Evaluating occupational health interventions : design, implementation, and effects

Sammanfattning: Background: Poor mental health, e.g. stress, anxiety, and depression, in the workplace is a challenge worldwide due to the individual suffering and its impact on sickness absence and productivity loss, causing societal costs. The World Health Organisation classifies stress as the health epidemic of the 21st century. Psychosocial working conditions, i.e., how work is organised and the social interplay at work, are health determinants. Thus, psychosocial hazards are one explanation for work-related mental ill-health. Interventions aiming to improve the psychosocial work environment are recommended. Still, there is a scarcity of studies evaluating occupational health interventions targeting psychosocial working conditions to prevent mental ill health. Also, the existing evidence of the effectiveness of such interventions is inconclusive. Implementation failure is described as one main obstacle to succeeding with these interventions. To tackle the global challenges of work-related stress, we need a better understanding of what can be done in the workplace to prevent employees from becoming ill due to workplace stressors. Aim: This thesis aims to contribute to knowledge on how stress-related ill health can be prevented in the workplace and develop our understanding of the design and implementation of occupational health interventions. Methods: This thesis comprises three papers that evaluate two occupational health interventions to improve the psychosocial work environment and mental health. The interventions were conducted within the human services (I) and construction industry (II & III), respectively. In study I, we applied an embedded mixed methods design to evaluate a participatory intervention to improve the psychosocial work environment and mental health (burnout and quality of sleep) within a municipality in Sweden. We utilised a controlled trial and a process evaluation exploring fidelity and participants’ reactions to the intervention activities, learning experiences, and changes in behaviours and work routines. We collected data through documentation, interviews and three waves of questionnaires. Differences in outcome variables (questionnaires) over time were calculated using t-tests for partially overlapping samples to handle partly different study populations at each time point caused by employee turnover and drop-out. We analysed the interview data by applying a thematic analysis. The second and third studies were conducted in a large Swedish construction company. In study II, we investigated the participants’ satisfaction with engaging in the co-creation process, perceived knowledge, and skill development through interviews. In total, eight men and four women participated. We applied a thematic analysis to analyse the data. In study III, we used a controlled trial to evaluate the potential effects of the co-created intervention on the psychosocial work environment and self-reported stress. We collected data on the outcomes with online questionnaires at baseline, 12, and 24 months. We also assessed adherence to the intervention and dose delivered (i.e., fidelity). Marginal means models adjusting for missing data patterns were applied to estimate potential differences in outcomes between groups over time. Findings: Neither of the interventions improved the long-term outcomes of burnout and stress, respectively. We found different effects of the municipality and construction industry interventions on the psychosocial working conditions. Within the municipality, we found detrimental effects of the intervention on social support from the manager, empowering leadership, control of work pacing, and role clarity. Within the construction industry, there was a noticeable improvement in role clarity for white-collar workers in the intervention group compared to the control group. The implementation fidelity, i.e., whether the intervention activities were delivered according to plan, was moderate in both projects. However, the process evaluation within the municipality project showed that the intervention activities led to few changes in attitudes, behaviours, and work routines. On the contrary, adherence to the construction industry intervention increased during the trial. The results of paper II showed that the co-creation participants reported increased learning about the psychosocial work environment and mental health. The respondents perceived the intervention and the implementation strategies as relevant and feasible. Thus, involving different stakeholders and allowing the organisation to decide the intervention activities and the implementation strategies seem to have enabled a good contextual fit. Conclusions: The program theory, i.e., intervention developed within the construction company, can potentially improve role clarity for white-collar workers. Three design principles stand out regarding their positive effect on the implementation: organisational capability and incentive systems to promote health, aligning the intervention with existing organisational objectives and practices and conducting a needs assessment. The co-creation process in the construction industry seems to have positively affected the above-mentioned design principles. Thus, co-creating occupational health interventions seems promising to improve the implementation.