OBJECTIVES

To investigate the efficacy of interventions for the prevention and treatment of low back pain in nurses.

DESIGN

Systematic review.

DATA SOURCES

The review was registered on the PROSPERO database (CRD42015026941) and followed the PRISMA statement guidelines. A two phase approach was used. In phase one, all randomised controlled trials included in the systematic review of Dawson et al. (2007) which reviewed interventions for low back pain in nurses until 2004 were selected. In phase two, relevant randomised controlled trials and cluster randomised controlled trials published from 2004 until December 2015 were identified by an electronic search of nine databases (Embase, CINAHL, SPORTDiscus, PsycARTICLES, Cochrane Library, Web of Science, PEDro, Scopus and MEDLINE). To be eligible, trials had to examine the efficacy of interventions either for the prevention or treatment of low back pain in nurses. Primary outcomes of interest were any measure of pain and/or disability.

REVIEW METHODS

Three reviewers independently assessed eligibility and two reviewers independently conducted a risk of bias assessment (Cochrane Back and Neck Group).

RESULTS

Four studies were retrieved from phase one. In phase two, 15,628 titles and abstracts were scanned. From these, 150 full-text studies were retrieved and ten were eligible. Fourteen studies (four from phase one, ten from phase two) were eligible for risk of bias assessment. The included trials were highly heterogeneous, differing in pain and disability outcome measures, types of intervention, types of control group and follow-up durations. Only four of the included studies (n=644 subjects) had a low risk of bias (≥6/12). Manual handling training and stress management in isolation were not effective in nurses with and without low back pain (risk of bias, 7/12, n=210); the addition of a stretching exercise intervention was better than only performing usual activities (risk of bias, 6/12, n=127); combining manual handling training and back school was better than passive physiotherapy (risk of bias, 7/12, n=124); and a multidimensional intervention (risk of bias, 7/12, n=183) was not superior to a general exercise program in reducing low back pain in nurses.

CONCLUSIONS

Only four relevant low risk of bias randomised controlled trials were found. At present there is no strong evidence of efficacy for any intervention in preventing or treating low back pain in nurses. Additional high quality randomised controlled trials are required. It may be worth exploring the efficacy of more individualised multidimensional interventions for low back pain in the nursing population.