Effectiveness and Cost-benefit Evaluation of a Comprehensive Workers' Health Surveillance Program for Sustainable Employability of Meat Processing Workers.

The stepped wedge design was changed after two plants closed, and plant B2 was added non-randomly, which could introduce allocation bias. Participation was voluntary, and invitation strategy differed by plant; some plants invited only workers aged 50 years and older, making the sample older and less generalizable. Participants differed from non-participants on age, gender, job tenure, and sickness absence, so the people studied may not represent the workforce. About 49.3% of primary and secondary outcome items were missing, and response rates fell over time, which reduced power. The POSE program was not fully implemented as intended; follow-up of recommendations was not registered, and only one out of six participants followed an individual intervention within regular health care. Sensitivity analyses showed that adjusting for time changed the primary outcomes, so the negative effects may partly reflect time trends rather than the program itself. Baseline sickness absence was already low, about 4.5% average yearly absence, which may have left little room for improvement. The program was available only to contracted employees, excluding about 30% temporary workers and 25% foreign workers, and materials were not available in foreign languages. Self-reported productivity may not reflect actual productivity because workers had to keep up with conveyor belts, and colleagues may compensate for lower reported productivity.

Funding was provided by ZonMw. The supplied text does not report other conflicts of interest.

Do not conclude that any single element of the POSE program, such as physiotherapist counselling, is ineffective. The program combined questionnaire, biometric measurements, functional capacity evaluation, counselling, and recommendations; it was not delivered as intended, and time effects changed the primary results.