The efficacy of workplace mental health promotion interventions in enhancing well-being and service outcomes among nurses: a systematic review

¹Department of Mental Health, Faculty of Health Sciences, University of Respati, Jakarta, Indonesia, ²Department of Mental Health, Faculty of Nursing, Universitas Indonesia, Jakarta, Indonesia, ³Department of Community Nursing, Faculty of Nursing, Universitas Indonesia, Jakarta, Indonesia, ⁴Dr. Cipto Mangunkusumo National Central Public Hospital, Jakarta, Indonesia, ⁵Department of Mental Health Nursing, National Health Polytechnic, Banten, Indonesia

^&Corresponding author
Thika Marliana, Department of Mental Health, Faculty of Health Sciences, University of Respati, Jakarta, Indonesia

Introduction    

Nurses play a central role in healthcare delivery, yet their work is frequently associated with high levels of stress, heavy workloads, and demanding clinical responsibilities. These challenges intensified during the COVID-19 pandemic and continue to affect nursing practice even after the acute phase of the pandemic has subsided [1]. In many countries, particularly those facing aging populations and workforce shortages, the demand for nursing services remains high [2]. As a result, nurses often experience long working hours, staffing shortages, and increasing clinical complexity.

Sustained exposure to these occupational stressors can negatively affect both physical and psychological health. Previous studies have reported high rates of professional burnout, psychological distress, anxiety, depression, insomnia, and reduced work ability among nurses [3]. Estimates suggest that between one-quarter and one-half of nurses experience symptoms of burnout or psychological strain. In addition, psychosocial work stressors such as high job demands, low decision latitude, and inadequate support are associated with reduced work capacity and increased risk of musculoskeletal and mental health disorders [4]. These consequences not only affect nurses' well-being but may also compromise quality of care, patient safety, and workforce sustainability.

Workplace mental health promotion interventions have been proposed as a strategy to address these challenges. Such interventions aim to reduce psychosocial and physical risk factors, improve working conditions, and enhance overall well-being and performance [5]. Previous systematic reviews have examined specific intervention types, particularly mindfulness-based approaches. Some reviews report small positive effects on burnout, well-being, and job performance, whereas others show inconsistent or minimal effects [6]. Moreover, evidence regarding other types of workplace interventions such as organizational changes, physical activity programs, psychosocial training, or multifaceted approaches remains fragmented [7]. The impact of these interventions on outcomes such as sickness absence, occupational injuries, and psychosocial risk factors is still unclear.

Individual-level interventions (mindfulness, resilience training, stress-management education, cognitive-behavioural approaches) target proximal psychological outcomes such as coping, emotional regulation, and perceived stress, whereas organizational-level interventions (workload modification, staffing changes, leadership support, and broader changes to the psychosocial work environment) target structural determinants of occupational stress. Because these intervention categories operate through different mechanisms and are likely to produce effects of different magnitudes and durations, the present review distinguishes between them in the synthesis.

Participation in workplace health programs is often reported to be low to moderate, and barriers such as workload, time constraints, and limited managerial support may reduce engagement [8]. However, evidence specifically addressing barriers and facilitators among nurses is limited.

Taken together, existing literature suggests that while workplace mental health promotion interventions show promise, the overall effectiveness across different outcomes and settings remains uncertain [9]. A particular concern is that much of the existing evidence is presented narratively without study-level summary tables that describe intervention characteristics, sample sizes, outcome measures, effect sizes, or follow-up durations, limiting readers' ability to assess the underlying evidence. A comprehensive synthesis of recent intervention studies is therefore needed to clarify what works, in which settings, and under what conditions. This review specifically aimed to synthesize study-level data on intervention characteristics and quantitative effects across predefined primary and secondary outcomes distinguish individual- from organizational-level interventions examine implementation factors and identify methodological limitations in the current evidence base. This systematic review aimed to synthesize current evidence on workplace mental health promotion interventions for nurses [10]. Specifically, the objectives were to: 1) Identify workplace interventions that improve nurses' work ability, well-being, perceived general health, job performance, and job satisfaction. Identify workplace interventions that reduce psychosocial stressors, professional burnout, sickness absence, and occupational injuries; 2) examine reported barriers and facilitators influencing nurses' participation in workplace mental health promotion programs. This article reports on the findings of a systematic review evaluating the effectiveness and implementation considerations of workplace mental health promotion interventions for nurses [11].

Methods     

Study design: this study was conducted as a systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. The review aimed to identify, evaluate, and synthesize evidence from randomized controlled trials assessing the effectiveness of workplace mental health promotion interventions among nurses. Although prospective registration in PROSPERO or another international registry was not completed before screening commenced, this is explicitly acknowledged as a limitation of the review. To strengthen transparency in the absence of prospective registration, the predefined research questions, eligibility criteria, search strategy (including database-specific strings in Annex 1), study selection procedures, primary and secondary outcomes, data extraction variables, and the planned approach to risk-of-bias assessment and narrative synthesis were established a priori and documented in an internal protocol that remained unchanged throughout the review process. Two reviewers independently performed study selection, data extraction, and risk-of-bias assessment to minimize selection and reporting bias. The authors recommend that future updates or follow-up reviews in this area be prospectively registered to enhance methodological transparency and reduce the potential for selective reporting.

