Backgrounds
In recent years, the COVID-19 pandemic has posed new challenges to nursing work, with both healthcare organizations and individual nurses facing uncertainty and complexity. Prior to the pandemic, nurse burnout was already a global challenge, and the additional stress associated with the COVID-19 pandemic may lead to workforce instability in the coming years [1]. Recent meta-analyses have indicated that scores for various dimensions of nurse burnout are higher than pre-pandemic levels and higher than those of other healthcare workers [2, 3]. Furthermore, studies have also indicated an increase in nurses’ turnover intention [4]. Therefore, it is evident that nurse burnout and turnover intention are crucial factors influencing the global shortage of nurses, needing immediate attention and resolution.
Burnout refers to a syndrome that individuals develop under prolonged and intense work pressure, comprising three dimensions: emotional exhaustion, depersonalization, and reduced personal accomplishment [5]. In May 2019, the World Health Organization included “burnout” in the International Classification of Diseases(ICD-10), highlighting its universality and severity [6]. Existing research indicates that occupational burnout can have negative effects on individuals’ physiological and psychological health. Physiologically, occupational burnout is associated with health issues such as structural and functional brain changes, cardiovascular diseases, musculoskeletal disorders, respiratory system diseases, headaches, infections, and others [7, 8]. Psychologically, it can lead to impaired cognitive functions such as working memory, executive function, and attention, as well as emotional changes such as anxiety and depression [9, 10]. Organizationally, occupational burnout may result in low job satisfaction, increased turnover intention, and higher employee costs [11], making it a significant factor contributing to nurse shortages.
Anzeige
The high turnover rate among nurses constitutes a significant factor contributing to the shortage of nurses. Turnover intention serves as an early indicator of actual turnover [12]. Therefore, one of the crucial approaches to addressing the nurse shortage is to reduce their turnover intention [13]。Turnover intention refers to the willingness to leave the job within a certain period [13]。Past research has identified nurse burnout as one of the strongest predictors of turnover intention [14]. Burnout negatively influences nurses’ daily behavioral adjustments, as employees experiencing burnout tend to avoid it instinctively and are inclined towards resignation [12]. Although the relationship between nurse burnout and turnover intention has been extensively studied, there are still several limitations. Firstly, most studies have focused on qualitative analysis or quantitative analysis based on traditional statistical methods, lacking a thorough understanding of this relationship. Secondly, existing research often overlooks the complex interactions among variables and network structures. Finally, exploration of the potential nonlinear relationships and interactions between nurse burnout and turnover intention remains limited. Therefore, it is imperative to employ more advanced and comprehensive methods to delve deeper into the intrinsic connection between nurse burnout and turnover intention.
In recent years, observation-based network analysis has been widely applied in psychopathology, assuming that mental disorders arise from direct interactions and causal relationships among symptoms [15, 16]. Compared to simple correlation methods, network analysis offers corresponding centrality and predictability indices for each node, assessing their importance and controllability within the entire network [15]. This holds significant potential in the field of nursing practice, where core nodes identified through network analysis could serve as potential targets for nursing interventions, aiding in the precise conduct of intervention studies. Additionally, network analysis can unveil how relationships between variables manifest. For instance, a recent network analysis study revealed the relationship between burnout, flourishing, and job satisfaction among HIV/AIDS healthcare workers in Western China [17]. The results found that “feeling frustrated at work” and “interested in daily activities” in occupational burnout were both central variables and bridges. These two factors may serve as intervention targets to alleviate the overall symptom level in the network and therefore prevent adverse health outcomes among healthcare workers. Therefore, from a methodological perspective, network analysis provides valuable insights for delving deeper into the exploration of nurses’ burnout and turnover intention.
