Alternative and more clinically applicable approach to the standardized mean difference effect size for meta-analysis of differentially scaled pain outcome data in anesthesia and pain medicine.

Background/purpose: The production of systematic reviews with meta-analysis in the field of analgesia and pain medicine has increased dramatically over the years and is increasingly used to guide clinical practice as well as decisions by others (policymakers, etc). A common metric for pooling data from studies that use different scales to assess the outcome of interest, for example, pain, is to convert the results from each study using the standardized mean difference (SMD) effect size. However, this is problematic because the SMD is not easy to interpret by the non-statistician. In this brief technical report, we describe how to easily rescale data into a common and more easily interpretable metric, including the provision of an easy-to-use Excel worksheet for rescaling one's own data.

Methods: Data from a previous meta-analysis of randomized controlled trials that examined the effects of transcutaneous electrical nerve stimulation on pain, assessed using different pain scales, were used. Using an Excel spreadsheet and selected formulas, data for each study were rescaled to a metric commonly used to assess pain in the clinical setting, 0-10. Results were then pooled using the inverse-variance heterogeneity model.

Results: Rescaling pain data to 0-10 were easily accomplished using this 'real-world' dataset.

Conclusion: Rescaling data into a more understandable metric intended for a wider variety of audiences is plausible. It is the hope that future systematic reviews that include a meta-analysis will use this approach when the results for an outcome of interest such as pain are reported using different scales.