Effectiveness of mindfulness-based interventions on burnout, resilience and sleep quality among nurses: a systematic review and meta-analysis of randomized controlled trials

The study addresses the high burden of psychological and physical challenges among nurses, including burnout, diminished resilience, and poor sleep quality arising from high workloads, emotional labor, and irregular schedules. Burnout is prevalent globally among nurses and is linked to anxiety, depression, and sleep disorders, negatively impacting job performance and quality of care. Resilience serves as a protective factor enabling adaptation to stress and is associated with improved well-being and professional quality of life. Sleep disorders are common in nurses, especially due to shift work, and contribute to chronic illness, fatigue, lower efficiency, and reduced care quality. While various interventions exist, mindfulness-based interventions (MBIs) such as MBSR and MBCT have shown promise in improving psychological outcomes. The research question is whether MBIs are effective in reducing burnout and improving resilience and sleep quality among nurses. The purpose is to synthesize RCT evidence to inform clinical practice and nursing management, given gaps in prior reviews that did not comprehensively assess resilience and sleep quality.

Prior systematic reviews reported benefits of MBIs for psychological distress, depression, stress, and burnout among nurses, but were limited by small samples, short follow-up, and lack of robust evidence. Some studies showed MBIs, including MBSR and MBCT, reduce anxiety, depression, and stress, with mixed findings on burnout dimensions (e.g., personal accomplishment). No previous meta-analysis comprehensively evaluated burnout, resilience, and sleep quality together in nurses, highlighting the need for this synthesis.

Design: Systematic review and meta-analysis conducted per PRISMA guidelines; protocol registered in PROSPERO (CRD42024610425). Data sources: PubMed, EMBASE, Scopus, Web of Science, PsycINFO, CINAHL, and CNKI searched from inception to November 2024 using MeSH and free-text terms related to nurses and mindfulness (e.g., MBSR, MBCT, mindfulness meditation). No restriction was applied on specific outcomes during search to maximize capture. Eligibility (PICOS): P Registered nurses with ≥1 year continuous work; I MBIs (e.g., MBSR, MBCT, ACT, mindfulness training/meditation/yoga, mindfulness eating/walking, loving-kindness meditation); C usual care/psychological education or no intervention; O at least one of burnout, resilience, or sleep quality; S RCTs in English or Chinese. Exclusions: Nursing students/other professionals; prior MBI exposure; non-mindfulness or combined therapies not isolating mindfulness; non-RCTs, protocols, theses, abstracts, reviews. Screening and data extraction: Duplicates removed in Zotero; two reviewers independently screened titles/abstracts and full texts, extracted study, participant, intervention, comparator, outcomes, and timing details into a standardized template with third-reviewer arbitration. Risk of bias: Assessed with Cochrane RoB 2 across randomization, deviations from intended interventions, missing data, outcome measurement, and selective reporting. Data synthesis: Continuous outcomes pooled using SMD or MD with 95% CI depending on measurement scales; heterogeneity assessed with I² (≥50% considered substantial). Random-effects models used when I²≥50%, fixed-effects otherwise. Sensitivity analyses assessed robustness. Subgroup analyses for burnout included country (developing vs developed), sample size (<100 vs ≥100), intervention type (MBSR, mindfulness therapy/interventions, mindfulness exercises), intervention length (>8 vs <8 weeks), and delivery mode (offline, online, hybrid). Meta-regression explored sources of heterogeneity where applicable. Publication bias was assessed via funnel plots and Egger’s test when ≥10 studies reported an outcome. Certainty of evidence: GRADE assessed risk of bias, inconsistency, indirectness, imprecision, and publication bias using GRADEpro.