Setting: eligible studies were conducted in workplace healthcare settings, including acute care hospitals, inpatient long-term care (LTC) facilities, home-based LTC settings, and cross-setting healthcare environments. Studies from any country were eligible. Only English-language publications were included, and the potential consequences of this restriction are discussed in detail in the Limitations subsection.

Participants: the review included studies involving registered nurses or nursing personnel working in clinical healthcare settings. Studies were eligible if they were randomized controlled trials evaluating workplace mental health promotion interventions and reported at least one relevant outcome related to work ability, well-being, perceived general health, job performance, job satisfaction, professional burnout, psychosocial risk factors, sickness absence, or occupational injuries. Studies published between January 2019 and 30 July 2024 were included.

Studies were excluded if they were non-randomized controlled trials, randomized trials without a control group, qualitative-only studies, studies involving informal caregivers, studies conducted among students or non-clinical nursing staff, or interventions consisting of a single brief session of less than 20 minutes with only immediate outcome measurement.

Variables: to improve transparency and support structured synthesis, review outcomes were predefined and categorized as primary and secondary outcomes before data extraction and are summarized in Table 1.

Primary outcomes (prioritized): well-being (psychological, occupational, and global); professional burnout (emotional exhaustion, depersonalization, personal accomplishment); work ability; job satisfaction; psychosocial risk factors (job demands, decision latitude, social support).

Secondary outcomes: perceived general health; job performance/work engagement; sickness absence; occupational injuries; implementation-related outcomes (participation rates, adherence, managerial support, feasibility, and intervention fidelity) (Table 1). This classification guided the narrative synthesis; primary outcomes were synthesized first and in greater detail, whereas secondary outcomes were addressed subsequently, allowing a hierarchical interpretation of intervention effectiveness.

Data sources and measurement: a comprehensive literature search was conducted in ProQuest, EBSCO, PubMed, and PubPsych from database inception to 30^th July, 2024. The search strategy combined Medical Subject Headings (MeSH) and text-based keywords. The core search structure was: ("mental health" OR "mental well-being") AND ("workplace" OR "occupational") AND ("nurs*") AND ("promotion" OR "intervention*"). Search syntax was adapted for each database as required. The complete electronic search strategies, including the database-specific search strings, Boolean operators, truncation symbols, field tags, and search limits used for each database are provided in Annex 1 to ensure full reproducibility, in accordance with PRISMA 2020 recommendations.

Reference lists of relevant review articles were manually screened to identify additional eligible studies. Gray literature (including theses, dissertations, conference proceedings, preprints, government reports, and trial registries) was not searched. This restriction was applied to ensure inclusion of peer-reviewed evidence but has methodological implications for the review's evidence base. Specifically, the exclusion of gray literature increases the risk of publication bias because studies reporting statistically significant or positive findings are more likely to be published in peer-reviewed journals than studies reporting null or negative results. In addition, studies published as conference proceedings may report preliminary findings that would otherwise enrich the evidence base. The exclusion of non-English publications further compounds this risk by potentially omitting relevant evidence from non-English-speaking contexts. The combined effect of these restrictions on the completeness of the evidence is discussed in the limitations subsection.

Study selection: all identified records were exported into reference management software, and duplicates were removed prior to screening. A total of 4,611 records were identified through database searches, and 4,418 records were removed prior to screening due to duplication, automation filtering, or ineligibility. Two reviewers independently screened 193 titles and abstracts against predefined inclusion and exclusion criteria.

Eighty-one reports were sought for retrieval. Thirty-four full-text reports were successfully retrieved and independently assessed for eligibility, whereas 47 reports could not be retrieved. This represents a substantial proportion (47 of 81, ≈58%) of the full-text records sought. The impact of this shortfall on the representativeness of the evidence base is addressed in the limitations subsection. Disagreements were resolved through discussion and, when necessary, consultation with a third reviewer. After exclusions, 14 studies were included in the final review. The selection process is presented in the PRISMA 2020 flow diagram (Figure 1).

Studies identified through manual screening of reference lists were subjected to the same eligibility assessment and selection procedures as records identified through database searches.

Data extraction: two reviewers independently extracted data using a standardized data extraction form developed a priori. Extracted information included study identifiers (author, year, country, setting); participant characteristics (sample size, professional role); detailed intervention descriptors (content, theoretical basis, format, duration, intensity, delivery mode); comparator description; outcome measures (validated instruments, time points); quantitative results (effect estimates, mean differences, standardized mean differences, relative risks, 95% confidence intervals, p-values where reported); follow-up duration; information required for risk-of-bias assessment; and any reported implementation outcomes (participation rates, adherence, fidelity, feasibility, managerial support). Disagreements were resolved through discussion, with consultation from a third reviewer when necessary. A list of studies excluded at the full-text stage together with the primary reason for exclusion is provided in Annex 1.

