Comparing the effectiveness of mind–body practices (MBPs) and various psychological methods on occupational stress among healthcare workers: a network meta-analysis of randomized controlled trials

Abstract

Objectives

The purpose of this study is to compare the efficacy of mind–body practices (MBPs) and multiple psychological methods, and identify the optimal method for relieving work-related stress among healthcare workers (HCWs) by network meta-analysis (NMA).

Methods

We applied six electronic databases, namely PubMed, Web of Science, Embase, PsycINFO, Cochrane, and Chinese National Knowledge Infrastructure to identify relevant RCTs from inception to September 16, 2023, and implemented a search strategy based on the PICOS principles. Data selection, extraction, and analysis of bias were carried out independently and in duplicate by separate researchers. State 16.0 was used to conduct NMA for comparing the effectiveness of various therapies.

Results

We identified 23 studies including MBPs and three different psychological therapies, namely mindfulness-related therapy (MRT), psychoeducational therapy (PT), and comprehensive therapy (CT), which were divided into eleven specific techniques, namely yoga, meditation techniques (MT), Qigong, muscle relaxation(MR), biofeedback therapy (BT), mindfulness-based interventions (MBIs), modified mindfulness-based stress reduction (MBSR-M), mindfulness-based interventions combined with others (MBIs-C), mindfulness-based awareness(MBA), PT and CT. Our NMA results of MBPs and three psychological therapies showed MBPs (SMD = -0.90, CrI:-1.26, -0.05, SUCRA = 99%) were effective for occupational stress in HCWs, followed by MRT(SMD = -0.48, CrI:-0.87, -0.08, SUCRA = 66.5%). NMA results of eleven specific techniques showed yoga (SMD = 1.36, CrI:0.91, 1.81, SUCRA = 97.5%) was the most effective technique in relieving the stress of HCWs, followed by MR (SMD = 1.36, CrI:0.91, 1.81, SUCRA = 87.3%).

Conclusions

Our study suggested MBPs may be the most effective intervention to improve the occupational stress of HCWs. Furthermore, yoga is likely to be the most optimal of MBPs. Hospital managers should attach importance to yoga in addressing occupational stress among medical workers.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-024-11437-7.

Key Points

• Previous studies have identified that various psychological methods can improve the occupational stress of healthcare workers (HCWs). Mind-body practices (MBPs) integrating physical exercise with mind-based practice have been proven to be a clear strategy to relieve stress and facilitate adaptation to extreme environments for HCWs in recent years. However, few studies have compared the efficacy of MBPs with that of psychological methods. In addition, it is unclear which specific method is optimal for stress reduction among HCWs.

• Network meta-analysis (NMA) can provide complete insights into the clinical efficacy of different methods and is regarded as the highest level of evidence in the intervention guidelines. Thus, this study aims to compare the efficacy of MBPs and psychological interventions and identify the optimal method for relieving work-related stress in HCWs using NMA.

• Our NMA indicates MBPs are more effective than other psychological interventions for occupational stress among HCWs. The best intervention among MBPs is yoga, followed by MR. Therefore, hospital managers should attach importance to yoga in workplace wellness programs. Researches targeting suitable yoga training plan for medical workers need to be explored.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-024-11437-7.

Introduction

Occupational stress has been recognized in all workplaces and acknowledged as a global phenomenon existing in developed and developing countries [1]. It is mainly caused by the imbalance between the objective needs of leaders or the public and the workers’ adaptive ability in an occupational environment [2]. The work pattern and intensity in the present society will continue to aggravate the development of stress symptoms and may even result in the occurrence of stress-related diseases. Occupational stress has a close relationship with mental disorders (i.e. depression [3], job burnout [4], and insomnia [5]) and physical diseases (i.e. cardiovascular disease [6], diabetes [7], and irritable bowel syndrome [8]). At the same time, economic losses caused by occupational stress are increasing annually [9].

Compared with the level of work-related stress among different fields, healthcare workers (HCWs) are widely regarded as one of the most high-risk professions [10]. Some studies proved that medical practitioners have experienced moderate to severe occupational stress [11–13]. A systematic review and meta-analysis assessed the prevalence of work-related stress among HCWs, which ranged from 50.8% to 54.2% [1]. The National Institute for Occupational Safety and Health (NIOSH) reported longer hours, night shift work, and hazards in the occupational environment as contributing to stress in hospitals [14], which may lead to lower job satisfaction, poor performance, and even medical accidents [15]. Thus, it is crucial to take effective and adequate measures to improve the work-related stress of HCWs.

