Mediation Role of Social Support and Burnout Between Financial Satisfaction and Turnover Intention in Primary Care Providers: A Cross-Sectional Study


 Background: Turnover intention is a major cause of reduced team morale and low work efficiency. It hinders work performance and reduces the quality of medical services. This study aimed to investigate the relationship between financial satisfaction and turnover intention and its mediators among primary care providers.Methods: Stratified sampling was used to select 1241 participants from three counties in Anhui province, China. Data were collected using self-administered questionnaire. Turnover intention was assessed with a turnover intention assessment scale. Social support and burnout were measured with the 12-item Perceived Social Support Scale and the Chinese version of the Maslach Burnout Inventory, respectively. Structural equation modeling was used for data analysis.Results: The findings showed high turnover intention among primary care providers (mean score 14.16±4.337), and most providers reported low financial satisfaction (mean score 2.49±0.990). The mean social support score was 64.93±13.229, and only 6.1% of primary care providers reported no burnout. Compared with participants with high financial satisfaction, those with low financial satisfaction were more likely to report higher turnover intention (β=−0.216, p<0.001), more social support (β=0.181, p<0.001), and more severe burnout (β=−0.123, p<0.05). Turnover intention may be related to social support (β=−0.147, p<0.001) and burnout (β=0.239, p<0.001). Furthermore, the effect of financial satisfaction on turnover intention was significantly mediated by social support (β=−0.027, p<0.001) and burnout (β=−0.029, p<0.05).Conclusions: Turnover intention is associated with financial satisfaction, with this association mediated by social support and burnout. A reasonable mechanism needs to be established to improve the financial satisfaction and social support and reduce burnout among primary care providers.


Background
Strengthening the primary healthcare system is one of the ve key targets of China's healthcare reform, which started in 2009 [1]. This healthcare reform substantially improved access to and affordability of primary healthcare, and reduced the disease burden [2]. However, primary healthcare providers face tremendous pressure caused by low levels of training and income. Although the Chinese government has invested many resources in healthcare reform, the reform of the drug system resulted in nancial losses for primary care providers, and incentives for primary care providers may be lacking [3]. Over the past decade, the number of primary care providers increased from 3.3 million to 3.8 million because of the continuing increase in demand for health services, but the proportion of primary care providers among all health workers declined from 40.0-32.6% [4]. The shortage of quali ed primary care providers remains a critical challenge in healthcare, and hinders the further development of the primary care system. Given the long time needed to train quali ed health workers, strategies that can be implemented to reduce burnout and turnover intention among primary care providers are key concerns in the current situation.
Financial incentives act as motivation that affects practice behaviors of primary care providers, and represents a policy option that can be introduced quickly to induce a behavioral response from primary care providers [5]. However, the same income level may be associated with differences in quality of life across different parts of China because of disparity in regional economic levels. Financial satisfaction is a subjective factor that can eliminate the differences associated with the level of regional economic development. Su cient nancial satisfaction can encourage primary care providers to provide better quality care to patients [6], and providers who provide the highest quality care may be those that are most likely to claim incentive payments [7]. Despite the increase in primary care providers' income in recent years, the salary gap between primary care providers and doctors in general hospitals has widened over the past decade, and primary care providers earn around half as much as general hospital doctors [8]. Higher income satisfaction may reduce turnover intention among primary care providers [9] and reduce the pressure from human resources shortages. However, the primary healthcare system cannot attract younger primary care providers because of the low wages and minimal bene ts [2]. Resolving the shortage of doctors working in primary care resulting from nancial dissatisfaction has become a problem for policymakers. Turnover intention is a major challenge in the primary healthcare system, and is de ned as the possibility of an employee resigning within a certain period [10]. Previous studies showed that the proportion of turnover intention among primary care providers in China ranged from 42.3-58.0% [11,12]. This high turnover rate is associated with economic losses. It has been estimated that the recruitment, training and productivity costs associated with turnover may be more than 5% of the total annual operating budget of a major medical center [13]. Furthermore, a heavier workload, reduced team morale, and lower work e ciency are unavoidable in the context of high turnover intention, and subsequently hinder work performance [14]. Turnover intention is affected by various factors, such as job satisfaction, work pressure, and burnout [11]. In addition, nancial dissatisfaction has frequently been associated with turnover intention. A study involving Chinese primary care providers found low nancial satisfaction was associated with turnover intention (odds ratio [OR] = 0.43, P < 0.001) [15]. However, it remains unclear whether there are other factors that mediate the association between income and turnover.
Few studies have investigated the mechanism linking nancial satisfaction with turnover intention among primary care providers. However, understanding this relationship is essential for developing effective measures to reduce turnover intention. Burnout is a prolonged response to chronic emotional and interpersonal job-related stressors, and is characterized by exhaustion, cynicism, and ine cacy [16]. A systematic review of the prevalence of burnout found that 67.0% of people had burnout [17]. As well as contributing to poor job performance, productivity decline, and high absence and turnover rates, burnout can exert a negative effect on a person's colleagues [18]. Furthermore, burnout also negatively impacts quality of sleep, job satisfaction, and coping strategies [19]. Many factors are known to contribute to burnout among primary care providers, including excessive workloads, clerical burdens, work-home con icts, and nancial dissatisfaction [20,21]. Improving income levels can help reduce burnout. Social support is generally de ned as individuals' perceptions of available material, emotional, or spiritual support from other individuals or groups of people [22]. A previous study showed that nancial satisfaction was associated with higher social support [23]. Social support has also been shown to be a predictor of burnout in Chinese healthcare providers [24]. In addition, social support was found to be a reason for turnover intention, and also acted as a partial mediator between job satisfaction, burnout, and turnover intention [25].
Therefore, this population-based study aimed to: a) explore the associations between nancial satisfaction and turnover intention among primary care providers, and b) test the extent to which the association between nancial satisfaction and turnover intention may be mediated by social support and burnout. Our ndings will contribute to improving work e ciency and alleviating the shortage of primary care providers.