Bias: risk of bias of included randomized controlled trials was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool. The following domains were evaluated: bias arising from the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and selection of reported results. Each study was categorized as low risk of bias, some concerns, or high risk of bias. Risk-of-bias assessment was performed independently by two reviewers, with disagreements resolved by consensus. Risk-of-bias judgments were considered during interpretation of the review findings and were used to contextualize study-level results; specifically, the synthesis explicitly notes where high risk or some concerns in particular domains (missing outcome data, deviations from intended interventions) may have inflated or otherwise influenced reported intervention effects. Study-by-study RoB 2 judgments across the five domains, together with the overall risk-of-bias classification, are presented in Table 2 and visually summarized in Figure 2 (traffic-light plot). Publication bias was not formally assessed (via funnel plots or Egger's test) due to heterogeneity in interventions and outcomes and the limited number of included studies.

Certainty of evidence: the certainty of the overall body of evidence across outcome domains could not be formally assessed using the grading of recommendations assessment, development and evaluation approach. Substantial heterogeneity existed in intervention characteristics, outcome measures, and follow-up periods, and the limited number of studies within individual outcome strata rendered a meaningful GRADE assessment infeasible at this stage. Consequently, methodological quality was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool, and the absence of a formal certainty-of-evidence assessment should be considered when interpreting the review findings. The authors recommend that future systematic reviews in this area incorporate a GRADE assessment once a sufficient number of studies with comparable interventions and outcomes are available.

Study size: the included randomized controlled trials varied in sample size, ranging from small pilot studies to larger trials. Sample size calculations and power analyses were reported inconsistently across studies. No power calculation was conducted at the review level because this study synthesized evidence from previously published trials.

Quantitative variables: outcome variables were extracted and summarized as reported in the original studies. Continuous outcomes (burnout scores, well-being measures) and categorical outcomes (sickness absence, occupational injuries) were recorded. Where available, effect estimates (mean differences, relative risks) and measures of precision (confidence intervals) were extracted to support interpretation. No transformations or re-categorization of variables were performed at the review level.

Statistical methods: the potential for quantitative meta-analysis was considered during protocol development. A meta-analysis was ultimately judged to be inappropriate for the full set of included studies because of substantial clinical and methodological heterogeneity, described below. Clinical heterogeneity (extent across included studies): diverse intervention approaches (mindfulness-based programs, resilience training, psychosocial interventions, organizational strategies, physical activity interventions, and multicomponent workplace programs); diverse healthcare settings (acute hospitals, inpatient long-term care facilities, mixed clinical environments) with varying staffing models, organizational cultures, and regulatory environments; diverse participant characteristics and professional roles. Methodological heterogeneity (extent across included studies): variable intervention intensity, duration, and delivery mode (individual, group-based, digital, or blended); variable follow-up periods ranging from weeks to several months; diverse outcome measurement instruments (different validated instruments for well-being, burnout, and job satisfaction), limiting the comparability of effect estimates.

The review team explicitly explored whether quantitative synthesis would have been feasible for subsets of studies. However, the number of studies within any single intervention-outcome stratum was small, outcome definitions and instruments were inconsistent across trials, and follow-up periods varied substantially. Pooling under these conditions was considered likely to compromise the validity and interpretability of pooled estimates and to potentially produce misleading summary effects. For these reasons, a narrative synthesis was undertaken. Sensitivity analyses (by restricting to low risk-of-bias studies or to specific intervention categories) were not performed because no quantitative synthesis was conducted; this is acknowledged as a limitation. The authors recommend that future reviews consider subgroup meta-analyses (by intervention category or setting) once the evidence base grows.

Sources of heterogeneity (narrative framework): to assist interpretation, heterogeneity across the included studies was conceptualized within a narrative framework comprising four principal sources: intervention characteristics diverse content (mindfulness, resilience, psychoeducation, physical activity, digital or blended programs, multicomponent workplace initiatives) targeting different psychological and organizational mechanisms; intervention duration and implementation intensity ranging from brief programmes over several weeks to more intensive interventions with extended follow-up, plausibly influencing both effect magnitude and durability; healthcare setting variations in organizational culture, staffing levels, workload, leadership support, and resource availability across acute hospitals, inpatient long-term care facilities, and mixed environments; outcome assessment diverse validated instruments and time points across studies of well-being, burnout, job satisfaction, work ability, psychosocial risk factors, sickness absence, and occupational injuries limiting direct comparison. Findings were grouped according to the predefined outcome domains, and the direction, consistency, and strength of evidence across studies were summarized descriptively.