The COVID-19 epidemic in 2019 made the public realize that healthcare workers were widely under immense physical and emotional strain, thus psychological interventions have been recommended to HCWs for stress reduction [16–18]. One popular approach is mindfulness training [19, 20], which involves a series of designed practices and aims to alleviate mental symptoms and facilitate emotional well-being. It transfers attention intentionally from the awareness of involuntary inner activities to current experiences and maintains a curious, open and accepting attitude, which helps alleviate negative emotion [21].

Three reviews of a population of HCWs showed that mindfulness-based programs are efficient interventions, which can reduce the mood disturbance and stress symptoms of HCWs [22–24]. Mindfulness training has been widely used to enhance psychological distress among different populations, such as patients with cancer, hypertensive persons, college students, and teachers [25, 26], especially for HCWs [23, 24]. Cognitive behavioral therapy (CBT) is another popular intervention that contributes to lowering the level of psychological disorders by helping individuals gain correct self-perception and self-confidence to deal with stress [27]. Previous studies have also shown that psychoeducational therapy can relieve stress and burnout among HCWs [28, 29].

Mind–body practices (MBPs), such as yoga, Qigong, Tai Chi, meditation technique (MT), muscle relaxation (MR), and biofeedback techniques (BT), combine physical movements, deep breathing, and mind-based practice in a specific pattern [30, 31]; this has been proven to be a clear strategy to relieve stress and facilitate adaptation to extreme environments among HCWs in recent years [32, 33]. A published meta-analysis reported the effectiveness of yoga, Qigong, and MR in stress reduction for doctors and nurses [34]. Among MBPs, yoga is an effective and low-cost practice that can improve the stress and well-being of HCWs [35, 36]. Clinical trials have reported the efficacy of MT for stress reduction among teachers and clinicians [37, 38]. MR, one of the MBPs, is opposite to tension and involves voluntary stretching and relaxation of muscle groups in the human body [39]. One study investigated the effectiveness of MR in improving the stress of nurses caring for patients with COVID-19 [40]. In addition, MBPs are convenient for HCWs and can be accessed online for practice.

Previous studies have identified that various psychological interventions and MBPs could improve occupational stress and enhance the well-being of HCWs [22]. However, few studies have compared the efficacy of MBPs with that of various psychological therapies. In addition, it is unclear which specific intervention is optimal for stress reduction among HCWs. Thus, the above limitations could be solved by the method of network meta-analysis (NMA) [41], which can compare the clinical efficacy of multiple methods and is regarded as the highest level of evidence in the intervention guidelines [42]. NAM is a novel analysis technique different from conventional meta-analysis, which builds a network graph that permits direct and indirect (no pairwise head-to-head) comparisons between different studies. Furthermore, NMA can be used to rank different categories of treatments based on different outcomes, which provides evidence-based data for policy-making [43].

To date, there is no study exploring the best psychological method to reduce the occupational stress of healthcare workers. Therefore, we performed an NMA to compare the efficacy of various psychological methods and identify the optimal approach for relieving work-related stress in HCWs.

Materials and methods

The study was reported following the Preferred Reporting Items for Systematic Reviews (PRISMA) and network meta-analysis guidelines. This study has been registered with PROSPERO (ID: CRD42023462614). The method of body stretch registered in PROSPERO was replaced with muscle relaxation in the study.

Search strategy and study selection

Six electronic databases, namely PubMed, Web of Science, Embase, PsycINFO, Cochrane, and Chinese National Knowledge Infrastructure (CNKI), were searched from inception to 16 September 2023. We combined relevant medical subject heading (MeSH) terms with free text words including healthcare workers, psychological interventions, mindfulness, and mind–body practices, with the Boolean search terms ‘OR’ and ‘AND’. The search strategy was reported in Supplementary 1. After consulting a medical librarian about the search strategy, we conducted the literature search under the supervision of one researcher (CW).

To prevent missing relevant literature, we conducted a manual search of related journals, conferences and references to obtain eligible studies. After the studies were obtained, two independent researchers (YJZ and JCS) assessed the titles and abstracts in duplicate, deleted irrelevant and repeated articles, and then read all articles to check for potentially eligible studies following the previous inclusion and exclusion criteria. To guarantee the agreement of the two researchers, we conducted a pilot process by five RCTs in examples before initiating the selection of included articles. Finally, the included articles were summarised and categorized into pre-designed Excel sheets. Controversial trials were resolved through consultation with a third researcher (YF).

