Structural and psychological empowerment in relation to nurse job satisfaction and burnout: a systematic review and meta-analysis

Abstract

Objective

To concurrently evaluate the impact of both structural empowerment (SE) and psychological empowerment (PE) on the dual outcomes of nurse job satisfaction and burnout, and to compare findings with published meta-analyses to clarify incremental value.

Methods

Systematic searches were conducted in databases including PubMed, Embase, Web of Science, CINAHL, and the Cochrane Library from inception until September 30, 2025. Cross-sectional studies reporting correlation coefficients between SE/PE and satisfaction/burnout were included. Effect sizes were pooled using random-effects models. Subgroup analyses by geographical region were performed, and publication bias was assessed using a triple-testing method.

Results

Thirty-seven studies involving 18,104 participants were included. SE demonstrated a moderate positive correlation with job satisfaction (r = 0.52), a negative correlation with emotional exhaustion (r = -0.25), but a positive correlation with depersonalization (r = 0.27). PE showed a weaker correlation with satisfaction (r = 0.33) and no significant associations with any burnout dimensions. Effect sizes were significantly larger in Chinese samples compared to those from Europe and North America. Trim-and-fill analyses indicated robust results.

Conclusion

Across predominantly cross-sectional studies, higher empowerment—particularly structural empowerment—was consistently associated with higher nurse job satisfaction and lower emotional exhaustion. Associations with other burnout dimensions were less consistent (including an unexpected positive correlation with depersonalization), warranting cautious interpretation and careful attention to measurement/scoring harmonization. Overall, the findings indicate correlational relationships rather than causal effects and can inform the design of context-appropriate empowerment strategies and future longitudinal research.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-026-14080-6.

Introduction

As the backbone of the healthcare system, nurses’ job satisfaction and burnout strongly influence care quality and patient experience [1]. Nurse shortages and turnover remain major global challenges, and job satisfaction (JS) and burnout (BO) are core correlates of retention intention and care quality [2]. Although global evidence documents substantial burnout burden and marked cross-regional variation [3, 4], and overall nurse job satisfaction remains generally low [5], these patterns highlight a practical need to identify modifiable levers that organizations can target to improve JS and reduce BO.

Empowerment has therefore attracted increasing attention as an intervenable construct in nursing management [6]. In nursing research, empowerment is commonly conceptualized as two complementary pathways: structural empowerment (SE), emphasizing access to information, resources, support, and opportunities within the organization [7], and psychological empowerment (PE), reflecting perceived meaning, competence, self-determination, and impact at work [8]. However, prior reviews often focused on a single empowerment dimension or examined satisfaction and burnout separately, leaving uncertainty about which pathway is more consistently associated with JS and BO and what this implies for intervention targets [9].

Accordingly, this systematic review and meta-analysis asked: (1) What are the pooled associations between SE and nurses’ JS and BO? and (2) What are the pooled associations between PE and nurses’ JS and BO? We hypothesized that SE would show a stronger positive association with JS than PE, and that both SE and PE would be inversely associated with BO, while burnout subdimensions were examined exploratorily given expected measurement variability.

Methods

Study design

The investigation was structured as a systematic review with integrated meta-analysis, adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Because all eligible primary studies reported cross-sectional associations, results are interpreted as correlational and do not support causal inference.

A review protocol was developed a priori but was not prospectively registered in PROSPERO. To minimize data-driven analytical decisions, the eligibility criteria, primary outcomes, planned subgroup analyses (geographical region), and sensitivity analyses were specified before data extraction and synthesis.

Search strategy

A comprehensive literature search was executed across five major electronic databases: PubMed, Embase, Web of Science, CINAHL, and Cochrane Library. The search encompassed all available records from database inception through September 30, 2025, with no filters applied for language or publication status. To optimize retrieval, the search methodology incorporated both controlled vocabulary (e.g., MeSH terms) and complementary free-text keywords, ensuring maximal search sensitivity while maintaining specificity.

The complete, database-specific search strategies (including full search strings for each database) are provided in Supplementary Table S1 to ensure reproducibility. Grey literature sources (e.g., theses, reports, preprints, and unpublished studies) and trial registries were not systematically searched; therefore, potentially relevant unpublished evidence may have been missed.

Search terms revolved around the four core concepts of this study: structural empowerment, psychological empowerment, job satisfaction, and burnout.