Results     

Participants: a total of 4,611 records were identified through database searches (ProQuest, EBSCO, PubMed, and Pub Psych). Before screening, 4,418 records were removed. The remaining 193 records were screened by title and abstract, and 112 were excluded, including 92 trials that were not randomized or did not include an appropriate control group. Full texts were sought for 81 reports, of which 47 (approximately 58% of sought reports) could not be retrieved. A total of 34 full-text reports were assessed for eligibility, and 20 were excluded for predefined reasons (including ineligible populations such as caregivers and students, qualitative- only studies, and single brief interventions with immediate outcome assessment only). Ultimately, 14 randomized controlled trials were included in the review. The selection process is presented in Figure 1.

Reasons for exclusion of full-text articles included ineligible study design, non-clinical populations, qualitative-only methodology, absence of an eligible comparison group, interventions not meeting predefined eligibility criteria, and insufficient outcome reporting. A detailed list of excluded studies and reasons for exclusion is provided in Annex 1.

Risk of bias assessment: the methodological quality of the 14 included randomized controlled trials was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool across five domains: (1) bias arising from the randomization process; (2) bias due to deviations from intended interventions; (3) bias due to missing outcome data; (4) bias in measurement of the outcome; and (5) bias in selection of the reported result.

Domain-level summary of findings: 1) randomization process: most studies were judged as low risk of bias; a smaller subset raised some concerns due to incomplete reporting of allocation concealment or baseline imbalances; 2) deviations from intended interventions: several studies were rated as having some concerns due to lack of blinding of participants or incomplete reporting of co-interventions; 3) missing outcome data: a subset of studies was rated as high risk of bias in this domain, primarily because of substantial attrition or differential loss to follow-up between arms; 4) measurement of the outcome: generally low risk of bias across the included studies, as most outcomes were based on validated self-report instruments with comparable administration in both arms; 5) selection of the reported result: generally low risk of bias, although concerns were noted for a small number of trials where pre-specified analysis plans were unavailable.

Overall distribution of risk-of-bias judgments across the 14 included trials: most studies were judged as having low risk of bias or some concerns, while a smaller number of studies were rated as high risk of bias, primarily because of missing outcome data, incomplete adherence to intended interventions, or insufficient reporting of allocation procedures. The interpretation of intervention effects in subsequent subsections explicitly considers these risk-of-bias judgments, particularly for outcomes and trials where high risk or persistent some concerns were noted. Detailed study-level risk-of-bias assessments are presented in Figure 2 (traffic-light plot) and Table 2.

Descriptive data: the 14 included trials evaluated workplace mental health promotion interventions among nurses across multiple healthcare settings. Studies were most commonly conducted in acute hospital settings (n = 7) and inpatient long-term care facilities (n = 4). Within the total of 14 trials, the predominant intervention categories were individual-level interventions (mindfulness- or stress-management-based, resilience training, and psychological skills training), with a smaller subset focused on organizational-level strategies (including workload reorganization and staffing adjustments), and a small number evaluating multicomponent or blended approaches. The remaining studies were conducted in mixed or cross-setting clinical environments. No eligible trials were identified in home-based long-term care settings.

Interventions varied in content and delivery (individual-focused, group-based, workplace-delivered, and digital or blended formats). Follow-up periods and outcome measurement tools differed across studies, which limited comparability. Participant and workplace characteristics were reported variably across trials; reporting of missing data also varied between studies. Although individual participant numbers, intervention dose, and follow-up duration varied across the included trials, a detailed study-level description is provided in Table 2 to enable appraisal of intervention characteristics and reported outcomes.

Table 2 summarizes the characteristics of the 14 included randomized controlled trials, including study setting, country, sample size, intervention characteristics (content, duration, delivery mode), comparator, outcome measures, follow-up duration, reported effect estimates (mean differences, standardized mean differences, 95% confidence intervals, and p-values), and principal findings. Considerable variation was observed across intervention approaches, outcome assessment tools, sample sizes, and follow-up periods, which contributed to the heterogeneity of the evidence and is reflected in the narrative synthesis (Table 3).

Outcome data: outcomes were synthesized according to the predefined primary outcome domains established during data extraction. Where reported in the included randomized controlled trials, quantitative findings including effect estimates, confidence intervals, standardized mean differences, adjusted between-group differences, and measures of statistical significance were extracted to facilitate comparison across studies. Owing to substantial clinical and methodological heterogeneity in intervention characteristics, outcome measures, and follow-up durations, a meta-analysis was not performed. Instead, findings were synthesized narratively by outcome domain. A detailed study-level summary of intervention characteristics, sample sizes, outcome measures, follow-up periods, and principal findings is presented in Table 3, while the corresponding risk-of-bias assessments are presented in Table 3. Quantitative findings, including effect estimates (mean differences, standardized mean differences (SMDs), adjusted between-group differences, and relative risks), 95% confidence intervals, and p-values were extracted to facilitate comparison across studies. Because pooled meta-analysis was not conducted, these quantitative findings are summarized narratively by outcome domain in the subsections below; numerical details for each study are presented in Table 3.