Inclusion and exclusion criteria and data extraction

To screen eligible trials, we implemented search strategies based on PICOS principles.

Participants

The participants in these trials were healthcare workers, including doctors, nurses and other professionals engaged in medical work (e.g., psychologists, dentists, speech therapists, paramedics, and occupational therapists). Studies involving medical college students were excluded.

Interventions

Psychological interventions included mindfulness-related training, such as mindfulness-based interventions (MBIs), mindfulness-based stress reduction (MBSR), modified MBSR, mindfulness-based awareness, mindfulness combined with other strategies, and CBT, psychoeducation therapy (PT) and other psychological methods. MBPs included yoga, meditation, muscle relaxation, biofeedback training, qigong and tai chi. Some studies combined two or three psychological methods; these were classified as comprehensive therapy (CT).

Comparisons

We compared all types of interventions with the control group or themselves. In general, the control group received the usual care, such as rest, entertainment activities, no intervention and regular health education.

Outcomes

Studies that used the Perceived Stress Scale (PSS) as a primary or secondary outcome and that contained extractable data to evaluate stress were included. Other scales assessing stress levels included the PSS-10, PSS-14, Depression, Anxiety and Stress Scale (DASS-21), Brief Job Stress Questionnaire (BJSQ), Occupational Stress Indicator (OSI) and Perceived Stress Questionnaire (PSQ).

Studies

The trials included in this study were RCTs and full-text articles. No limitations were imposed on the publication time, but the languages of the trials were English and Chinese. Protocols without relevant data or results were excluded. we conducted this by manual screening during the process of extracting the basic characteristics of each included trial.

Outcome measurement and quality appraisal

The data contained general information on the literature (e.g. first author, publication time, country), the characteristics of participants (e.g. age, gender, occupation), further details of the intervention and control groups (e.g. intervention methods, sample numbers, way, frequency and period) and the measurement scale of outcome, which were extracted using an Excel sheet by two researchers (YJZ and CW). There was a pilot process before initiating the quality assessment.

The outcome of the extracted studies was the efficacy of various methods on occupational stress among HCWs compared to each other or a control group. Each trial yielded different results, which were measured using different scales. The investigators extracted the value of the mean (M) and standard deviation (SD) of each scale at baseline before intervention and after the last intervention to analyse the change in value after treatment.

Two researchers independently (YJZ and CW) evaluated the quality of the included randomized controlled trial (RCT) using the revised Cochrane Risk of Bias (ROB: version 2) from seven domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other biases. All studies were classified as high, low, or unclear in each of the seven domains, and the assessment outcome of ROB was presented using the Review Manager. Any discrepancies during data extraction were resolved by discussion, and consent was obtained.

Statistical analysis

The data reported in this study were continuous variables, and the mean square deviation (MD) and 95% confidence interval (CI) were used as effect measures. The methodological quality of the literature was assessed using Review Manager 5.3 and a bias risk assessment map for all trials was charted.

STATA16.0 was used for network meta-analysis. We used network diagrams to analyse the intervention network's geometry and potential biases. In network diagrams, each blue node indicated one intervention, and each edge indicated direct comparisons between different nodes. The bigger the blue node, the larger the size of the participants, and the thicker of the edge, the more comparisons between different interventions.

A node-cutting method was adopted to determine the consistency between direct and indirect comparisons of the various interventions when there was a closed-loop structure. The inconsistency factors (IF) and 95% CI and significant P values were calculated [44]. If P > 0.05, IF value < 1, and 95% CI’ s contained 0. The consistency of each closed loop was good. otherwise, inconsistency is considered to exist, and the inconsistency model was used.

As an index for predicting probability by ranking the targeted intervention, the surface under the cumulative ranking curve (SUCRA) was represented as a simple numerical summary statistics cumulative ranking probability map for each approach, with SUCRA values ranging from 0 to 100%; the closer of the SUCRA value to 100%, the better the intervention performance. We also used Stata 16.0 to draw funnel plots to test for publication bias by Begg’s and Egger’s tests.

