Reporting of rehabilitation intervention for low back pain in randomized controlled trials: Is the treatment fully replicable?

Study Design. Methodological review of randomized controlled trials (RCTs). Objective. To assess the quality of reporting of rehabilitation interventions for mechanical low back pain (LBP) in published RCTs. Summary of Background Data. Reporting of interventions in RCTs often focused on the outcome value and failed to describe interventions adequately. Methods. We systematically searched for all RCTs in Cochrane systematic reviews on LBP published in the Cochrane Database of Systematic Reviews until December 2013. The description of rehabilitation interventions of each RCT was evaluated independently by 2 of the investigators, using an ad hoc checklist of 7 items. The primary outcome was the number of items reported in sufficient details to be replicable in a new RCT or in everyday practice. Results. We found 11 systematic reviews, including 220 eligible RCTs, on LBP. Of those, 185 RCTs were included. The median publication year was 1998 (I-III quartiles, 1990 to 2004). The most reported items were the characteristics of participants (91.3%; 95% confidence interval [CI], 87.3-95.4), the intervention providers (81.1%; 95% CI, 75.4-86.7), and the intervention schedule (69.7%; 95% CI, 63-76). Based on the description of the intervention, less than one fifth would be replicable clinically. The proportion of trials providing all essential information about the participants and interventions increased from 14% (n=7) in 1971 to 1980 to 20% (n=75) in 2001 to 2010. Conclusion. Despite the remarkable amount of energy spent producing RCTs in LBP rehabilitation, the majority of RCTs failed to report sufficient information that would allow the intervention to be replicated in clinical practice. Improving the quality of intervention description is urgently needed to better transfer research into rehabilitation practices.