Study Design and Population
We conducted a cross-sectional study in three counties in Anhui province. Anhui province is a major province in the east of China with a large population. Anhui is divided into northern, central, and southern parts because of economic, cultural, and demographic differences. Anhui was the pilot province for the primary healthcare system in China. The primary care providers in this study included medical staff from community health service centers, township health centers, and village clinics.
A random sampling method was used to select one county from the northern, central, and southern areas of Anhui province (three counties in total). A questionnaire was conducted with primary care providers in the selected counties and towns. Each participant completed a self-administered questionnaire independently and anonymously, but research staff was available to address any questions. In each district, the questionnaires were collected by research staff immediately after they were completed by primary care providers. The exclusion criteria for our study were: a) worked for less than 1 year; b) selected the same option for all questions; and c) incomplete answers to the questionnaire. For the present analysis, we included 1214 participants with complete information for the variables analyzed.
Several training sessions were conducted for research staff before the investigation. The study was approved by research ethics committee, and all participants voluntarily participated in the survey after providing their informed consent.

Financial satisfaction
The independent variable was primary care providers' nancial satisfaction. Financial satisfaction was measured using one item that was developed and used in a previous study [26]. Participants were asked: "How satis ed are you with your nancial situation?" Response options were: "very satis ed," "satis ed," "somewhat satis ed," "dissatis ed," and "very dissatis ed."

Turnover intention
The dependent variable was primary care providers' turnover intention. Turnover intention was measured with a modi ed version of a turnover intention assessment scale, which consisted of six items. The Chinese version of this scale was developed by Mickaeland et al. [27]. Responses were on a four-point Likert ranging from 1 = "never" to 4 = "always." The total score of the six items was computed as the score for turnover intention, which ranged from 6 to 24; a higher score indicated a stronger turnover intention. A previous study conducted among general practitioners reported the Cronbach's α for the scale was 0.740 [28], indicating that the scale had good internal consistency.

Social support
The 12-item Perceived Social Support Scale was used to measure perceived social support [29]. This instrument contains 12 items on three subscales: family (items 3, 4, 8, and 11), friends (items 6, 7, 9, and 12), and others (items 1, 2, 5, and 10). The items were rated with a seven-point Likert scale from 1 ("de nitely disagree") to 7 ("de nitely agree"). The total score is an equally weighted sum of the 12 items, ranging from 12 to 84. A higher score indicates more social support. The Cronbach's α was 0.940 in this study, which indicated that the questionnaire was applicable. Burnout Burnout was measured using the Chinese version of the Maslach Burnout Inventory [30], which was revised by Li et al. and used in a previous study [31]. This adapted version consists of 15 items grouped on three dimensions, each with ve items: emotional exhaustion, cynicism, and e cacy. Responses were on a seven-point scale from 1 (never) to 7 (always), and total scores ranged from 15 to 105. Burnout was divided into four grades: no burnout, slight burnout, moderate burnout, and severe burnout. The Cronbach's α for the scale was 0.816 [31].

Covariables
To control for potential confounding variables, age (years), gender, and marital status were examined as adjustment variables.