The search terms included: (1) Structural Empowerment: “structural empowerment”, “organizational empowerment”, “work empowerment”, “empowering work environment”; (2) Psychological Empowerment: “psychological empowerment”, “perceived empowerment”, “employee empowerment”; (3) Job Satisfaction: “job satisfaction”, “work satisfaction”, “satisfaction”; (4) Burnout: “burnout”, “emotional exhaustion”, “depersonalization”, “cynicism”, “professional burnout”.

Additionally, the reference lists of included studies were scrutinized (i.e., “snowballing”) to identify potentially omitted literature.

Inclusion criteria

(1) Study Type: Observational studies (e.g., cross-sectional, cohort, case-control) reporting relationships between the variables of interest were included. Conference abstracts, reviews, case reports, and theoretical articles were excluded.

(2) Participants: The study population must comprise registered nurses engaged in clinical nursing practice, irrespective of department, years of experience, or educational level. Studies including physicians, other healthcare professionals, or nursing students were excluded unless data specific to nurses could be extracted separately.

(3) Measurement Tools: Included studies were required to employ psychometrically validated instruments for assessing all core constructs. Standardized measures such as the Conditions of Work Effectiveness Questionnaire (CWEQ) for structural empowerment, the Psychological Empowerment Scale (PES) for psychological empowerment, the Minnesota Satisfaction Questionnaire (MSQ) for job satisfaction, and the Maslach Burnout Inventory (MBI) for burnout represented acceptable assessment tools.

(4) Data Reporting: Studies must have reported the correlation coefficient (r), sample size (n), or other statistics convertible to r for at least one pair of the variables of interest.

Exclusion criteria were: (1) Studies irrelevant to the core research question (relationship between empowerment and job satisfaction/burnout) or failing to measure any core variable; (2) Studies lacking sufficient data for effect size calculation, even after attempting to contact the authors (e.g., reporting only p-values without specific statistics or confidence intervals); (3) Duplicate publications or studies with substantial data overlap from the same research team; in such cases, the most comprehensive publication with the largest sample size was included.

Study selection and data extraction

The literature screening process adhered to a four-phase framework: identification, screening, eligibility, and inclusion. During the identification phase, records retrieved from databases were consolidated in EndNote, with duplicate entries removed. In the screening phase, two independent reviewers evaluated titles and abstracts against predetermined eligibility criteria, excluding clearly non-conforming studies. Records deemed potentially relevant advanced to the eligibility phase, where full-text articles underwent independent assessment by both reviewers. Any discrepancies in inclusion decisions were resolved through consensus or arbitration by a third reviewer. The inclusion phase culminated in the finalization of studies for quantitative synthesis.

A standardized data extraction form was developed a priori using Microsoft Excel. Extracted elements encompassed: (1) bibliographic details (first author, publication year, country/region); (2) methodological attributes (study design, sample size); (3) population characteristics (age, gender, professional experience, clinical department); (4) measurement instruments (specific scales used for each construct); (5) outcome statistics (correlation coefficients with corresponding 95% confidence intervals and p-values for all relevant variable pairs); (6) risk-of-bias/quality assessment items; and (7) for burnout outcomes, the exact instrument version/subscale (e.g., MBI-HSS vs. MBI-GS) and scoring direction. Before pooling, correlation coefficients were harmonized so that higher burnout scores consistently indicated higher burnout; when a study reported reverse-scored or positively framed burnout indicators, correlations were reverse-coded (multiplied by -1) prior to synthesis. Extracted study-level effect sizes (correlations), sample sizes, empowerment type, measurement instruments, and burnout dimensions are presented in Supplementary Table S2. Instrument details and direction harmonization decisions are reported in Supplementary Table S3. When key statistics required to compute effect sizes were not reported, corresponding authors were contacted by email (up to two attempts, two weeks apart). If data remained unavailable, the study was excluded and the reason documented in the PRISMA flow diagram.

Quality assessment

Methodological rigor was evaluated using the Newcastle-Ottawa Scale (NOS) with an a priori adaptation for cross-sectional studies. The adapted NOS covered three domains: Selection (maximum 5 points), Comparability (maximum 2 points), and Outcome/Exposure (maximum 3 points), yielding a total score range of 0–10. Two reviewers independently rated each study, with disagreements resolved by discussion or a third reviewer. For transparency, item-level ratings for each study are provided in Supplementary Table S4 rather than total scores alone. Studies were categorized as high quality (≥ 7 points), moderate quality (5–6 points), or low quality (≤ 4 points).