Results

Study characteristics

We identified 2,285 records using the described search strategy and six records using manual research. After removing duplicate and irrelevant records, we obtained 1660 records, of which 1588 were filtered based on titles and abstracts. Of the 72 studies under full-text evaluation, 8 studies were irrelevant interventions, 12 trials were not medical workers; 26 studies did not have the relevant outcome; two records had incomplete data, and one record was not an RCT. Finally, 23 trials that met our inclusion criteria were included in this network meta-analysis [45–67]. Figure 1 showed the flowchart for screening the RCTs and the final result.

Fig. 1.

Literature review flowchart. MRT: Mindfulness Related Therapy; MBPs: Mind–body practices; PT: Psychoeducational Therapy; CT: Comprehensive Therapy; MBSR-M: Mindfulness based stress reduction-modified; MBIs: Mindfulness based interventions; MBIs-C: Mindfulness based interventions combined other strategies; MBA: Mindfulness based awareness; MT: Meditation Techniques; BT: Biofeedback Techniques; MR: Muscle Relaxation

The 23 studies included 4,363 participants using MBPs and three psychological interventions (MRT, PT and CT), which were divided into 11 specific methods (MBIs, MBSR-M, MBIs-C, MBA, yoga, MT, qigong, MR, BT, PT and CT). The included studies were published between 2011 and 2023. Table 1 summarized the key characteristics of participants and interventions of the 23 RCTs. The studies were from multiple regions: the United States had the most, with five records, followed by Germany and Turkey with three studies, respectively; two studies were completed in Japan, Taiwan and China, respectively; one trial was in Australia, Canada, Spain and India, respectively. Regarding the participants, nine studies focused on nurses, and six studies focused on doctors. The remaining eight studies did not report the specific profession in detail. For the intervention approach, 16 studies were conducted in person, four in digital and three in mixed methods. The age range of all participants was between 18 and 60 years.

Table 1.

