Corrigendum to: Effectiveness of psychological interventions for chronic pain on health care use and work absence
Following publication of our systematic review and meta-analysis of effects on healthcare use and work loss of psychological interventions for chronic pain,2 Dr Rob Smeets contacted us to ask why his cost-effectiveness study of 2009,3 a follow-up of a randomized controlled outcome study of 2008,4 had been excluded. On investigation we discovered that its exclusion was our error, for which we apologize to Dr Smeets and his fellow authors, and to readers of PAIN. The 2009 paper was overlooked since an earlier report of the same study (2008) did not report changes in healthcare use or work, so was excluded from our initial shortlist of studies, leading in turn to the exclusion of the 2009 follow-up in which cost-effectiveness data on healthcare use and on work were provided.
Since the purpose of a systematic review is to include all eligible studies, the faulty selection should be rectified by addition of Smeets et al3 results to the systematic review and meta-analysis. The study met all inclusion criteria: it was a published report of a randomized controlled trial; its design had a psychological treatment delivered by qualified staff as the primary focus; at least one of the treatment(s) had defined psychological content; participants were adults (mean age 42, s.d. 10 years) with chronic low back pain of at least 3 months' duration (actual mean around 5 years); and there were more than 10 participants in each arm at the end of treatment. Methodological quality was high.
Changes to results are shown below, in bold, with a revised abstract with changes also in bold. The systematic review now includes 19, not 18 studies, with 16 (rather than 15) providing data on healthcare use, and 10 (rather than 9) on work loss. We used 2 of the 3 arms of the Smeets et al 20084 and 20093 trials: the graded activity with problem solving (GAP) as intervention (N = 52) and active physical treatment (APT) as the control (N = 52). This increases the overall N in the trials from 2253 to 2357, with no notable changes to personal and medical details of the total sample.