Statistical analysis

The primary effect measure was the Pearson correlation coefficient (r) with its 95% confidence interval. For each included study, correlation coefficients (r) and 95% CIs were calculated or derived for the four variable pairs: structural empowerment with job satisfaction, structural empowerment with burnout, psychological empowerment with job satisfaction, and psychological empowerment with burnout. For studies not reporting CIs, Fisher’s z-transformation was applied to normalize the correlation coefficients and stabilize variances before constructing CIs.

Given anticipated population heterogeneity across studies, random-effects models were employed for all meta-analyses, providing more conservative and generalizable results. Between-study heterogeneity was quantified using Cochran’s Q (with p-values), I², and τ²; these indices are reported for each pooled estimate and summarized in Supplementary Table S5. To assess potential publication bias (the tendency for positive results to be published more readily), multiple methods were used. Initially, funnel plots were visually inspected for asymmetry. Subsequently, Egger’s linear regression test and Begg’s rank correlation test were conducted quantitatively (where appropriate); a p-value > 0.05 suggested no statistically detectable small-study effects. If bias was suggested, the trim-and-fill method was applied to impute potentially missing studies and recalculate the pooled effect size.

To explore heterogeneity, we conducted subgroup analyses by geographical region. In addition, we pre-specified sensitivity analyses including (a) leave-one-out analyses for outcomes with ≥ 5 studies, (b) excluding studies rated as low quality (NOS ≤ 4), and (c) restricting analyses to comparable burnout measures (e.g., MBI-based outcomes; and where possible, stratifying by MBI version such as HSS vs. GS). Meta-regression was planned for outcomes with sufficient studies (typically ≥ 10) and consistent reporting of candidate moderators; however, because potential moderators (e.g., department type, staffing intensity, or detailed cultural indicators) were inconsistently reported across studies, formal meta-regression was not undertaken and findings are presented as exploratory. All statistical analyses and figure generation were performed using R software.

Results

Study selection process

The database search identified 873 records. After removal of duplicates, 312 unique records remained for title/abstract screening. Of these, 270 records were excluded as clearly irrelevant, leaving 42 full-text articles assessed for eligibility. Five studies were excluded at the full-text stage because effect size data could not be obtained despite author contact, resulting in 37 studies included in the quantitative synthesis (Fig. 1).

Fig. 1.

Flow diagram of the study selection process

Characteristics of included studies

The 37 included studies encompassed a total of 18,104 participants from various countries and regions, including Canada, Spain, Italy, Iran, China, Sweden, Thailand, Malaysia, and Pakistan. Sample sizes ranged from 65 to 3,156. Participant ages varied widely, from 18 to 44 years, though some studies did not report specific age data. The methodological quality, assessed using the NOS, was generally high, with the majority of studies (n = 28) scoring 7 points or above (Table 1).

Table 1.