Although numerous studies have indicated a close relationship between burnout and turnover intention, employing network analysis to investigate this relationship is crucial. Network analysis enables us to identify central variables and bridges within the networks of burnout and turnover intention. These nodes may serve as crucial predictors or influencing factors of burnout and turnover intention, making them pivotal intervention targets. By targeting these core nodes for intervention, managers can devise more effective human resource management strategies, optimize workflow processes, and specifically reduce nurses’ levels of burnout, thereby decreasing their turnover intention and improving the quality of nursing care. Therefore, this study aims to explore the relationship between burnout and turnover intention among nurses using network analysis, offering potential solutions to address the issue of nurses leaving their jobs due to burnout.
Study design, setting, and participants
This study employed a cross-sectional research design and was conducted at three comprehensive tertiary hospitals in Hangzhou, China, which are similar in size, management style, and medical standards. The sample size was calculated using G*Power software (version 3.1). Based on a correlation analysis, referencing past studies with a correlation coefficient of 0.26 [18], α = 0.05, and a power (1-β err prob) of 0.95, the minimum required sample size was determined to be 155 nurses. Accounting for a potential 10% rate of non-responsive or invalid questionnaires, the sample size was adjusted to at least 170 nurses.
Anzeige
Inclusion and exclusion criteria
Inclusion criteria were as follows: (1) registered nurses, (2) with at least 1 year of work experience, and (3) providing informed consent. Exclusion criteria were: (1) nursing interns/residents, (2) a history of mental or psychological disorders, and (3) on leave during the survey period.
Data collection
From September to November 2023, the present study used an online questionnaire survey for data collection. Researchers proactively contacted nursing departments in major hospitals, distributing the online questionnaire links to data collectors, and conducted training sessions on ethics and questionnaire completion guidelines. Following the training, data collectors distributed the links to nurses, who autonomously responded to the questionnaire. Each questionnaire began with a section that outlined the purpose of the study, the procedures involved, potential risks, and benefits. This introduction ensured that all participants provided informed consent. It also highlighted that the data would be anonymized and participants could withdraw from the study at any time without any penalties; if nurses chose not to participate, no data would be collected and the response process would be terminated. To enhance response accuracy, we restricted each IP address to one response. Following data collection, researchers downloaded the data for analysis. Initially, 584 questionnaires were collected. After removing 24 that did not meet the inclusion criteria, a total of 560 valid questionnaires remained, yielding an effective response rate of 95.89%.
Measures
Social demographic information
Social demographic information was collected through a self-administered questionnaire. This included gender, age, educational background, years of working, professional title, with or without children, marital status, job category, and monthly night shift frequency.
Burnout
In this study, the Chinese version of the Maslach Burnout Inventory-General Survey (MBI-GS) was utilized to measure burnout among nurses. This scale, developed by Maslach and adapted by Chinese scholar Li Chaoping [19, 20], has been widely used in China [21]. The scale comprises three dimensions: emotional exhaustion, depersonalization, and diminished personal accomplishment, consisting of 15 items. Responses are rated on a 7-point scale ranging from 0 (never) to 6 (every day), with reverse scoring applied to the dimension of reduced personal accomplishment and regular scoring for the other dimensions. The total score ranges from 0 to 90, with higher scores indicating greater severity of work-related burnout. In this study, the Cronbach’s α coefficient for the scale was 0.924. MBI 1–15 see Table 1.
