Eight studies (n=750) were included: 4 randomised controlled trials (RCTs; n=234), 3 cohort studies (n=463) and one case series (n=53).
Overall, the quality of studies was rated as fair (5 studies scored at least 32 out of 48 on validity); the scores ranged from 18 to 43. Methodological limitations included small sample sizes, the lack of standardised outcome measures and statistical analysis, inadequate reporting of the intervention and results, and the lack of control groups.
Exercises (1 RCT, 33 workers): this RCT found that individualised physical therapy and group exercises both reduced pain, with greater reductions found with individualised therapy.
In-house or worksite physical therapy (1 cohort study, 152 workers): this study found that in-house therapy reduced time lost from work compared with outside therapy. The baseline characteristics of the treatment groups were not compared and there was no description of the outside therapy.
Worksite analysis (1 cohort study, 43 workers): this study found that worksite analysis reduced absence from work in comparison with no worksite analysis, but there was a low (46%) response rate from one of the worksites.
Nurse case managers' training on accommodations (1 RCT, 101 claimants; 1 case series, 53 claimants): both studies suggested that training changed the practice of accommodations, but the data were incomplete and were not analysed statistically.
Ergonomic modifications (2 RCTs, 100 computer users; 1 cohort study, 268 users): both RCTs found that using alternative geometry keyboards may reduce hand pain; the cohort study found that regular breaks may improve recovery, but the outcome measure was not standardised and there was potential for intervention bias.