Basic characteristics of included studies

ID	Author	Region/Country	Sample Size	Age	NOS
1	Dahinten VS 2016 [10]	Canada	1007	42	9
2	Trillo A 2025 [11]	Spain	150	46	8
3	Kelly C 2022 [12]	Italy	83	25–30	9
4	Jafarian ASR 2023 [13]	Iran	200	33.3	9
5	Laschinger HK 2013 [14]	Canada	272	43.89	8
6	Zhang SL 2014 [15]	China	203	27	6
7	Meng L 2016 [16]	China	244	18–30	8
8	Guo JJ 2013 [15]	China	350	30.21	7
9	Hochwälder J 2007 [17]	Sweden	838	42.6	8
10	Spence L HK 2009 [18]	Canada	612	-	7
11	Manojlovich M 2002 [19]	Canada	347	40	8
12	Spence LHK 2011 [20]	Canada	3156	42	7
13	Laschinger HK 2001 [21]	Canada	273	44	8
14	Wang X 2013 [22]	Thailand	385	18–35	8
15	Guo JJ 2016 [23]	China	1002	30.21	7
16	Orgambídez RA 2017 [24]	Spain	297	37.42	9
17	Jin C 2025 [25]	China	1225	26–30	6
18	Permarupan PY 2020 [26]	Malaysia	432	31–41	8
19	Ardabili FS 2020 [27]	Iran	138	33.5	7
20	Ouyang YQ 2015 [28]	China	726	26–30	6
21	Tourangeau A 2010 [29]	Canada	675	44	6
22	Ahmad N 2010 [30]	Malaysia / UK (England)	388/168	21–30	7
23	Orgambídez A 2024 [31]	Spain	439	40.72	8
24	Asif M 2019 [32]	Pakistan	386	21–30	9
25	Bawafaa E 2015 [33]	Canada	1216	41.5	7
26	Cai CF 2009 [34]	China	189	30.45	6
27	Cai CF 2011 [35]	China	208	30.4	7
28	Choi S 2019 [36]	South Korea	208	28.8	7
29	de Almeida H 2019 [37]	Spain	151	44.04	7
30	Lautizi M 2009 [38]	Italy	120	42	8
31	Li IC 2013 [39]	China	65	34.4	9
32	Lyden C 2018 [40]	USA (Portland, OR)	141	-	9
33	Ning S 2009 [41]	China	598	30.77	6
34	Ta’an WF 2022 [42]	Jordan	126	30	7
35	Wong CA 2013 [43]	Canada	280	43.4	8
36	Yang J 2014 [44]	China	524	30.2	7
37	Janighorban M 2020 [45]	Iran	282	34.78	8

Structural empowerment and job satisfaction

This analysis included data from 22 independent cohorts (reported in 21 articles), involving 11,087 participants. The correlation coefficients (r) between structural empowerment and job satisfaction ranged from 0.34 to 0.76. Significant heterogeneity was observed; therefore, a random-effects model was employed. The meta-analysis revealed a moderate positive correlation between structural empowerment and job satisfaction: pooled r = 0.52 (95% CI: 0.46 to 0.72). The prediction interval (0.25 to 0.72) suggests robustness of the average effect (Fig. 2). Heterogeneity statistics (I², τ², Q, p) for this analysis are summarized in Supplementary Table S5.

Fig. 2.

Forest plot of the correlation between structural empowerment and job satisfaction

A three-pronged approach was employed to evaluate publication bias: visual examination of funnel plots, Egger’s regression test, and trim-and-fill analysis. While Egger’s test showed no significant bias (t = 1.788, p = 0.089), the trim-and-fill method indicated the potential absence of 9 studies, resulting in an adjusted pooled r of 0.46, a minimal change indicating robust findings (Figure 3A and B).

Fig. 3.

Funnel plots for publication bias regarding structural empowerment and job satisfaction. (A: Original studies; B: After trim-and-fill adjustment)

Subgroup analysis was conducted to explore the relationship between structural empowerment and job satisfaction across different geographical regions. The number of studies was highest for Canada (n = 7), followed by China (n = 5). The pooled correlation coefficients varied significantly across regions. (Table 2)

Table 2.

Subgroup analysis of structural empowerment and job satisfaction by region

Region	Number of Studies	Sample size	I2	Pooled r	Lower 95%CI	Upper 95%CI
Canada	7	7293	80.6%	0.46	0.40	0.51
Italy	2	203	89.1%	0.65	-0.97	1.00
Malaysia	1	388	-	0.34	0.25	0.42
UK (England)	1	168	-	0.62	0.52	0.71
Spain	2	439	0	0.47	0.00	0.77
Pakistan	1	386	-	0.48	0.40	0.55
China	5	1584	88.8%	0.57	0.42	0.69
South Korea	1	208	-	0.57	0.47	0.66
USA (Portland, OR)	1	141	-	0.73	0.64	0.80
Jordan	1	126	-	0.37	0.21	0.51

Psychological empowerment and job satisfaction

This analysis incorporated data from 10 independent cohorts (reported in 9 articles), comprising 4,103 participants. A random-effects meta-analysis, employed due to significant heterogeneity, synthesized correlations spanning 0.13 to 0.57, revealing a weak positive pooled association between psychological empowerment and job satisfaction (r = 0.33, 95% CI: 0.22–0.44). The prediction interval (-0.04 to 0.62) indicates that the true effect may vary across settings but corroborates the direction of the average relationship (Fig. 4). Heterogeneity statistics (I², τ², Q, p) are summarized in Supplementary Table S5.

Fig. 4.

