Effects of interventions implemented by occupational health professionals to prevent work-related stress complaints: a systematic review

Work-related mental health problems are a significant and growing public health concern, accounting for 18.6% of all reported work-related health problems in the EU in 2020. These complaints affect workers’ well-being, performance, work ability, absenteeism and healthcare costs. Many countries engage occupational health professionals and services to promote and protect workers’ health, prevent sickness absence and support organisational performance. Occupational health professionals provide preventive and curative services, including health surveillance, assessments and consultations; they identify psychosocial risk factors, offer support and resources, and implement preventive interventions. Prior systematic reviews have often assessed interventions (eg, exercise, mindfulness, e-health, primary care) without explicitly addressing occupational risk factors or workplace context, potentially overlooking key causes of work-related stress and burnout. This review aims to summarise the effectiveness of preventive interventions delivered by occupational health professionals that target work-related stress complaints.

Previous systematic reviews have explored activities and interventions to prevent work-related mental health complaints but largely focused on interventions that do not explicitly account for occupational risk factors (eg, job demands, work attitudes, social support). Examples include physical exercise, mindfulness, and care delivered outside occupational health (primary care) or online e-health programs. While these can improve mental health and reduce stress symptoms, they often lack integration with workplace context and may miss causes of work-related stress and burnout. Evidence suggests organisational-level interventions can be more effective than individual-focused ones, and combined individual-organisational approaches may yield longer-lasting results. However, controlled organisational interventions targeting stress and burnout are relatively scarce in the literature, and the roles of occupational health professionals in implementing preventive interventions have been underrepresented.

Design and reporting followed PRISMA guidelines. A comprehensive search strategy was developed with a librarian, tested, and applied to PubMed, Embase, PsycInfo and Medline on May 2, 2023, and repeated on August 13, 2023. Search terms covered worker/employee/workplace; prevention/intervention; occupational health/medicine; occupational physician; occupational health service; occupational health nursing; stress/burnout/well-being/distress/job stress/mental health; trial design terms; and language (Dutch, English). Reference lists of included studies and related reviews were checked (snowballing). Additional handsearching was conducted in Google Scholar and the Cochrane Library. Eligibility (PICO): - Population: working populations not on sick leave. - Intervention: delivered by an occupational health professional (broadly defined to include professionals labelled with ’occupational health’ or ’occupational medicine’); universal, selective, and indicated prevention; individual or organisational level. - Control: a control group (no/minimal intervention or care as usual). - Outcomes: self-reported work-related stress complaints (distress, work-related fatigue, burnout). Selection: Records were imported into Rayyan; duplicates removed; titles/abstracts screened, followed by full-text screening. One researcher screened, second checked; disagreements resolved by discussion; authors were contacted if delivery by occupational health professionals was unclear. PRISMA flow diagram summarised identification and selection. Data extraction: One researcher extracted into a predefined Excel form (study characteristics, design, PICO, professional involved, participants, data collection/unit, main results, conclusions). Short-term outcomes were defined as measured up to 1 year after baseline; long-term outcomes >1 year. Risk of bias: Two researchers independently assessed risk of bias using RoB-2 for randomised trials (five domains) including intention-to-treat considerations, and ROBINS-I for non-randomised studies (seven domains). Disagreements were discussed; domain definitions provided. Analysis: Due to heterogeneity in methods and effect measures, meta-analysis was not feasible. A qualitative narrative synthesis described key findings, grouped studies by prevention type (universal, selective, indicated), and identified consistencies and differences across studies.