Characteristics of included studies

Numbers	Publication	Recruiting area	Participants	Treatment and sample size	Age	Outcomes	Ways	Outcome Time	Frequency
1 [45]	Lin, 2018	China	Nurses	MBSR-M = 55 PT = 55	30.86 ± 7.49/ 30.20 ± 6.09	PSS-14	Mixed	8-week	2-h weekly group sessions and 20 min daily for 6 days a week
2 [46]	Fendel, 2021	Germany	Resident physicians	MBSR-M = 76 UC = 71	31.04 (3.39)/ 31.00(3.49)	PSS-10	in person	8-week	135 min weekly group sessions and 6-h silent retreat daily
3 [47]	Ireland, 2017	Australia	Medical practitioners	MBIs = 23 UC = 21	26.88 ± 4.79	PSS-10	in person	10-week	1-h training weekly
4 [48]	Li, 2022	China	Healthcare professionals	MBA = 87 UC = 47	23.0–49.0/ 24.0–48.0	PSS-10	in person	16-day	15 min and once daily
5 [49]	Purdie, 2023	USA	Pediatric residents	MBA = 27 UC = 39	25–34	PSS-14	Mixed	6-week	2-h weekly, 6-group sessions curriculum
6 [50]	Strauss, 2021	UK	Healthcare worker	MBIs-C = 115 UC = 119	42.95 ± 10.05/ 44.92 ± 10.68	DASS-21	in person	8-week	2-h weekly session and 40 min daily at home
7 [51]	Taylor, 2022	UK	Healthcare workers	MBIs-C = 1095 UC = 1087	40.42 ± 10.92/ 40.64 ± 11.02	DASS-21	digital	4.5-month	10-min practice daily for 10 days
8 [52]	Mistretta, 2018	USA	Health care workers	MBIs-C = 22 UC = 15	48.2 ± 11.6/ 46.1 ± 10.5	DASS-21	in person	6-week	Six weekly 2-h sessions daily
9 [53]	Miyoshi, 2019	Japan	Night shift nurses	Yoga = 20 UC = 20	28.7 ± 4.9	BJSQ	digital	4-week	3 times weekly for approximately 5 to 15 min each time
10 [54]	Mandal, 2021	India	Nurses	Yoga = 58 UC = 55	35.0 ± 7.9/ 32.5 ± 6.8	PSS-10	in person	12- week	50 min once and twice weekly
11 [55]	Celi, 2022	Turkey	Nurse	Yoga = 51 UC = 50	28.86 ± 6.79 / 28.86 ± 6.06	PSS-14	digital	4 weeks	40 min once and twice weekly
12 [56]	Lin, 2015	Taiwan	Healthcare professionals	Yoga = 30 UC = 30	32.07 ± 7.54/ 29.77 ± 6.89	Work-related stress	in person	12-week	1-h weekly
13 [57]	Joshi, 2022	USA	Healthcare workers	MT = 41 UC = 39	39 ± 12/41 ± 9	BSI-18	in person	3-month	75 min sessions and self-practice of 20 min twice daily
14 [58]	Griffith, 2008	USA	Hospital staff	Qigong = 16 UC = 21	52 ± 9/50 ± 10	PSS-10	Mixed	6-week	1.5-h once and twice weekly
15 [59]	Akyurek, 2020	Turkey	Nurses	MR = 15 UC = 15	33.53 ± 13.67 / 41.47 ± 9.39	VAS	in person	5-week	35 min exercise and twice weekly
16 [60]	Ozgundondu, 2019	Turkey	Nurses	MR = 28 UC = 28	24.61 ± 2.61/ 27.75 ± 4.75	PSS-10	in person	8-week	20 min training and once weekly
17 [61]	Hsieh, 2020	Taiwan	Psychiatric nurses	BT = 49 UC = 39	38.45 ± 9.23/ 35.61 ± 7.47	OSI	in person	6-week	1-h weekly
18 [62]	Lemaire, 2011	Canada	Physician	BT = 21 UC = 19	47.8 ± 8.5/ 44.8 ± 8.2	POQA-R	in person	28-day	5 min and 3 times a day
19 [63]	Kubota, 2016	Japan	Oncology nurses	PT = 50 UC = 46	38.9 ± 5.4/ 40.0 ± 5.4	NJSS	in person	2-week	8-h program and once weekly
20 [64]	Sawye, 2023	USA	Nurse leaders	PT = 45 UC = 32	43.97 ± 10.6/ 44.76 ± 9.24	PSS-10	in person	1-month	Nine weekly 90 min group sessions
21 [65]	Fiol-DeRoque, 2021	Spain	Healthcare workers	CT = 221 UC = 215	42.07 ± 11.0/ 40.62 ± 9.6	DASS-21	digital	2-week	France language
22 [66]	Mache, 2018	Germany	Junior physicians	CT = 35 UC = 35	27.3 ± 2.5/ 27.1 ± 2.1	PSQ-20	in person	12-week	1.5 h training weekly
23 [67]	Mache, 2015	Germany	Junior Physicians	CT = 42 UC = 43	/	PSQ-30	in person	12-week	2-h weekly

BJSQ The Brief Job Stress Questionnaire, BSI-18 18-item Brief Symptom Inventory, VAS Visual Analog Scale, OSI Occupational Stress Indicator, POQA-R Personal and Organizational Quality Assessment–Revised, NJSS Nursing Job Stressor Scale, DASS-21 Depression, Anxiety, and Stress Scale, PSQ Perceived Stress Questionnaire, PSS Perceived stress scale, MBPs Mind–body practices, MRT Mindfulness related therapy, PT Psychoeducational therapy, CT Comprehensive therapy, MBIs Mindfulness-based interventions, MBSR-M Modified mindfulness-based stress reduction, MBIs-C Mindfulness-based interventions combined with other strategies, MBA Mindfulness based awareness, MT Meditation techniques, MR Muscle relaxation, BT Biofeedback therapy

Risk of bias assessment

Among the 23 studies, 19 described the allocation methods; 20 reported allocation concealment, and eight mentioned blinding of participants and researchers and outcome assessment. In three trials, the risk of bias in the blinding of participants was high. All records contained complete outcome data. Figure 2 showed the risk of bias assessment.

Fig. 2.

Risk of bias summary and risk of bias graph

Inconsistency check

To analyse the consistency between the direct and indirect evidence of the four different therapies and 11 specific techniques, we used two node-splitting analysis models. The results showed that both therapies (P = 0.811) and techniques (P = 0.680) had no differences in consistency.

Evidence network

Figures 3 and 4 showed the evidence network plot of the four therapies and 11 techniques, respectively. In the network geometry, each node represented a method, the size of which depended on the number of participants included in this method. The lines joining the nodes represented the direct relationships between the therapies. In addition, the thickness of the lines was weighted based on the direct evidence available between them. Figure 3 showed that MBPs had the largest sample size in the intervention group, followed by MRT and CT. The PT group comprised a small number of participants. One closed loop was observed between MRT and PT, representing direct evidence. Figure 4 showed that the sample size of MBIS-C was the highest in the intervention group, followed by CT, PT and yoga. Qigong was the smallest. One closed loop was also observed between MBSR-M and PT.