Forest plot of the correlation between psychological empowerment and job satisfaction

Assessment for publication bias included a funnel plot, Egger’s test, and the trim-and-fill method. Egger’s test showed no significant bias (t = 0.884, p = 0.403). However, imputation of 4 potentially missing studies via the trim-and-fill method reduced the pooled r to 0.21, indicating that the results for this relationship may be less robust. Funnel plots depicting publication bias for psychological empowerment and job satisfaction are shown in Fig. 5A (original) and B (trim-and-fill adjusted).

Fig. 5.

Funnel plots for publication bias regarding psychological empowerment and job satisfaction. (A: Original studies; B: After trim-and-fill adjustment)

Subgroup analysis by region was performed to further investigate the relationship between psychological empowerment and job satisfaction. Canada contributed the most studies (n = 3), followed by China and Spain (n = 2 each). Significant differences in the pooled correlation coefficients were observed among regions (Table 3).

Table 3.

Subgroup analysis of psychological empowerment and job satisfaction by region

Region	Number of Studies	Sample size	I2	Pooled r	Lower 95%CI	Upper 95%CI
Canada	3	2029	95.3%	0.30	-0.10	0.61
Iran	1	138	-	0.17	0.00	0.33
Malaysia	1	388	-	0.33	0.24	0.42
UK (England)	1	168	-	0.57	0.46	0.67
Spain	2	589	93.5%	0.36	-0.96	0.99
China	2	791	7.3%	0.25	-0.28	0.66

Structural empowerment and burnout

Three studies reported correlations between structural empowerment and the overall MBI score, encompassing 741 participants. Correlation coefficients ranged from − 0.46 to 0.04. The presence of substantial heterogeneity necessitated the application of a random-effects model for all meta-analytic computations. The meta-analysis indicated no significant association: pooled r = -0.28 (95% CI: -0.76 to 0.40). The prediction interval (-0.93 to 0.80) suggests instability in the average effect (Fig. 6).

Fig. 6.

Forest plot of the correlation between structural empowerment and the total MBI score

Visual assessment of the funnel plot identified asymmetry, suggesting the possible presence of publication bias. The Trim-and-Fill method imputed zero missing studies; however, the limited number of included studies and underlying heterogeneity prevent definitive conclusions regarding publication bias risk (Fig. 7).

Fig. 7.

Funnel plot for publication bias regarding structural empowerment and the total MBI score

Analyses of burnout subdimensions revealed distinct associations. Nine studies examined cynicism, yielding a pooled r of -0.33 (95% CI: -0.34 to -0.09; prediction interval: -0.56 to 0.18). Nine studies examined emotional exhaustion, showing a pooled r of -0.25 (95% CI: -0.31 to -0.18; prediction interval: -0.42 to 0.05). Six studies examined depersonalization, demonstrating a pooled r of 0.27 (95% CI: 0.12 to 0.41; prediction interval: -0.13 to 0.60). Analysis of burnout subdimensions revealed that structural empowerment was inversely associated with emotional exhaustion but positively linked to depersonalization. However, the prediction interval for depersonalization crossing zero indicates limited generalizability, requiring cautious interpretation (Fig. 8). All burnout outcomes were harmonized to a common direction prior to pooling (Supplementary Table S3). To address potential measurement-related inconsistencies, we conducted sensitivity analyses excluding studies with unclear scoring direction or using different burnout instruments/MBI versions; results are summarized in Supplementary Table S6.

Fig. 8.

Forest plots of the correlations between structural empowerment and key burnout dimensions. A: Cynicism; B: Emotional exhaustion; C: Depersonalization

Egger’s tests revealed no significant publication bias for any subdimension (all p > 0.513). Trim-and-Fill analysis for cynicism imputed 2 studies, reducing the pooled r from − 0.22 to -0.29, indicating robustness. For emotional exhaustion, imputation of 3 studies reduced the pooled r from − 0.25 to -0.20, suggesting relative robustness. For depersonalization, imputation of 3 studies reduced the pooled r from 0.27 to 0.13, indicating potential instability (Fig. 9).

Fig. 9.