- Included studies: 9 total (5 RCTs; 4 quasi-experimental controlled). Interventions: cognitive behavioural/stress management training (4); consultations/screening by occupational health professionals (3); e-health/online interventions (3); peer support (1). Implementing professionals included occupational physicians, occupational nurses, physiotherapists, psychologists, OHS social workers, and occupational medicine trainers. - Effect summary: Four studies showed no significant effects; five showed mixed, short-term positive effects or benefits in subgroups with high adherence. Interventions of greater intensity or multiple sessions/components tended to be more effective than one-time or low-intensity interventions. Study-specific outcomes (selected): - Ojala et al (Finland): Cognitive behavioural programme led by occupational physiotherapist, psychologist and nurse. At 9 months, lower burnout (p=0.023) and exhaustion (p<0.001) in intervention vs control. - Umanodan et al (Japan): Six-session worksite stress-management training with OP and occupational nurses. Intention-to-treat: no effect on psychological distress at 6 months (p>0.05). Per-protocol (complete attendance): favourable effect (p=0.049). - Kawaharada et al (Japan): Three SMT sessions led by OP and occupational nurse. No significant effects on stress response subscales at immediate post-test or 1-month (e.g., vigour p=0.569, anger p=0.088, fatigue p=0.279, anxiety p=0.924, depression p=0.263, somatic stress response p=0.631). - Limm et al (Germany): Stress management tailored to managerial needs addressing organisational stressors. At 12 months, lower perceived stress reactivity (SRS) in intervention (p=0.016). No significant cortisol effect; α-amylase lower in intervention at 12 months (p=0.075). - Gärtner et al (Netherlands): Online screening plus voluntary preventive OP consultations for nurses/allied health. No significant effects on distress, work-related fatigue or PTSD at 3/6 months (p>0.05). Improvement in work functioning at 6 months (p=0.04). - Steel et al (Belgium): Electronic health survey with selective OP follow-up vs periodic face-to-face screening. At up to 19 months, no significant effects on mental health (GHQ; p>0.01). - de Boer et al (Netherlands): Occupational health programme with supervisor/manager consultations and individual action plan for at-risk workers. At 6 months, less emotional exhaustion (p<0.05) and emotional distance (p<0.01); at 2 years, no differences except less emotional distance (p<0.05). Baseline differences favoured control on some well-being measures. - Peterson et al (Sweden): Peer support groups led by OHS physicians, social workers, physiotherapists for workers with high exhaustion. At 12 months, greater reduction in exhaustion vs control (p=0.040). - Notenbomer et al (Netherlands): E-health plus OP care for employees with frequent sickness absence. No effect on burnout at 1 year (p=0.300). Risk of bias: All RCTs had high risk of bias in one or more RoB-2 domains (overall high), largely due to impossibility of blinding and reliance on self-reported outcomes. All non-randomised studies had serious risk of bias (ROBINS-I), commonly due to confounding and participant selection; one study had >30% missing data.

Overall, findings were mixed: five studies showed positive or short-term benefits, while four showed no significant effects. Several factors may explain these results: - Intervention intensity and duration: One-time or low-intensity interventions did not yield significant effects, whereas multi-session or multi-component interventions more often produced favourable outcomes, suggesting higher intensity/frequency is beneficial. - Outcome measure heterogeneity: Diverse stress-related outcomes and instruments hinder comparability. A core outcome set for work-related stress and burnout is recommended to standardise assessments. - Focus on individual-level interventions: Most included studies targeted individuals rather than organisational factors. Evidence suggests organisational-level interventions, or combined individual-organisational approaches, can be more effective and longer-lasting, and occupational health professionals are well-positioned to bridge individual needs and organisational change. - Implementation quality and adherence: Poor implementation or limited adherence may reduce effectiveness; reporting on fidelity and compliance was often lacking. Involving occupational health professionals directly in implementation may support sustainability. - Methodological constraints: High risk of bias, especially due to lack of blinding and complex workplace contexts, is common in occupational health trials. While controlled trials remain the gold standard, alternative designs may be necessary to address practical and ethical constraints. These insights underscore the need for longer-term assessments, standardisation of stress outcomes, more organisational interventions, and better reporting on implementation and adherence.

Five of nine studies reported positive short-term outcomes, while four found no significant effects. All studies exhibited high risk of bias. Intervention intensity appears to influence effectiveness, but the overall evidence base for preventive interventions delivered by occupational health professionals to address work-related stress remains limited. Further high-quality research is needed, with standardised outcomes, attention to organisational-level strategies, and improved implementation reporting, to deepen understanding of intervention effectiveness and mechanisms.

- High risk of bias across all included studies (blinding infeasible; self-reported outcomes; confounding and selection issues in non-randomised designs; missing data). - Heterogeneity of interventions, populations, designs, outcomes and baseline values precluded meta-analysis and limits generalisability. - Predominantly individual-level interventions; scarcity of controlled organisational-level interventions. - Limited reporting on implementation fidelity and participant adherence, impeding interpretation of null effects versus poor compliance. - Potential misclassification/exclusion of studies due to unclear reporting of who delivered the intervention. - Possible missed studies despite comprehensive database search and handsearching. - Short duration or low intensity common among interventions; limited long-term follow-up and sustainability evidence.