Group Data or Categorical Data for Outcomes of Pain Treatment?

Early in 2018 the Spine Section of Pain Medicine adopted a policy to require authors to provide categorical data in studies of pain treatment. This requirement did not preclude reporting group data for descriptive or analytical purposes, but it stipulated that categorical data must at least be added. Group data describe samples in terms of the mean value for pain scores (or other outcome variables) along with its standard deviation. These parameters are then used to test for statically significant differences, be they before and after treatment or between different treatments. A P value of 0.05 typically serves as the threshold for significant differences. In contrast, categorical data describe the numbers and proportions of patients who achieve defined categories of outcome. In particular, those proportions indicate the success rate of treatment in achieving those outcomes. In comparative studies, respective success rates can be compared by comparing their 95% confidence intervals. If an index treatment has a greater success rate and if its confidence intervals do not overlap with those of the comparison treatment, it can be concluded with statistical confidence that the index treatment is more often effective than the comparison treatment. In observational outcome studies, the 95% confidence intervals indicate to readers the possible range of values in which they expect their success rate to fall should they adopt the treatment.