Systematic Review of Workplace Interventions to Support Young Workers' Safety, Work Environment and Health.

Only 33 studies were found, and only 6 of them met more than 85% of the quality criteria; the other 27 were medium quality. Most outcome areas rest on one to five studies. The best evidence synthesis counts a small study the same as a large one, so a trial of a few dozen people carries as much weight as a trial of thousands. The studies differed so much in who took part, what was delivered and what was measured that the authors could not pool the results, so there is no overall estimate of effect and no recommendation for any single intervention. Most included studies were of health-related professions such as nurses, junior doctors, surgeons and anaesthetists, so the findings mainly describe that group rather than young workers generally. Only three studies included workers under 18, so the findings mostly describe 18- to 29-year-olds, and young workers under 18 often work under different legal rules. Thirteen of the 33 studies were not randomised, which the authors acknowledge can introduce bias in blinding and in how groups were formed. The authors state the review may be affected by publication bias, that searching as widely as they did may have made studies with positive results more likely to be included, and that restricting included studies to English or Scandinavian languages may leave out available evidence. The stakeholders who shaped the review came from four Danish industries, so they do not represent all sectors of work.

The text supplied names one funder, the Danish Working Environment Research Fund, and reports no author conflicts of interest.

The three positive results in this review do not mean those programmes work. Interventions aimed at attitudes and beliefs, behaviour-based interventions and multifaceted skin-protection programmes each rested on a small number of medium-quality studies, and the review's own message for every one of them is that there is not enough evidence to guide practice. Reading them as ready-to-use approaches goes further than the evidence allows.