Well-being and stress-related outcomes: well-being and psychological stress were the most frequently evaluated outcomes across the included trials. Several randomized controlled trials reported statistically significant short-term improvements in psychological well-being or reductions in perceived stress following workplace mental health promotion interventions, whereas several other trials found no statistically significant differences between intervention and control groups. Where effect estimates were reported, the magnitude of effect on short-term well-being varied, ranging from small to moderate, with confidence intervals spanning both clinically modest and more substantial improvements, and several trials reporting non-significant differences. Quantitative findings, including effect estimates, confidence intervals, and p-values, are summarized in Table 2 by study.

Professional burnout: burnout was commonly assessed using validated instruments, most frequently the Maslach Burnout Inventory. Some interventions demonstrated improvements, particularly in emotional exhaustion, whereas findings for depersonalization and personal accomplishment were less consistent. Where reported, intervention effects on burnout dimensions ranged from small to moderate; however, several studies reported no statistically significant between-group differences. This pattern was observed across both individual-level and multicomponent interventions.

Job satisfaction, work engagement, and job performance: job satisfaction, work engagement, and job performance were evaluated in a smaller subset of trials. Some interventions produced statistically significant improvements in these outcomes, while others demonstrated no meaningful differences compared with control conditions. Where effect estimates were available, reported changes in job satisfaction tended to be small to moderate and short-term; confidence intervals frequently included the null value. Variability in intervention design, implementation strategies, and follow-up duration contributed to inconsistent findings across studies.

Work ability, perceived general health, and psychosocial risk factors: relatively few studies evaluated work ability, perceived general health, or psychosocial workplace risk factors. The available evidence showed inconsistent effects across intervention types and healthcare settings, with most reported effect estimates either non-significant or accompanied by wide confidence intervals, limiting confidence in sustained improvements over time.

Sickness absence and occupational injuries: sickness absence and occupational injuries were infrequently reported across the included trials. The small number of studies assessing these outcomes, together with variation in outcome definitions and follow-up periods, and the typical absence of adequately powered analyses for these relatively rare events, prevented robust conclusions regarding the effectiveness of workplace mental health promotion interventions for these occupational outcomes.

Overall synthesis: overall, the evidence indicates that workplace mental health promotion interventions may provide modest short-term improvements in primary outcomes (well-being and selected burnout dimensions) and modest, inconsistent effects on secondary outcomes (job performance, work ability, perceived general health, sickness absence, occupational injuries). The narrative consistency of effects across trials was greater for psychological well-being and emotional exhaustion than for more distal occupational outcomes, where evidence was sparse and effect estimates frequently non-significant. However, considerable heterogeneity in intervention characteristics, implementation approaches, outcome measures, and follow-up periods limited direct comparison across studies and precluded quantitative pooling of results.

Implementation characteristics (substantially expanded): implementation outcomes were reported inconsistently across the included randomized controlled trials, and interpretation is constrained by incomplete reporting. Key implementation findings are as follows.

Participation and adherence: several studies described participant recruitment, attendance, or programme completion rates. Where participation rates were reported, they varied across studies, with workload, shift scheduling, staffing shortages, competing clinical responsibilities, and limited time availability commonly cited as barriers, particularly during pandemic-era or high-acuity periods. Programme completion was generally above 70% in trials that reported it, although denominators and definitions varied. Facilitators: where implementation factors were described, managerial encouragement, protected time for participation, organisational support, and flexible intervention delivery (including digital formats) were identified as important facilitators of engagement.

Intervention fidelity: few studies formally evaluated intervention fidelity or assessed whether programmes were delivered consistently as intended. Where fidelity was reported, it was typically described narratively rather than through formal indices. Managerial support, feasibility, and acceptability: reporting of managerial support, formal feasibility assessments, and participant acceptability was rare. A small subset of trials noted that managerial endorsement and protected time were associated with higher participation, but these observations were not derived from formal analyses.

Overall, the limited, non-standardized, and frequently narrative reporting of implementation characteristics across the 14 included trials restricts conclusions regarding the practical implementation, scalability, and sustainability of workplace mental health promotion interventions for nurses. This limitation is acknowledged in the Discussion and informs the priorities identified for future research.

Discussion     

Principal findings: this systematic review synthesized evidence from 14 randomized controlled trials evaluating workplace mental health promotion interventions among nurses [12]. Outcomes were synthesized hierarchically by predefined primary and secondary domains, intervention effects were interpreted in light of study-level risk-of-bias assessments, and individual-level and organizational-level interventions were distinguished [13].

Primary outcomes: across the primary outcome domains (well-being, burnout, work ability, job satisfaction, psychosocial risk factors), short-term improvements were observed in a subset of studies for well-being and selected burnout dimensions (notably emotional exhaustion), but findings were inconsistent and rarely sustained beyond immediate follow-up periods [14]. Reported effect estimates ranged from small to moderate, and several trials reported non-significant differences between intervention and control arms.

Secondary outcomes: evidence regarding job performance, perceived general health, sickness absence, occupational injuries, and implementation outcomes was limited and generally inconsistent.