Fig. 3.

Network of evidence of four therapies. MBPs: Mind–body practices; MRT: Mindfulness related therapy; PT: Psychoeducation therapy; CT: Comprehensive therapy; UC: Usual Care

Fig. 4.

Network of evidence of eleven techniques. PT: Psychoeducational therapy; CT: Comprehensive therapy; MBIs: Mindfulness-based interventions; MBSR-M: Modified mindfulness-based stress reduction; MBIs-C: Mindfulness-based interventions combined with other strategies; MBA: Mindfulness based awareness; MT: Meditation techniques; MR: Muscle relaxation; BT: Biofeedback therapy; UC: Usual Care

Network meta-analysis

We performed a network meta-analysis of the included studies in terms of reducing occupational stress between the four therapies. The results showed that the efficacy of MBPs (SMD = -0.90, CrI = -1.26, -0.05) and MRT (SMD = -0.48, CrI = -0.87, -0.08) was better than that of the control group. However, CT (SMD = -0.47, CrI = -1.11, 0.16) and PT (SMD = -0.31, CrI = -1.09, 0.48) reported contrary results (Fig. 5). A SUCRA line was drawn to rank the availability of each treatment, which showed that MBP had the highest probability in terms of stress reduction compared with the other three psychological methods. The ranking was MBPs (SUCRA = 99.0%) > MRT (SUCRA = 66.5%) > CT (SUCRA = 33.8%) > PT (SUCRA = 33.4%) > UC (SUCRA = 17.4%) (Fig. 6).

Fig. 5.

Relative effect sizes of efficacy at post-treatment. Significant pairwise comparisons are highlighted in orange. MBPs: Mind–body practices; MRT: Mindfulness related therapy; PT: Psychoeducation therapy; CT: Comprehensive therapy; UC: Usual Care

Fig. 6.

Overall ranking based on SUCRA line. MBPs: Mind–body practices; MRT: Mindfulness related therapy; PT: Psychoeducation therapy; CT: Comprehensive therapy; UC: Usual Care

To identify the specific techniqueto relieve the stress of the medical staff, we divided the MBPs and other psychological interventions into 11 techniques and conducted another NMA. The results demonstrated that yoga (SMD = 1.36, CrI = 0.91, 1.81) and MR (SMD = 1.36, CrI = 0.91, 1.81) were more effective than that of the control group. However, Qigong (SMD = -0.84, CrI = -2.00, 0.32), MBSR-M (SMD = -0.64, CrI = -1.43, 0.14), MBIs (SMD = -0.54, CrI = -1.66, 058), CT (SMD = -0.47, CrI = -1.06, 0.12), MT (SMD = -0.44, CrI = 0–1.49, 0.60), MBIs-C (SMD = -0.43, CrI = -1.01, 0.16), PT (SMD = -0.37, CrI = -1.03, 028), BT (SMD = -0.3, CrI = -1.06, 0.47) and MBA (SMD = -0.05, CrI = -0.79, 0.69) had no statistical difference compared with the control group (Fig. 7). A SUCRA line was drawn to rank the efficacy of 11 techniques, which reported that yoga had the best probability compared to the other 10 approaches. The ranking was yoga (SUCRA = 95.7%) > MR (SUCRA = 87.3%) > qigong (SUCRA = 65.6%) > MBSR-M (SUCRA = 57.5%) > MBIs (SUCRA = 50.0%), CT (SUCRA = 47.1%) > MT (SUCRA = 46.0%) > MBIs-C (SUCRA = 45.0%) > PT (SUCRA = 40.8%) > BT (SUCRA = 35.8%) > MBA (SUCRA = 20.8%) > Control (SUCRA = 13.1%) (Fig. 8).

Fig. 7.

Overall ranking based on SUCRA line. PT: Psychoeducational therapy; CT: Comprehensive therapy; MBIs: Mindfulness-based interventions; MBSR-M: Modified mindfulness-based stress reduction; MBIs-C: Mindfulness-based interventions combined with other strategies; MBA: Mindfulness based awareness; MT: Meditation techniques; MR: Muscle relaxation; BT: Biofeedback therapy; UC: Usual Care

Fig. 8.

