Objectives To investigate the feasibility of recruitment, adherence and likely effectiveness of an e-learning intervention for managers to improve employees’ well-being and reduce sickness absence.

Methods The GEM Study (guided e-learning for managers) was a mixed methods pilot cluster randomised trial. Employees were recruited from four mental health services prior to randomising three services to the intervention and one to no-intervention control. Intervention managers received a facilitated e-learning programme on work-related stress. Main outcomes were Warwick Edinburgh Mental Wellbeing Scale (WEMWBS), 12-item GHQ and sickness absence <21 days from human resources. 35 in-depth interviews were undertaken with key informants, managers and employees, and additional observational data collected.

Results 424 of 649 (65%) employees approached consented, of whom 350 provided WEMWBS at baseline and 284 at follow-up; 41 managers out of 49 were recruited from the three intervention clusters and 21 adhered to the intervention. WEMWBS scores fell from 50.4–49.0 in the control (n=59) and 51.0–49.9 in the intervention (n=225), giving an intervention effect of 0.5 (95% CI −3.2 to 4.2). 120/225 intervention employees had a manager who was adherent to the intervention. HR data on sickness absence (n=393) showed no evidence of effect. There were no effects on GHQ score or work characteristics. Online quiz knowledge scores increased across the study in adherent managers. Qualitative data provided a rich picture of the context within which the intervention took place and managers’ and employees’ experiences of it.

Conclusions A small benefit from the intervention on well-being was explained by the mixed methods approach, implicating a low intervention uptake by managers and suggesting that education alone may be insufficient. A full trial of the guided e-learning intervention and economic evaluation is feasible. Future research should include more active encouragement of manager motivation, reflection and behaviour change.