Funnel plots for publication bias regarding structural empowerment and key burnout dimensions. A: SE vs. Cynicism (Original); B: SE vs. Cynicism (Trim-and-fill adjusted); C: SE vs. Emotional Exhaustion (Original); D: SE vs. Emotional Exhaustion (Trim-and-fill adjusted); E: SE vs. Depersonalization (Original); F: SE vs. Depersonalization (Trim-and-fill adjusted)

Psychological empowerment and burnout

Four studies investigated the correlation between psychological empowerment and the overall MBI score (n = 2,101). Correlation coefficients ranged from − 0.58 to 0.00. Significant heterogeneity warranted a random-effects model, showing no significant association: pooled r = -0.36 (95% CI: -0.70 to 0.11). The prediction interval (-0.89 to 0.60) indicates substantial uncertainty (Fig. 10).

Fig. 10.

Forest plot of the correlation between psychological empowerment and the total MBI score

Funnel plot asymmetry suggested potential publication bias. Trim-and-Fill analysis imputed one missing study, increasing the pooled r to -0.28 and indicating result instability (Fig. 11).

Fig. 11.

Funnel plots for publication bias regarding psychological empowerment and the total MBI score. (A: Original studies; B: After trim-and-fill adjustment)

Subdimension analysis included three studies on cynicism (pooled r = -0.21; 95% CI: -0.51 to 0.13; prediction interval: -0.69 to 0.40), five studies on emotional exhaustion (pooled r = -0.36; 95% CI: -0.66 to 0.05; prediction interval: -0.88 to 0.57), and three studies on depersonalization (pooled r = 0.34; 95% CI: -0.23 to 0.73; prediction interval: -0.65 to 0.90). While negative and positive trends were observed for emotional exhaustion and depersonalization respectively, none reached statistical significance (Fig. 12). As above, instrument versions and scoring direction were extracted and harmonized (Supplementary Table S3), and sensitivity analyses addressing instrument differences are provided in Supplementary Table S6.

Fig. 12.

Forest plot of the correlations between psychological empowerment and key burnout dimensions

For these subdimensions, limited study numbers precluded formal Egger’s tests. Visual funnel plot inspection and Trim-and-Fill analysis were conducted. For cynicism and depersonalization, zero studies were imputed, though publication bias cannot be fully excluded. For emotional exhaustion, imputation of one study reduced the pooled r from − 0.36 to -0.24, indicating instability. The corresponding funnel plots are presented in Fig. 13.

Fig. 13.

Funnel plots for publication bias regarding psychological empowerment and key burnout dimensions. A: PE vs. Cynicism (Original); B: PE vs. Emotional Exhaustion (Original); C: PE vs. Emotional Exhaustion (Trim-and-fill adjusted); D: PE vs. Depersonalization (Original)

Discussion

This review highlights a nuanced pattern across empowerment domains and outcomes. Structural empowerment showed a moderate positive association with job satisfaction and an inverse association with emotional exhaustion, whereas psychological empowerment showed a weaker association with satisfaction and no statistically significant association with overall burnout. Importantly, associations differed across burnout dimensions and exhibited substantial heterogeneity, indicating that empowerment effects are not uniform across contexts or outcomes.

The unexpected positive association between structural empowerment and depersonalization warrants careful interpretation. From a job demands–resources perspective, structural empowerment may increase nurses’ decision latitude and access to resources, but may also be accompanied by higher role demands, responsibility, and exposure to conflict or emotionally taxing situations. Additionally, structural empowerment may enhance innovative behavior through perceptions of decent work [46], yet may also inadvertently contribute to emotional distancing under high demands. In such settings, depersonalization (or emotional distancing) can function as a short-term coping strategy to manage sustained emotional labor. Alternatively, this pattern may reflect measurement and contextual artifacts (e.g., differences in burnout instruments or scoring approaches, or unit cultures in which empowerment is implemented alongside intensified workload). Although we harmonized scoring directions prior to pooling, the attenuation observed after trim-and-fill for depersonalization suggests additional uncertainty and reinforces the need for transparent reporting of burnout measures and for prospective designs that can disentangle directionality.

Geographical differences observed in subgroup analyses may reflect broader cultural and organizational context. For example, cross-national differences in power distance, collectivism–individualism, and uncertainty avoidance could shape how empowerment is perceived and how autonomy and accountability translate into satisfaction or burnout symptoms [47]. Additionally, cultural factors may moderate the relationship between structural empowerment and compassion fatigue, as observed in Chinese nurses [48]. Because the primary studies rarely reported standardized cultural indicators, we treated region as a coarse proxy and did not conduct formal culture-based moderator analyses; future meta-analyses should incorporate explicit cultural metrics and organizational characteristics to test these hypotheses.