Intervention-level findings: across the 14 included trials, individual-level interventions (mindfulness, resilience training, stress-management education, and similar psychological approaches) predominated. Organizational-level interventions (workload modifications, staffing changes, leadership-focused support, or changes to the psychosocial work environment) were comparatively underrepresented. As described separately in the subsections below, this imbalance, together with the different mechanisms through which the two intervention categories are expected to act, has implications for interpretation.

Setting-related findings: most trials were conducted in acute hospital or inpatient long-term care settings. No eligible randomized controlled trials were identified in home-based long-term care settings; consequently, no evidence-based conclusions can currently be drawn regarding the effectiveness of workplace mental health promotion interventions in these settings, and this finding represents an absence of evidence rather than evidence of ineffectiveness [15]. This is discussed further in the Generalisability subsection.

Collectively, these findings suggest that workplace mental health promotion interventions may offer limited short-term benefits for nurses, but the magnitude, consistency, and long-term sustainability of effects remain uncertain. An additional finding of this review is the limited reporting of implementation outcomes. Although barriers and facilitators to programme participation were predefined review objectives, relatively few randomized controlled trials systematically evaluated implementation characteristics such as recruitment, participation rates, adherence, intervention fidelity, feasibility, managerial support, or programme acceptability. Consequently, it remains difficult to determine whether variability in intervention effectiveness reflects differences in intervention content, implementation quality, or contextual workplace factors.

Interpretation of these findings should also consider the methodological quality of the included studies. While several trials demonstrated a low risk of bias, others exhibited methodological concerns particularly in the domains of missing outcome data and deviations from intended interventions that may have influenced the magnitude and consistency of reported intervention effects.

Strengths and limitations: this review has several strengths. It was conducted in accordance with PRISMA 2020 guidelines and included only randomized controlled trials, thereby prioritizing higher- quality evidence. Study selection and risk-of-bias assessment were performed independently by two reviewers using the Cochrane RoB 2 tool. Multiple international databases were searched using a structured strategy, increasing the likelihood of comprehensive study identification. Complete database-specific search strings are provided in Annex 1 to support reproducibility. Detailed study characteristics and quantitative findings are presented in Table 2.

Several limitations should be considered when interpreting the findings. First language and gray-literature restrictions. Restricting inclusion to English-language publications may have introduced language bias and limited representation of evidence from non-English-speaking contexts, potentially affecting global generalisability. Furthermore, the exclusion of gray literature (including conference proceedings, dissertations, preprints, government reports, and trial registries) may have increased the risk of publication bias, because studies reporting statistically significant or positive findings are more likely to be published in peer-reviewed journals than studies with null or negative results. Conference abstracts and trial registry records in particular may contain pilot or preliminary findings that could otherwise inform the evidence base. These restrictions may have reduced the comprehensiveness of the evidence base, and their potential influence on the magnitude and direction of synthesized effects should be explicitly considered when interpreting the conclusions of this review.

Second unretrieved reports: of the 81 reports sought for full-text retrieval, 47 (approximately 58%) could not be retrieved. Although eligibility could not be confirmed for these reports, their non-retrieval raises the possibility of selection bias if the missing studies differed systematically from the included studies (in intervention type, setting, or reported effect). Strategies to mitigate the risk of non-retrieval including multiple retrieval attempts, interlibrary loan requests, and author contact are described in the methods; nonetheless, the inability to retrieve the majority of sought full-text reports is a limitation that should be considered when interpreting the scope and representativeness of the evidence base.

Third heterogeneity and the absence of meta-analysis: substantial heterogeneity in intervention design, delivery format, intensity, outcome measures, and follow-up duration precluded quantitative synthesis. The justification for narrative rather than quantitative synthesis included clinical heterogeneity (diverse intervention approaches and healthcare settings), methodological heterogeneity (variation in intensity, duration, delivery mode, and outcome instruments), and small, non-comparable strata within individual intervention-outcome categories. Consequently, conclusions are based on narrative interpretation rather than pooled effect estimates, limiting precision in estimating overall effect magnitude and precluding formal subgroup analyses.

Fourth limited number of included trials: the relatively small number of included trials (n = 14) reduces the robustness of conclusions, particularly for less frequently reported outcomes such as sickness absence and occupational injuries. Fifth risk of bias: although several randomized controlled trials demonstrated a low overall risk of bias, others presented some concerns or high risk in specific domains particularly missing outcome data, deviations from intended interventions, and aspects of the randomization process. These methodological limitations reduce confidence in some reported intervention effects and should be considered when interpreting the overall evidence.

Sixth absence of GRADE evaluation: the certainty of the overall body of evidence was not formally assessed using the Grading of Recommendations Assessment, Development and Evaluation approach. Although the methodological quality of individual studies was evaluated using the Cochrane RoB 2 tool, the absence of a GRADE assessment means that the overall certainty of evidence across outcome domains could not be established. Consequently, confidence in the consistency and strength of the cumulative evidence remains uncertain. Future systematic reviews should incorporate GRADE assessments, together with quantitative synthesis (including subgroup or meta-regression analyses) where sufficient methodological homogeneity exists, to provide a more comprehensive evaluation of the certainty of evidence and to strengthen evidence-based recommendations.