Statistical Analysis
Descriptive statistics were used to summarize the distribution of participant characteristics, the frequency of occurrence of categorical variables (percentages), and the means and standard deviations (SD)of continuous variables.
Structural equation modeling (SEM) is an optimal statistical technique for evaluating a priori models, identifying mediators, and elucidating direct and indirect paths between variables. We used structural equation modeling to investigate the associations between nancial satisfaction, turnover intention, and potential mediator mechanisms. Mediation analysis, which is part of structural equation analysis, was used to analyze the relationships between exposure, outcome, and mediating variables. In Fig. 1, nancial satisfaction was treated as an exogenous variable, estimating two latent variables (social support and burnout).
Maximum likelihood was used to estimate model parameters. The model t was estimated by: a) ratio of chi-square value to degrees of freedom (χ 2 /df) ≤ 3; b) comparative t index (CFI) ≥ 0.90; c) standardized root mean square residual (SRMR) < 0.05; and d) root mean square error of approximation (RMSEA) < 0.05 [32]. Descriptive analyses were performed using SPSS 25
We found that 6.1% of participants had no burnout, 45.0% had slight burnout, 37.6% had moderate burnout, and 11.3% had severe burnout. The mean nancial satisfaction score was 2.49 (SD = 0.990) out of a possible total score of 5. The mean score for turnover intention was 14.16 (SD = 4.337), indicating that participants had a moderate turnover intention. Furthermore, the mean social support score was 64.93 (SD = 13.229) ( Table 1).

Structural Equation Modeling
Models were built to estimate the relationships between nancial satisfaction, turnover intention, social support, and burnout (Fig. 1). The model t indices and standardized direct effects among four variables and indirect effects of two mediation variables are shown in Tables 2 and 3, after adjusting age, gender, and marital status. The model t indices were evaluated as: χ 2 /df = 2.909, CFI = 0.978, SRMR = 0.039, RMSEA = 0.040 and 90% con dence interval (CI) = 0.033 to 0.046, which showed the model had good t.

Relationships Among Variables
In the SEM, nancial satisfaction had a signi cant direct effect on turnover intention (β=−0.216, 95% CI − 0.279 to − 0.152); participants with higher nancial satisfaction were more likely to report low turnover intention. Financial satisfaction was signi cantly associated with two mediating variables. Higher nancial satisfaction was associated with an increased likelihood of having burnout (β=−0.123, 95% CI − 0.210 to − 0.017) compared with low nancial satisfaction. In terms of social support, nancial satisfaction often re ected higher social support (β = 0.181, 95% CI 0.125-0.234) for primary care providers.
Associations were also found between turnover intention and the two mediating variables. A higher social support score was associated with lower turnover intention (β=−0.147, 95% CI − 0.217 to − 0.080). In addition, turnover intention was signi cantly in uenced by burnout (β = 0.239, 95% CI 0.140 to 0.328), and participants with higher burnout were more likely to report higher turnout intention.
Furthermore, the association between two the mediating variables (social support and burnout) was signi cant, with low social support indicating higher burnout (β=−0.270, 95% CI − 0.355 to − 0.178). The direct and indirect effects were signi cant before and after the covariates were added into the model. More details are shown in Table 2.

Mediation Role of Social Support and Burnout
The multi-serial mediator model demonstrated that social support played a potential mediating role in the association between nancial satisfaction and turnover intention (β=−0.027, 95% CI − 0.042 to − 0.013). Burnout also displayed a mediator effect in the relationship between nancial satisfaction and turnover intention (β=−0.029, 95% CI − 0.059 to − 0.003). In addition, the indirect effect on nancial satisfaction and turnover intention, mediated by social support to burnout, was signi cant (β=−0.012, 95% CI − 0.019 to − 0.005). Overall, the three indirect pathways between nancial satisfaction and turnover intention were not greater than the direct pathway. In terms of three potential mediating pathways, burnout was greater than the other two pathways.

