Effect of Reassuring Information About Musculoskeletal and Mental Health Complaints at the Workplace: A Cluster Randomized Trial of the atWork Intervention.

Both groups received an active atWork programme, and there was no untreated control group, so the trial cannot show whether either programme reduced sick leave compared with usual workplace practice. The analysis was per protocol only, not intention to treat, because intention-to-treat data were not available. There was no blinding to group assignment, because everyone at the workplace received the same programme content. Several statistical tests were performed without adjustment for multiple comparisons. The two interventions were similar, and the authors state this may have made the trial insufficiently powered to detect differences on general sick leave. Some kindergartens withdrew after randomization, and sick-leave data were missing for one of the 93 participating kindergartens. Follow-up respondents were more likely to be women and to have higher age and education than non-responders. The sample was limited to private kindergarten employees, was mostly female, and came from one occupational group, so results may not transfer to other workplaces or occupations. The primary outcome used only physician-certified sick leave, not self-certified absence. Baseline differences were found between groups for some mental-health beliefs and directive social support.

The supplied text reports funding from Vestfold Hospital Trust and no additional conflict-of-interest declaration.

Do not read the 5.7% reduction in sick leave in the modified group as evidence that the programme works. That reduction was not statistically significant, the original group had a 7.5% increase, and the primary comparison between groups was null. Because both groups received an active atWork programme and there was no untreated control, the trial cannot show that either programme reduced sick leave compared with usual practice.