Relative effect sizes of efficacy at post-treatment. Significant pairwise comparisons are highlighted in orange. PT: Psychoeducational therapy; CT: Comprehensive therapy; MBIs: Mindfulness-based interventions; MBSR-M: Modified mindfulness-based stress reduction; MBIs-C: Mindfulness-based interventions combined with other strategies; MBA: Mindfulness based awareness; MT: Meditation techniques; MR: Muscle relaxation; BT: Biofeedback therapy; UC: Usual Care

Publication bias assessment

We performed a biased assessment of four interventions and 11 specific techniques in terms of reducing occupational stress among HCWs and drew a comparison-correction funnel chart (Figs. 9 and 10). The results showed that the symmetry was acceptable; however, publication bias affected this study.

Fig. 9.

A UC (Usual care); B MBPs (Mind–body practices) C CT (Comprehensive therapy); D MRT (Mindfulness related therapy); E PE (Psychoeducation therapy)

Fig. 10.

A UC (Usual care); B BT (Biofeedback therapy) C CT (Comprehensive therapy) D MBSR-M (Modified mindfulness-based stress reduction) E MBIs(Mindfulness-based interventions) F MBA(Mindfulness based awareness); G MBIs-C(Mindfulness-based interventions combined with other strategies); H PT(Psychoeducational therapy) I Qigong J MR(Muscle relaxation) K MT(Meditation techniques); L yoga

Discussion

Although numerous stress-reduction techniques for HCWs have been discussed in previous studies, they lack network integration of the effects of multiple methods. The integration of direct and indirect comparisons in our network meta-analysis gave rise to great statistical precision in the scientific comparisons of various treatments for occupational stress. According to our study, MBPs had the highest probability of being the best intervention for stress reduction among the four therapies, followed by MRTs. Furthermore, we divided the four kinds of methods into 11 specific techniques and found that yoga showed a great advantage compared to other techniques. MR was suboptimal and could be used as an important complementary and assistive method for stress induction.

By investing in emotions and focusing on physical movements, MBPs guide the body to produce a relaxed and self-healing feedback response, thereby stimulating a virtuous cycle of psychological influence and physiological assistance and achieving the maximum value of positive physical and mental development [68]. Several systematic reviews covering this topic have proven that MBPs were beneficial in improving individuals’ physical function and well-being [68–70]. In our study, MBPs were more effective than other psychological treatments in reducing the stress of HCWs, which was according to a pilot study [71]. When individuals are exposed to a stressful environment, their sympathetic nervous system is triggered, and many inflammatory neurotransmitters, such as cytokines, are released into the human body, leading to negative emotions produced by humans. MBPs have a regulatory effect on the sympathetic nervous system and the hypothalamus–pituitary–adrenal axis [72, 73] and can reduce the reaction of inflammatory neurotransmitters in individuals during stress [74]. This evidence implies that MBPs may be a promising approach for the intervention of occupational stress.

Our NMA revealed that MRT was another best psychological intervention to relieve stress and improve the well-being of HCWs, which had been verified in previous reviews [75, 76]. Three reviews provided evidence to support the effectiveness of MBSR in reducing the occupational stress of medical workers [22, 77, 78]. A meta-analysis by Burton et al. proved that MBSR, MBSR-M and MBIs had a significant effect on the psychological stress experienced by HCWs [78]. In addition to MBSR, Van der Riet et al. proved that mindfulness-based awareness (MBA) had a positive impact on work-related stress in nurses [79]. MRT in healthcare environments could reduce providers’ perceived stress by stimulating self-care, establishing resilience and improving stress management. Medical staff who were unsuitable to MBPs due to physical sickness, pain or discomfort can choose MRT to regulate mental disorders.