Seventh absence of prospective registration: the review protocol was not prospectively registered in an international registry such as PROSPERO. Although the review methods were predefined before study screening and were applied consistently throughout the review, prospective registration would have enhanced methodological transparency and reduced concerns regarding selective reporting. Future systematic reviews in this area should be prospectively registered, and the present authors provide the rationale that the review was initiated before protocol registration was considered; this rationale is offered transparently and should itself be considered when interpreting the review's methodological rigor.

Finally, the decision not to perform a meta-analysis should be interpreted in light of the substantial clinical and methodological diversity across the included trials. Although this approach limited quantitative estimation of overall intervention effects, narrative synthesis was considered the most appropriate method for accurately representing the available evidence without producing potentially misleading pooled estimates.

Interpretation in context of existing evidence: the pattern of modest and predominantly short-term effects identified in this review is consistent with findings from broader systematic reviews conducted among healthcare professionals [16]. Previous evidence has shown that workplace mental health interventions can improve psychological well-being and reduce burnout, although the magnitude and sustainability of these effects vary across intervention types and implementation contexts [17].

Individual-level interventions: most included studies evaluated individual-level interventions, including mindfulness training, resilience enhancement programmes, stress-management education, cognitive-behavioural approaches, and other psychological skills training. These interventions primarily targeted proximal psychological outcomes such as stress, emotional exhaustion, resilience, and perceived well-being. Across the included trials, individual-level interventions generally demonstrated modest short-term improvements in well-being and selected dimensions of burnout. However, improvements were inconsistent across studies and were rarely maintained during longer follow-up periods.

These findings are consistent with previous systematic reviews suggesting that individual psychological interventions are effective for improving coping capacity but have limited influence on broader workplace conditions that contribute to occupational stress [18]. Because most included trials in this category reported small-to-moderate, short-term effects on individual-level psychological outcomes, the present review supports the view that individual-level strategies alone are unlikely to produce sustained improvements in broader occupational outcomes (such as work ability, sickness absence, and occupational injuries) that are driven by structural workplace factors.

Organizational-level interventions: in contrast, relatively few included studies evaluated organizational-level strategies such as modifications to work processes, workload management, staffing arrangements, leadership support, or broader psychosocial work environment improvements. Organizational interventions aim to address structural determinants of occupational stress, including workload, scheduling, decision latitude, staffing, and leadership support, rather than individual coping behaviours [19]. Because comparatively few randomized controlled trials examined these approaches in nursing populations during the review period, evidence regarding their effectiveness remains limited. Nevertheless, existing literature suggests that organizational changes may have greater potential to produce sustained improvements in work environments, reduce psychosocial risk factors, and improve workforce outcomes when implemented alongside individual-level interventions. The present findings are consistent with this view but cannot confirm it directly because of the limited number of organizational-level trials in the included evidence base; this represents a clear priority area for future research and is reflected in the future research subsection [20].

Heterogeneity narrative framework: to facilitate interpretation of the variability across studies, the findings can be understood within a narrative framework based on four principal sources of heterogeneity, consistent with the framework described in the methods section: intervention characteristics; intervention duration and implementation intensity; healthcare setting; and outcome measurement [21]. The observed heterogeneity is attributable to the cumulative effect of these four sources rather than to a single underlying factor, which is reflected in the decision to synthesize evidence narratively and supports the case for greater standardization of intervention components, outcome measures, and follow-up periods in future research. Taken together, the findings indicate that workplace mental health promotion is unlikely to be optimised through individual-level strategies alone [22]. Future interventions should integrate psychological support with organisational initiatives that address workload, staffing, leadership, and workplace culture to achieve more sustainable improvements in nurses' mental health and broader occupational outcomes.

Generalisability: the limited evidence for work ability, general health, sickness absence, and occupational injuries suggests that mental health promotion strategies alone may be insufficient to influence more distal occupational or service-related outcomes. Multifaceted interventions integrating psychosocial support with organizational and structural changes may be necessary to produce broader and more sustained improvements [23].

Most included trials were conducted in acute hospital and inpatient long-term care settings. Therefore, findings are primarily applicable to institutional healthcare environments. No eligible randomized controlled trials were identified in home-based long-term care settings [24]. This is an evidence gap, that is, the absence of eligible studies rather than empirical evidence of reduced effectiveness; consequently, no conclusions can currently be drawn regarding the effectiveness of workplace mental health promotion interventions in these settings.

This finding represents an absence of evidence rather than evidence of lack of effectiveness and highlights an important priority for future research [25]. In addition, variations in healthcare systems, staffing models, workplace cultures, and regulatory environments across countries may influence both implementation and effectiveness. Caution is therefore warranted when applying these findings across different healthcare contexts, including home-based long-term care, primary care, and community settings, and conclusions should be drawn conservatively.