Table 1
Centrality measures of Burnout and turnover intention indicators
Items | Item Description | Betweenness | Closeness | Strength |
|---|---|---|---|---|
MBI 1 | Work makes me feel physically and mentally exhausted. | 5 | 0.015 | 6.090 |
MBI 2 | When I finish work, I feel completely drained. | 0 | 0.015 | 6.334 |
MBI 3 | Getting up in the morning to face another day of work, I feel extremely tired. | 9 | 0.016 | 6.527 |
MBI 4 | Working all day is indeed very stressful for me. | 0 | 0.011 | 3.400 |
MBI 5 | Work makes me feel like I’m on the verge of a breakdown. | 0 | 0.014 | 5.436 |
MBI 6 | Since starting this job, I’ve become increasingly disinterested in it. | 0 | 0.011 | 3.494 |
MBI 7 | I’m not as enthusiastic about work as I used to be. | 4 | 0.014 | 5.883 |
MBI 8 | I doubt the significance of the work I do. | 8 | 0.015 | 5.918 |
MBI 9 | I’m becoming less concerned about whether my work makes a contribution. | 0 | 0.010 | 3.149 |
MBI 10 | I am able to effectively solve problems that arise in my work. | 3 | 0.019 | 9.125 |
MBI 11 | I feel that I am making a valuable contribution to my organization. | 6 | 0.019 | 8.888 |
MBI 12 | I believe I am skilled at my job. | 0 | 0.014 | 5.887 |
MBI 13 | I feel very pleased when I accomplish tasks at work. | 8 | 0.016 | 6.525 |
MBI 14 | I feel that I have completed a lot of valuable work. | 24 | 0.020 | 9.698 |
MBI 15 | I am confident in my ability to efficiently complete tasks. | 17 | 0.019 | 8.866 |
TIQ 1 | Are you considering resigning from your current job? | 5 | 0.011 | 4.578 |
TIQ 2 | Are you interested in seeking other jobs with a similar nature? | 0 | 0.012 | 4.313 |
TIQ 3 | Are you interested in exploring jobs with a different nature? | 0 | 0.011 | 3.496 |
TIQ 4 | Considering your current situation and qualifications, how likely do you think it is to find a suitable position in other organizations? | 5 | 0.016 | 6.977 |
TIQ 5 | If you were aware of a suitable job opening in another organization right now, how confident are you in your chances of securing that job? | 0 | 0.011 | 4.017 |
TIQ 6 | Would you consider resigning from your current job? | 1 | 0.011 | 4.225 |
Turnover intention
In this study, the Chinese version of the Turnover Intention Questionnaire (TIQ) was employed to measure nurses’ turnover intention. This scale, adapted from the work of Michaels and Spector [22], was translated, adapted, and validated by Chinese researchers for assessing turnover intention among nurses [23]. The questionnaire consists of three dimensions: Turnover Intention I, representing the willingness to quit current job; Turnover Intention II, indicating the desire to seek other job opportunities; and Turnover Intention III, reflecting readiness for a new job, each dimension comprising two items. Participants used a four-point scale to rate each item, ranging from 1 (indicating “never”) to 4 (indicating “always”). The total score ranges from 6 to 24, with higher scores indicating stronger turnover intention among nurses. In this study, the Cronbach’s α coefficient for the questionnaire was 0.924. TIQ 1–6 see Table 1.
Statistical analysis
This study utilized IBM SPSS Statistics version 26.0 for descriptive statistical analysis. Categorical variables were described using frequencies and percentages, while continuous variables were described using means and standard deviations.
Network estimation and visualization
For symptom network analysis, this study employed R software version 4.3.2. Gaussian graphical models (GGM) were utilized to estimate symptom networks, employing the graphical least absolute shrinkage and selection operator (glasso) process to construct regularized partial correlation networks. Specifically, the “glasso” package (version 1.11) and the “qgraph” package (version 1.9.8) were used to build and visualize weighted networks of nurse burnout and turnover intention. In the networks, each “node” represented an item of nurse burnout or turnover intention, while each “edge” represented the partial correlation coefficient between node pairs. Considering the ordinal distribution of scores, partial correlation coefficients were calculated using Spearman’s correlation while controlling for all other nodes. To minimize spurious correlations, both lasso and the Extended Bayesian Information Criterion (EBIC) were used to obtain sparse regularized networks. Network visualization was performed using the Fruchterman-Reingold algorithm, with nodes with more connections positioned closer to the network center. Node connections’ strength was indicated by edge width, with nodes of different scales (emotional exhaustion, depersonalization, reduced personal accomplishment, Turnover Intention I, II, III) visualized in different colors to highlight various symptom clusters.
