Objective To evaluate the effectiveness of a comprehensive workers’ health surveillance (WHS) program on aspects of sustainable employability and cost-benefit. Methods A cluster randomized stepped wedge trial was performed in a Dutch meat processing company from february 2012 until march 2015. In total 305 workers participated in the trial. Outcomes were retrieved during a WHS program, by multiple questionnaires, and from company registries. Primary outcomes were sickness absence, work ability, and productivity. Secondary outcomes were health, vitality, and psychosocial workload. Data were analyzed with linear and logistic multilevel models. Cost-benefit analyses from the employer’s perspective were performed as well. Results Primary outcomes sickness absence (OR = 1.40), work ability (B = −0.63) and productivity (OR = 0.71) were better in the control condition. Secondary outcomes did not or minimally differ between conditions. Of the 12 secondary outcomes, the only outcome that scored better in the experimental condition was meaning of work (B = 0.18). Controlling for confounders did not or minimally change the results. However, our stepped wedge design did not enable adjustment for confounding in the last two periods of the trial. The WHS program resulted in higher costs for the employer on the short and middle term. Conclusions Primary outcomes did not improve after program implementation and secondary outcomes remained equal after implementation. The program was not cost-beneficial after 1–3 year follow-up. Main limitation that may have contributed to absence of positive effects may be program failure, because interventions were not deployed as intended.