Implications for practice and policy: the findings indicate that workplace mental health promotion interventions for nurses may yield modest short-term improvements in well-being and burnout-related outcomes. However, variability in intervention design and implementation limits definitive conclusions regarding optimal strategies. Healthcare organizations should consider context-specific, evidence-informed approaches that integrate both individual-level and organizational-level components. Successful implementation is likely to depend not only on intervention content but also on organisational readiness, leadership commitment, managerial support, protected participation time, adequate staffing, flexible programme delivery, and monitoring of intervention fidelity. Attention to these implementation factors may improve workforce engagement, programme adherence, scalability, and long-term sustainability of workplace mental health promotion initiatives. Attention to implementation processes, workforce engagement, and sustainability may enhance the effectiveness of workplace mental health programs.

Future research: future research should prioritize adequately powered randomized controlled trials with longer follow-up periods to assess sustainability of effects. Greater standardization of outcome measures would enhance comparability across studies and support future quantitative synthesis. Both individual-level and organizational-level interventions should be evaluated in adequately powered trials to clarify their relative and combined impact, and head-to-head or factorial designs would strengthen inference. Systematic assessment of implementation processes, including documented participation rates, adherence assessed through formal indices, intervention fidelity, feasibility, acceptability, adoption, and sustainability, is also needed across the full review base rather than in a small subset of trials. Future randomized controlled trials should incorporate established implementation science frameworks and report implementation outcomes alongside effectiveness outcomes to improve understanding of factors influencing intervention success and facilitate translation into routine nursing practice. Importantly, research in underrepresented settings, particularly home-based long-term care, as well as primary care and community nursing, is essential to address current evidence gaps. Improved methodological rigor and transparent reporting will strengthen the evidence base and inform policy and practice more effectively. Future reviews should also incorporate formal GRADE certainty assessments, standardized reporting of intervention characteristics and outcome measures, and meta-analyses or subgroup analyses (for example, by intervention category or setting) where sufficient methodological homogeneity permits quantitative synthesis.

Conclusion     

This systematic review synthesised evidence from 14 randomized controlled trials evaluating workplace mental health promotion interventions among nurses. Interventions were categorized as individual-level or organizational-level, outcomes were prioritized as primary and secondary, and findings were interpreted in light of study-level risk-of-bias assessments. Overall, interventions demonstrated modest and predominantly short-term improvements in primary well-being and burnout-related outcomes, with inconsistent and limited evidence regarding job-related, work ability, general health, sickness absence, and occupational injury outcomes. The certainty of the underlying evidence could not be formally quantified in the absence of a GRADE assessment. Evidence regarding implementation processes was also limited because relatively few studies comprehensively reported participation rates, intervention adherence, implementation fidelity, feasibility, managerial support, or other factors influencing programme delivery and sustainability. No eligible randomized controlled trials were identified in home-based long-term care settings. Therefore, the current evidence is insufficient to determine whether workplace mental health promotion interventions are effective in these settings. This absence of evidence should not be interpreted as evidence of lack of effectiveness but rather as an important research gap requiring future investigation. These findings suggest that workplace mental health promotion interventions may provide limited short-term benefits for nurses, but the magnitude, sustainability, and broader occupational impact remain uncertain. Future research should focus on adequately powered trials with longer follow-up periods, standardized outcome measures, and greater emphasis on organizational-level strategies and implementation processes. Expanding research into underrepresented settings, particularly home-based long-term care, is essential to inform comprehensive and context-specific workplace mental health strategies.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

All of the authors contributed to concept analysis; Thika Marliana, Marchellino Edmond Aprico Limanauw, and Oriza Sativa Manurung arranged the manuscript; Budi Anna Keliat, Herni Susanti, Yossie Susanti Eka Putri, and Henny Permatasari reviewed the manuscript and checked for final submission. All the authors read and approved the final version of this manuscript.

Acknowledgments     

The authors would like to thank the University of Indonesia for the support and funding provided to facilitate the completion of this research.

Tables and figures     

Table 1: primary and secondary outcome domains and prioritization rationale for a systematic review of workplace mental health promotion among nurses in clinical healthcare settings (eligible publications: January 2019-30 July 2024)

Table 2: Cochrane Risk of Bias 2 (RoB 2) assessments of 14 randomized controlled trials of workplace mental health promotion among nurses in clinical healthcare settings (January 2019-30 July 2024)

Table 3: characteristics, interventions, comparators, outcomes, and follow-up of workplace mental health promotion trials among nurses in clinical healthcare settings (January 2019-30 July 2024)

Figure 1: preferred reporting items for systematic reviews and meta-analyses (PRISMA) 2020 flow diagram of workplace mental health promotion trials among nurses in clinical healthcare settings (January 2019-30 July 2024)

Figure 2: Cochrane Risk of Bias 2 (RoB 2) traffic-light plot for 14 workplace mental health promotion trials among nurses in clinical healthcare settings (January 2019-30 July 2024)

Annex     

Annex 1: supplementary materials (PDF KB)

References