Anzeige
Node centrality
Node centrality serves as an indicator for identifying core symptoms from a mechanistic perspective. We conducted centrality analysis using strength centrality (Str), betweenness centrality (Bet), closeness centrality (Clo), and expected influence (Exp). Higher centrality indices indicate greater importance from a mechanistic perspective. Strength centrality refers to the sum of edge weights connected to a node in the network, while betweenness centrality measures a node’s frequency of occurrence on the shortest paths between all pairs of nodes, indicating its role and influence in connecting other nodes. Closeness centrality is an indicator of the distance between a node and other nodes, measured by the reciprocal of the sum of shortest path distances from that node to all other nodes. Expected influence quantifies a node’s influence across the entire network by calculating the expected impact of each node on other nodes.
Network stability assessment
To assess the accuracy and stability of the results, we employed bootstrapping based on the “bootnet” package (version 1.5.6). Through 1000 nonparametric bootstrap samples, we repeatedly estimated the network model and calculated 95% confidence intervals for edge weights to assess edge stability. Node stability was evaluated by iteratively removing samples during bootstrapping to calculate stability coefficients. Node stability coefficients indicate the proportion of times the parameter correlation value between the reconstructed network and the original network, when a certain proportion of samples are removed, exceeds 0.7, with a minimum threshold of 0.25, indicating relatively stable node indices when exceeding 0.5.
Results
Demographic characteristics
The data show that the majority of nurses participating in this study were female (95.4%), with the largest age group being 36 years and older (38.9%). Regarding educational background, most nurses held a bachelor’s degree (95.2%). Concerning years of working, a majority of the nurses had been working for 5 to 15 years (60.2%). In terms of professional titles, nurses with the title of “Registered Nurse” were the most prevalent (53.0%). Marital status data indicated that most nurses were married (71.6%) and had children (67.3%). Regarding job category, most nurses held permanent positions (64.6%). Monthly night shift data revealed that the number of night shifts per nurse was relatively uniform, with fewer than five shifts being the most common (36.8%). See Table S1.
Network structure
Figure 1 presents the network model of nurse burnout and turnover intentions, in which all edges indicate positive correlations. Within the burnout symptoms, the strongest connection is between MBI 13 and MBI 14 (0.934), followed by MBI 14 and MBI 15 (0.930), and MBI 13 and MBI 15 (0.917). Among the TIQ symptoms, the most robust link is between TIQ1 and TIQ6 (0.871), with subsequent strongest edges being TIQ1 and TIQ3 (0.846), and TIQ3 and TIQ6 (0.808).
Anzeige
Fig. 1
Network structure of burnout and turnover intention among nurses. Note: Entries starting with MBI-GS are related to occupational burnout (brown nodes represent emotional exhaustion, blue nodes represent depersonalization, and green nodes represent reduced personal accomplishment). Entries starting with TIQ are related to turnover intention (pink represents the possibility of quitting the job, yellow represents the motivation to find another job, and light blue represents the possibility of obtaining an external job). Positive and negative partial correlations are indicated by green and red edges, respectively. Closer node proximity and thicker edges indicate stronger relationships between symptoms
×
![]()
Accuracy and stability analysis results
The comprehensive analysis of the closeness (Clo), betweenness (Bet), and strength (Str) of each item within the model reveals that node MBI 14 exhibits high closeness and betweenness, with the highest strength value (Bet = 24, Clo = 0.020, Str = 9.698), identifying it as one of the most influential symptoms in the model. Similarly, MBI 15 shows high closeness and betweenness, and substantial strength (Bet = 17, Clo = 0.019 Str = 8.866), marking it as an influential symptom within the model.
Figure 2 presents line graphs of the centrality indices for symptoms within the nurse burnout and turnover intention symptom cluster network. The results indicate that MBI 14 possesses the highest closeness, betweenness, strength, and expected influence. Conversely, TIQ4, TIQ5, and MBI 10 display the lowest strength values, indicating a marginal effect within the network.
