Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: the need to present minimal important differences units in meta-analyses

BACKGROUND: The results of meta-analyses are all too often elusive, making it difficult to interpret their relevance for clinical practice. Reporting them in minimal important difference (MID) units could improve the interpretation of evidence in meta-analyses. The aim of this study was to compare, via calculation of MID units, outcomes after multidisciplinary biopsychosocial rehabilitation (MBR) versus usual care for pain relief in chronic low back pain (LBP). METHODS: We re-analyzed the data of a published Cochrane review on MBR. To attribute a MID to each pain instrument, we first searched the literature for MIDs. The MID was imputed for instruments without an established MID. We compared outcomes after MBR versus usual care for chronic LBP in the short (< 3 months), mid (> 3 and < 12 months), and long (≥12 months) term. The results of the meta-analyses are reported in MID units and interpreted as follows: if the overall effect size was greater than 1, many patients gained clinically important benefits, if it lay between 0.5 and 1.0, an appreciable number benefited, and if it fell below 0.5 few did. RESULTS: Improvement in back pain was observed in an appreciable number of patients in the short- and medium-term after MBR: the MID was lower but still close to 1 (0.75 and 0.86 MID units, respectively). MBR probably had little or no benefit for the majority of patients in the long-term, where the MID approached 0 (0.27 MID units, confidence interval 0.07–0.48). CONCLUSIONS: Meta-analyses expressed in MID units may offer better insight into the clinical relevance of MBR: the intervention is highly recommended for reducing pain in the short- and medium-term but cannot be recommended for long-term pain reduction since the benefit decays rapidly. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12955-018-0924-9) contains supplementary material, which is available to authorized users.