Fifteen studies (n=6,593 participants) were included in the review, comprising 10 randomised controlled trials (RCTs) and five clinical controlled trials (CCTs). Four (n= 357 participants) of the 15 included studies were rated as high quality, but the remaining studies were judged to be of low quality with methodological limitations, including problems with randomisation, blinding, compliance reporting and follow-up.
Incidence of low back pain: Statistically significant beneficial effects were observed in four low quality RCTs and three CCTs on the incidence of low back pain with exercise interventions. Compliance with treatment was not reported in the RCTs reporting this outcome.
Intensity of low back pain: Of the six studies that evaluated this outcome, two high quality RCTs and one low quality RCT reported significant improvements in low back pain intensity with exercise interventions. One high quality trial reported improvements in office workers after 15 weeks of light resistance training of a five minute duration during work time (p=0.02). Compliance was 69%. In a second high quality RCT, railroad workers significantly improved on visual analogue scores of pain intensity after a twice-weekly programme of specific strength, balance stretching, and lumbar neutral zone exercises (p=0.052). However, compliance was poor during the last six months of the study, with 27% remaining in guided training at 12 months.
Impact of low back pain and disability: The outcomes evaluated were sick leave, activity interference and cost of low back pain. Of four trials that evaluated sick leave, two low-quality RCTs reported statistically significant benefits of exercise for nursing staff and manufacturing factory workers. One RCT and one CCT each observed a significant decrease in activity interference with exercise interventions. Another study reported significant (p=0.028) improvements in self-estimated work ability.