Fig. 2
Comparison of network centrality measures across different indicators
×
![]()
Anzeige
Overall, MBI-G6 and TIQ1 emerge as the most influential nodes in our network model, playing a pivotal role in connecting different parts of the network and in disseminating influence. Other nodes, such as TIQ4, TIQ5, and MBI-GS10, appear to have a more limited role in the network. See Table 1.
The results demonstrate that the CS coefficients for bridge strength, closeness, and strength are all 0.75, suggesting that the network’s structure remains substantially unchanged when 75% of the sample is omitted, indicating good stability. The CS coefficient for betweenness is 0.284, which is below 0.5 but above 0.25, denoting acceptable stability for betweenness.
The accuracy estimates for the network edge weights of nurse job burnout and turnover intention, as shown in Fig. 3, indicate high precision, as evidenced by the narrow grey intervals.
Fig. 3
Bootstrap mean and sample values of network edges in burnout and turnover intention analysis
×
![]()
The accuracy of the centrality indices, shown in Fig. 4, is high for bridge strength, closeness, and strength, followed by betweenness. These findings indicate that the network model for nurse job burnout and turnover intention is both accurate and stable.
Fig. 4
Average correlation with original sample for different centrality measures across
×
![]()
Discussion
The results of this study indicate that MBI 14(I feel that I have completed a lot of valuable work) has the highest centrality in the network, followed by MBI 15(I am confident in my ability to efficiently complete tasks.) and MBI 3 (Getting up in the morning to face another day of work, I feel extremely tired.) , which are core variables in the network of nurse job burnout and turnover intention. Nodes MBI 6 (Since starting this job, I’ve become increasingly disinterested in it.) and TIQ 1 (Work makes me feel physically and mentally exhausted.) have the highest bridging centrality in the network and are considered bridge variables connecting MBI 7 (I’m not as enthusiastic about work as I used to be.) and TIQ 3 (Are you interested in exploring jobs with a different nature?) . Additionally, significant strong ties identified in the current network include the connections between MBI 13 (I feel very pleased when I accomplish tasks at work.) and MBI 14, node MBI 14 and MBI 15, and TIQ 1 and TIQ 6 (Would you consider resigning from your current job?) .
Network analysis reveals that the perception of meaningful work (MBI-14: “I have accomplished many valuable tasks”) is a key factor linking job burnout to turnover intentions. Perceived work value acts as positive feedback, enhancing personal accomplishment and effectively reducing nurses’ turnover intentions. This finding aligns with past research [24], which suggests that accomplishing valuable work boosts nurses’ confidence and reinforces their professional identity. This commitment to their roles mitigates burnout and subsequently reduces turnover intentions [25]. Furthermore, following the COVID-19 pandemic, Chinese nurses took on critical roles in fighting the outbreak and protecting the nation, realizing the importance of their positions. This realization fostered a heightened sense of dedication [26], thereby enhancing their loyalty and reducing their turnover intentions. Therefore, a key strategy to reduce nurse turnover is to highlight the value of their work, potentially a primary focus for future managerial interventions.
Additionally, MBI-15, “I am confident in my ability to handle my work effectively,” is a key variable in the network model. This finding aligns with Luo ‘s research on job burnout and self-efficacy [27]. Self-efficacy, a cognitive process where individuals learn new behaviors through environmental and social influences to improve future outcomes, serves as a psychological foundation for performing duties and achieving organizational goals [28]. Nurses with high self-efficacy value their professional benefits, closely align their personal and career values and goals, and immerse themselves in nursing, thereby reducing their work stress and job burnout [29]. Furthermore, self-efficacy effectively helps individuals regulate emotions. Following the COVID-19 pandemic, nurses with high self-efficacy are likely to maintain emotional stability, calmly and confidently handling complex situations [30], thus mitigating job burnout. Therefore, enhancing nurses’ self-efficacy should be incorporated into nursing management to effectively reduce their turnover intentions and decrease attrition.