- Studies included: 16 RCTs (2018–2024) from China, Jordan, Spain, USA, and Iran; total n=1384 (IG=715, CG=669); sample sizes 40–151. Interventions included MBSR, MBCT, ACT, mindfulness exercises (meditation, yoga, loving-kindness), delivered offline, online, or hybrid over 4–12 weeks. Outcomes measured with MBI/MBI-HSS/ProQoL (burnout), CD-RISC (resilience), PSQI/GSQ (sleep quality). - Burnout: 13 studies (n=1020) showed significant reduction with MBIs (SMD = -1.43, 95% CI -1.94 to -0.92; Z=5.52; P<0.001) with high heterogeneity (I²=92%). - Resilience: 6 studies (n=458) showed improvement (MD = 9.78, 95% CI 0.38 to 19.17; Z=2.04; P=0.04) with high heterogeneity (I²=98%). - Sleep quality: 3 studies (n=352) indicated better sleep with MBIs (SMD = -1.10, 95% CI -1.79 to -0.41; Z=3.14; P=0.002) with high heterogeneity (I²=89%). - Risk of bias: In RoB 2, 75% low risk for randomization; deviations from intended interventions often had some concerns due to blinding limits; two studies rated high risk overall; only two studies low risk overall; most had some concerns. - GRADE certainty: Burnout low; Resilience moderate; Sleep quality low. - Subgroup findings for burnout: • Country: Significant effects in both developing (SMD -1.78) and developed (SMD -0.64) settings; heterogeneity high. • Sample size: Significant for samples <100 (SMD -1.60); not significant for ≥100 (SMD -1.08; P=0.07). • Intervention type: MBSR (SMD -3.28), mindfulness therapy/interventions (SMD -0.72; I²=63%), mindfulness exercises (SMD -0.76; I²=0%); heterogeneity persisted overall. • Length: >8 weeks showed larger effect (SMD -1.91) than <8 weeks (SMD -0.71). • Mode: Hybrid online+offline (SMD -2.19) greater than online (SMD -0.78) and offline (SMD -1.72). - Resilience subgroup: MBSR (MD 16.71) and mindfulness exercises (MD 8.75) improved resilience; mindfulness therapy/interventions did not (MD 0.34). Hybrid delivery most effective (MD 16.71); offline also effective (MD 8.11). - Publication bias: Funnel plot asymmetry and Egger’s test (P<0.05) indicated risk of publication bias for burnout. Meta-regression did not identify sources of heterogeneity.

The findings support the research question that mindfulness-based interventions reduce burnout and improve resilience and sleep quality among nurses. Effects were consistent across various MBI formats, with MBSR and longer, hybrid-delivered programs tending to yield larger benefits. Potential mechanisms include neurophysiological changes (e.g., modulation of anterior cingulate and insula activity) and psychological shifts from a doing to a being mode that enhance present-moment awareness and emotional regulation, thereby lessening burnout, bolstering resilience, and reducing pre-sleep arousal. Despite robustness in sensitivity analyses, high heterogeneity, risks of bias (notably limited blinding and missing data), and publication bias warrant cautious interpretation. Subgroup patterns suggest program design factors (length >8 weeks, hybrid delivery, MBSR content) may enhance effectiveness and inform implementation in clinical settings to support nurse well-being and performance. However, the certainty of evidence is moderate-to-low, emphasizing the need for more rigorous, adequately powered, and transparently reported RCTs.

This systematic review and meta-analysis indicates that mindfulness-based interventions effectively reduce burnout and improve resilience and sleep quality in nurses, offering practical support for initiatives to enhance nurse mental health, quality of life, and nursing management. The main contributions include the first integrated synthesis of these three outcomes in nurse populations, identification of program characteristics associated with greater effects (e.g., MBSR, longer duration, hybrid delivery), and an evidence map of current RCT quality. Future research should prioritize high-quality, multicenter RCTs with adequate sample sizes, standardized outcomes and follow-up, rigorous allocation concealment and blinding of assessors, fidelity monitoring, and exploration of dose–response and implementation strategies across diverse cultural and work settings.

- High heterogeneity across studies that was not explained by subgroup or meta-regression analyses. - Risk of bias concerns, including limited blinding of participants/personnel, missing outcome data, and potential selective reporting; only a small proportion of studies were overall low risk. - Evidence of publication bias for burnout and a language restriction to English and Chinese, potentially excluding relevant trials. - Variability in intervention content, delivery modes, durations, and outcome measures, and limited follow-up in many studies, which may affect generalizability and precision of pooled estimates.