specific Functional Scale, and Hannover Functional Ability Questionnaire. Within each trial, the change score between baseline and short-term follow-up was computed. These change scores were then converted from one outcome measure to another with linear regression and ridge-reroughed linear models. They were also tested for correlation and responsiveness (the ability to detect changes in condition, measured by standardised response means).
Conclusion
Most of the correlations between outcome measurements were high (>0.5), but responsiveness was often significantly different. Moreover, the accuracy of the conversions was poor, e.g. >40% of predictions from the conversion of RMDQ to MVK were incorrect by more than 10 points. The lack of agreement between different outcome measures suggests that conversion between different disability measurements is not recommended.