This study found strong correlations between MBI 13 and MBI 14, MBI 14 and MBI 15 (0.930), and MBI 13 and MBI 15 (0.917), suggesting that bridge nodes or edges link different symptom clusters. Similar to findings from past research, nurses’ sense of achievement has a higher capacity for information dissemination across the network [31]. The nursing sector is highly competitive with frequent peer evaluations related to professional advancement and skills competitions. Concurrent demands to maintain medical and patient relationships, along with a large workload, predispose clinical nurses to negative psychological states and self-deprecation, which can diminish their sense of personal achievement [32]. Enhancing nurses’ sense of accomplishment at work fosters personal value realization, potential activation, and self-recognition. This enables bridge nodes or edges to play a crucial role in accelerating information dissemination, effectively reducing depersonalization and emotional exhaustion among nurses, and alleviating burnout [33].
Strong connections between TIQ 1 and TIQ 6, TIQ 1 and TIQ 3 (0.846), and TIQ 3 and TIQ 6 (0.808) indicate that bridge nodes or edges link different symptom clusters, with the likelihood of quitting a current job serving as a critical medium for information transmission in the network. Research indicates that young nurses, being in the ascent phase of their careers, have high expectations for their professional development [34]. If their current employment does not meet these needs, they are more likely to experience job burnout and seek better opportunities elsewhere [35]. Therefore, new nurses should be given a supportive working environment, so as to have a good development prospect, improve the enthusiasm and stability of work, the likelihood of nurses leaving their current jobs can be reduced. This strengthens organizational commitment and loyalty, enabling bridge nodes or edges to facilitate rapid information transmission and effectively decrease nurses’ motivation to seek external employment opportunities [36].
Implication
This study provides significant insights for nursing management. MBI 3, which shows high centrality within the network, highlights that nurses’ physical and mental fatigue is a core issue in job burnout. This finding suggests that managers can alleviate nurse burnout by strengthening psychological support—such as establishing mental health hotlines, offering regular wellness workshops, and setting up relaxation areas in the workplace to help staff manage work-related stress [37]. Additionally, MBI 14 serves as a key node within the network, indicating that recognizing nurses’ contributions can foster positive emotions about their work. Hospitals can enhance job satisfaction by providing clear career development paths and implementing recognition programs. For example, quarterly awards such as “Outstanding Nurse” or “Exemplary Nursing Contribution Award” could be established, and success stories could be shared in regular meetings to highlight nurses’ achievements [38]. MBI 6 and TIQ 1 exhibit high bridging centrality, underscoring the importance of a supportive team atmosphere in reducing turnover intentions among nurses. Managers can create and maintain a positive work and learning environment and culture [37, 39].
Limitations
This network analysis study has several limitations that warrant consideration. Firstly, due to the cross-sectional design of this research, it is not possible to entirely rule out the potential for changes in turnover intentions due to fluctuations in job burnout. Secondly, the measurement scales used in this study were self-reported, and although anonymity and other measures were emphasized to mitigate bias, response biases cannot be completely eliminated. Additionally, the sample derived from three hospitals in Hangzhou, Zhejiang Province, China, may limit the representativeness of the findings. Lastly, this study did not perform subgroup analyses, which may limit the generalizability and depth of the findings. Future research should aim to diversify the sample and include more detailed analyses to better understand the factors affecting nurse burnout and turnover intentions across various healthcare settings and specialties.
Conclusion
In summary, this study reveals that MBI 14, MBI 15, and MBI 3 are central to the network of nurse job burnout and turnover intentions. The strong link between MBI 13 and MBI 14 further suggests that managers should focus on these connections, enhancing nurses’ sense of achievement and self-confidence to reduce turnover rates.
Acknowledgements
We sincerely thank all the nurses who participated.
Declarations
Ethical approval
Prior to data collection, this study received approval from the Ethics Committee of the Affiliated Hospital of Hangzhou Normal University (ID: 2023(E2)-KS-121). Informed consent was obtained from all participants.
Consent for publication
No applicable.
Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