Our analysis of 11 techniques showed that yoga was significant in relieving occupational stress among medical staff. This finding was in line with a meta-analysis that reported the efficacy of yoga in improving stress and promoting positive psychological adjustment compared to other interventions [34]. Referred to as a mind–body practice, yoga integrates physical exercise with mind components to achieve mental health [80], which can be explained from two aspects. First, yoga is a typical aerobic exercise that effectively reduces weight, blood lipids and body fat. Studies have confirmed that people who engaged in long-term yoga training showed a significant decrease in BMI compared to those who did not [81, 82]. In particular, yoga exercise has a significant regulatory effect on abdominal and buttock fat accumulation in healthy individuals [83]. Research has suggested that weight loss not only shapes the perfect physique of the human body but also plays an important role in overcoming social and physical anxiety, improving physical self-esteem and enhancing resilience to stress [84]. However, modulation of the autonomic nervous system appears to play a critical role. Second, yoga breathing is an effective way to regulate the imbalanced autonomic nervous system, and conscious breathing control may positively affect autonomic nervous functions such as heart rate variability (HRV) and cardiac vagal tension, thereby affecting emotions, stress and cognition and improving psychological disorders [85]. Studies have also documented a reduction in heart rate, HRV, and breath rate after yoga practice. Sarika et al. [86] proposed that cardiac vagal nerve tension served as a marker for regulating emotions and psychological adaptability. In addition, our study found that MR was another effective approach, which was inconsistent with a randomized clinical trial [87]. MR refers to building a condition of moderation contrary to tension [88]. Jacobson reported that MR affected the body by increasing respiratory ventilation, blood flow volume and epinephrine secretion levels, and it assisted individuals in improving their physical and mental health [89].

Given the evidence from our study and findings from previous research, health guidance provided by hospital managers on MBPs for healthcare workers in daily clinical practices would be more conducive to relieving occupational stress and promoting well-being. Furthermore, yoga is simple to operate and flexible in time and space, which makes it more suitable for application in medical staff. Another study agreed with our results by suggesting that acceptance has grown for recommending the use of yoga alongside psychological treatment in clinical practice [90].

The most substantial advantage of this study was to integrate previous studies that only discussed mind–body practice or psychological therapies for medical workers with the current study, providing evidence for managing occupational stress among HCWs. Moreover, mind–body practices and psychological interventions are multiple and complex, and few NMAs have been conducted on work-related stress; this was the unique aspect of this study, which presented more scientific evidence for stress reduction.

Our study had some limitations. First, although a systematic and comprehensive search strategy was carried out, it is evident that articles that covered this topic were not included in this study for using different keywords or other languages. Second, the literature on some interventions included in our study was limited, and only one record on Qigong, MBIs and MR was available. Third, the overall quality of the trials included in this NMA was medium, mainly due to a lack of blinding. Finally, the outcome of occupational stress among healthcare workers also had objective indexes, such as salivary amylase, cortisol and physiological index. we only involved psychological indicators.

Conclusion and recommendations

The occupational stress among HCWs is one of the obstacles in providing high-quality healthcare services and promoting health for the public. Occupational stress increases the workforce's mental health issues, as well as threatening the safety of patients. Our NMA indicated that MBPs were more effective than other psychological therapies for occupational stress among HCWs. The perfect intervention was yoga, followed by MR. The findings can be helpful to policymakers or hospital managers in resolving the work-related stress suffered by medical workers. Further researches are expected to explore the in‐depth effects of occupational stress on patients and personalized yoga practice to address occupational stress more effectively in HCWs.

Abbreviations

NMA

Network meta-analysis

HCWs

Healthcare workers

MBPs

Mind–body practices

MRT

Mindfulness related therapy

PT

Psychoeducational therapy

CT

Comprehensive therapy

MBIs

Mindfulness-based interventions

MBSR

Mindfulness-based stress reduction

MBIs-C

Mindfulness-based interventions combined with other strategies

MBA

Mindfulness based awareness

MT

Meditation techniques

MR

Muscle relaxation

BT

Biofeedback therapy

CBT

Cognitive-behavioral therapy

CNKI

Chinese National Knowledge Infrastructure

PRISMA

Preferred Reporting Items for Systematic Reviews

Authors’ contributions

Study design: YJZ, WDH, HJL and JCS; Data collection: YJZ and CW; Data analysis: YF and CW; Manuscript writing: YJZ, CW and JCS.

Funding

We acknowledge the financial support from National Science Foundation of China (72101262), Degree and Graduate Education Project of Shaanxi Province (SXGERC2023116), the Science and Technology of Shaanxi Provincial Department Projects Fund(2023-YBSF-616), Education Department of Shaanxi Province Project Fund(22JK0343) and the Development Mechanism and Adjustment Strategy of Nursing Staff Burnout in the post-epidemic Era (2023KXKT018).

Availability of data and materials

Data supporting our findings are contained within the manuscript.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.