Discussion
This study presented critical information regarding the current pro le of nancial satisfaction, turnover intention, social support, and burnout and their relationships among primary care providers in China. Our results showed that nancial satisfaction was strongly related to turnover intention, and individuals who were dissatis ed with their nancial situation had a higher risk for turnover intention However, this correlation occurred through both direct and indirect effects. Financial satisfaction had a signi cant indirect effect on turnover intention mediated by social support and burnout. Moreover, the pathway from social support to burnout played a mediating role between nancial satisfaction and turnover intention.
The results showed that primary care providers were generally dissatis ed with their nancial situation. This nding was consistent with a previous study that found primary care providers were the least satis ed with their income [33]. Since 1980, primary care providers in China have been able to charge a 15% markup on drug sales, but this was abolished in 2017 because of over-prescribing of drugs. The government increased its subsidy for primary healthcare institutions, but this policy substantially reduced the incomes of primary care providers [2]. The mean score for turnover intention was 14.16, which was similar to previous results for general practitioners [28]. The possible interpretation for the high-level turnover intention among Chinese primary care providers is their heavy workload, low income level, and few professional development opportunities [15]. Moreover, 11.3% of participants reported severe burnout, which was higher than that reported among Chinese primary care doctors [34] and primary care providers in the US [35]. The mean score for social support in this study was 64.93, which was lower than reported in a previous study involving Chinese physicians [25].
To our knowledge, this is the rst population-based study among Chinese primary care providers that investigated the relationship of nancial satisfaction and turnover intention as well as the mediating roles of social support and burnout. Prior studies observed correlations between nancial satisfaction and turnover intention [9,28]. However, those studies used different participant groups (e.g., specialist physicians) or analysis methods (e.g., traditional regressions and correlations). Although those studies differed from ours in terms of speci c details, the results regarding the negative relationship between nancial satisfaction and turnover intention were consistent, which con rms the results of our study.
This study found that nancial satisfaction was associated with an increased likelihood of high social support among primary care providers. A study conducted among primary healthcare patients in Brazil suggested that people with a high income were more likely to receive more social support than those with a low income [36]. The workloads of primary care providers have surged since the introduction of the basic public health services program [37]. Furthermore, compared with workers with a low income, those with a high income may be able to get more support from coworkers [38]. However, income levels and workloads appear to be mismatched. Financial dissatisfaction may also reduce the desire for social intercourse, and increase pressure in family life. Support from social networks may therefore be reduced by nancial dissatisfaction and lack of communication. We found a correlation between low social support and high risk for turnover intention, which was consistent with a previous survey of Chinese specialist physicians and emergency room nurses [25,39]. Social relationships are established by primary care providers in daily life and also at work (e.g., physician-nurse, physician-leadership, and physician-patient relationships). Primary care providers can also receive encouragement and support from their daily life networks (e.g., family and friends) or from work networks (e.g., leaders and colleagues) when faced with increased workloads. These factors may explain the mediating role of social support in nancial satisfaction and turnover intention observed in this study. Financial dissatisfaction may lead to decreased social support from family, friends, and colleagues because of lack of communication and increased life pressure, and turnover intention increases when there is a lack of social support. Financial dissatisfaction means a higher level of burnout for primary care providers; although this has not been con rmed in primary care providers, it has been described in previous studies conducted among Korean doctors [9] and Chinese nurses [40]. Moreover, an association between burnout and turnover intention was observed in this study, which was consistent with earlier ndings for physicians in the United States [41]. Factors predicting burnout include low reward, excessive workload, low organizational status, and low support [42]. Medical and basic public health services provided by primary care providers that resulted in excessive workloads did not increase their income. Furthermore, there is no professional rank promotion system for primary care providers in China, which means it is di cult for them to gain a higher organizational status [43]. This may explain the associations between nancial satisfaction, burnout, and turnover intention in primary care providers in this study. A mediating role of burnout between nancial satisfaction and turnover intention was found in our study. Financial satisfaction was a predictor factor of burnout, and low nancial satisfaction was re ected in a high level of burnout; that is, the higher level of burnout, the greater risk for turnover intention. In addition, we found a multi-serial mediation role of burnout through social support between nancial satisfaction and turnover intention. Financial dissatisfaction may reduce social support through decreased social intercourse and increased family life pressure. Decreases in social support, especially support from colleagues, may increase the risk for burnout, thereby leading to increased turnover intention.
We selected age, gender, and marital status as covariables to control potential confounding variables in the SEM. The direct and indirect effects were signi cant before and after the covariates were added in the model.
The strengths of this research include the use of SEM, which is an analytical method suitable for evaluation and measurement that can eliminate measurement errors for variables that are di cult to measure, such as social support, burnout, and turnover intention. This study also showed a direct effect among four study variables and the mediation linking nancial satisfaction with turnover intention, which is the rst time this has been shown among primary care providers.
This study had several limitations. First, the measurements of nancial satisfaction, turnover intention, social support, and burnout were obtained using self-administered questionnaires, meaning self-report bias might have impacted the results. Second, this study was a cross-sectional study, and the interpretation of causal inferences on the results is limited. Finally, the sample was selected from one province, so the extrapolation of conclusions to the national level could be challenged. These limitations need to be addressed in further research.

Conclusion
Our study indicates that nancial dissatisfaction, low social support, and burnout can aggravate turnover intention. Furthermore, social support and burnout mediate the association between nancial dissatisfaction and turnover intention. Our study offers an exploration of possible pathways between nancial dissatisfaction and turnover intention, and may have implications for reducing turnover intention among primary care providers. Ultimately, our research ndings may help to alleviate the shortage of primary care providers and improve the quality of medical services. First, remuneration mechanisms need to be established to increase the income of primary care providers; for example, income could be determined by workload or professional level. Second, suitable professional promotion and organizational management systems need to be established to encourage retention among primary care providers. Finally, measures to improve social support and reduce burnout are needed, such as decreased workloads and improved organizational status for primary care providers. Abbreviations SEM Structural equation modeling.

Declarations
Ethics approval and consent to participate The study was approved by the Ethical Committee of Anhui Medical University (AMUREC:20170260).
Written informed consent was provided by all participants prior to answering the questionnaire, including consent for de-identi ed data to be used in publications arising from the research.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.