Our results align broadly with previous meta-analyses, reinforcing the significance of empowerment for nurse occupational health. Yesilba’s meta-analysis of 11,078 nurses reported a significant positive correlation between SE and job satisfaction (r = 0.559) [9], closely matching our pooled estimate (r = 0.52) and supporting the stable, cross-cultural impact of SE. For PE, Gu et al.‘s (2022) meta-analysis of 7,664 nurses found a moderate positive correlation with job satisfaction (r = 0.55) [49], slightly higher than our finding (r = 0.33). This discrepancy may be attributable to the predominant focus on Asian populations in Gu et al.‘s work, whereas our analysis incorporated more Western studies, suggesting cultural differences may moderate the PE-satisfaction relationship. Additionally, Şenol Çelik et al. (2023) reported a significant negative correlation between SE and emotional exhaustion, while PE was suggested to indirectly alleviate burnout by enhancing competence and autonomy [50]. Our study confirms the SE-emotional exhaustion link but found no significant direct relationship between PE and overall burnout, possibly due to PE’s indirect mechanism and highlighting the need for future research to explore mediating variables like self-efficacy and organizational support.

Unlike previous meta-analyses that examined either SE or PE in isolation, or focused solely on either satisfaction or burnout, our study concurrently integrates both empowerment constructs and both outcomes. This approach yields novel, nuanced evidence: firstly, SE demonstrates the strongest association with improved job satisfaction; secondly, it may paradoxically intensify depersonalization in high-demand environments; thirdly, the effect sizes differ markedly between China and Western countries. In essence, this research refines the simplistic mantra that ‘empowerment is beneficial’ into a precise, operational guideline emphasizing the need for dimension-specific, outcome-specific, and culture-specific empowerment strategies—an advancement not previously achieved in the meta-analytic literature.

Several limitations warrant consideration. First, the predominance of cross-sectional studies precludes causal inference and may be vulnerable to common-method bias. Second, although core eligibility criteria and analyses were specified a priori, the protocol was not prospectively registered. Third, grey literature and trial registries were not systematically searched, which may increase the risk of publication and selective reporting bias. Indeed, trim-and-fill adjustments materially changed some pooled estimates (notably psychological empowerment–job satisfaction and depersonalization-related outcomes), indicating that small-study effects cannot be ruled out. Fourth, substantial heterogeneity and wide prediction intervals for several outcomes suggest that pooled averages may not generalize to all settings; while we conducted pre-specified sensitivity analyses, moderator testing was limited because key contextual variables (e.g., department type, staffing, and cultural indicators) were inconsistently reported and meta-regression was not feasible. Future research should prioritize longitudinal and intervention designs, report measurement/scoring direction transparently, and collect standardized contextual variables to enable robust moderator analyses.

Conclusion

This systematic review and meta-analysis indicates that structural empowerment is more consistently associated with higher nurse job satisfaction than psychological empowerment. Associations with burnout were dimension-specific: structural empowerment was linked to lower emotional exhaustion, whereas associations with cynicism and depersonalization were less stable and, in the case of depersonalization, unexpectedly positive. Given the cross-sectional nature of the evidence base and the presence of publication bias signals for some outcomes, these findings should be interpreted as correlational and context-dependent rather than causal.

For practice, empowerment initiatives that improve access to resources, information, support, and opportunities may enhance satisfaction and reduce emotional exhaustion, but implementation should be accompanied by safeguards that monitor workload, role strain, and interpersonal climate to mitigate potential unintended consequences. Future research should (1) use longitudinal and intervention designs to test causality, (2) compare empowerment interventions across units and cultures, (3) examine mediators such as self-efficacy, psychological safety, and staffing adequacy, and (4) standardize and transparently report burnout instruments, scoring direction, and analytic decisions to improve comparability across studies.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

Yang Li and Xuyan Liu contributed to the conception and design of the study, acquisition of data, or analysis and interpretation of data; Ying Liu, Yuying Zhao and Zhao mei Meng contributed to the drafting the article or making critical revisions related to the relevant intellectual content of the manuscript; All authors validated and final approved of the version of the article to be published.

Funding

Not received.

Data availability

The datasets generated during and/or analyzed during the current study are available in the manuscript.

Declarations

Ethical approval

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.