The aim of this study was to compare two approaches for a worker's health surveillance (WHS) mental module on work functioning and work-related mental health. Nurses and allied health professionals from one organisation were cluster-randomised at ward level to e-mental health care (EMH) (N = 579) or occupational physician care (OP) (N = 591). Both groups received screening and personalised feedback on impaired work functioning and mental health. Positively screened participants received an invitation to follow a self-help EMH intervention, or for a consultation with an OP. The primary outcome was impaired work functioning. Follow-up was performed after 3 and 6 months. Linear mixed models were applied to determine differences. Non-inferiority of the EMH-care approach was demonstrated if the mean absolute improvement on work functioning in the OP-care group was ≤10 points higher than the EMH-care group. Analyses were performed on the positively screened participants (almost 80 %) (EMH N = 75; OP N = 108) and all participants (EMH N = 98; OP N = 142). Both groups improved over time regarding impaired work functioning. A considerable percentage of participants had improved relevantly at follow-up regarding work functioning (3 months: EMH 30 %, OP 46 %; 6 months: EMH 36 %, OP 41 %) compared to baseline. No statistically significant differences were found between the groups, and the difference did not exceed the pre-defined criterion for non-inferiority. The OP-care approach for a WHS mental module trended towards better performance in targeting work functioning, but our findings indicate that the EMH-care approach was non-inferior. However, the high dropout rate and low compliance to EMH interventions should be taken into account.