Materials and Methods
Participants and Procedures
The subjects of the investigation were all candidates who signed up for the medical licensing examination at a site in Changchun, Jilin Province. Our study procedures were approved by the Medical Ethics Committee of the School of Public Health Jilin University (No. 20181102) too. The investigators were a postgraduate team from the School of Public Health at a university. We contacted the Changchun Medical Examination Center and asked it to cooperate with us in completing the research work. During the investigation, the investigator explained in advance the purpose and filling requirements of the questionnaire and informed the candidates that participation was voluntary. The questionnaires were then distributed to the candidates and retrieved immediately upon completion.
After data collection, all paper questionnaires were manually entered into a computer database, and items with empty or invalid data were dropped from further analysis. Ultimately, 1778 complete and valid responses were obtained. In this sample, most of the respondents were women (60.74%), were majoring in clinical medicine (81.72%), had a college degree (80.82%), and were studying or working in tertiary hospitals (64.64%), and the mean age was 29.08(±5.57).
The 52 factors from the pre-investigation were used to create the main part of the questionnaire. The question was "If a PMHCI was recruiting medical personnel, please rate the importance of the following factors in their influence on your choice to seek employment at a PMHCI based on your actual situation", and we used a 5-point Likert scale to measure the factors, with anchors ranging from 1 (very unimportant) to 5 (very important). Further, the sociodemographic characteristics of the respondents were collected by the questionnaire (gender, age, degree, major, and study or work institution).
To explore the factors influencing medical personnel to seek employment at a PMHCI and their relationship, we need to not only explore the underlying structure of the factors but also conduct a hypothesis test on this structure. Therefore, in the analysis process, half of the sample was chosen randomly for exploratory factor analysis (EFA), which was used to extract common factors from the items and obtain concise and representative factors and then put forward a hypothesized model. Then, for the other half of the sample, confirmatory factor analysis (CFA) was adopted to test the reliability and validity of the dimensions, and finally, structural equation modeling (SEM) was applied to test the hypothesis.
Specifically, EFA was used for the first sample, in which principal component analysis (PCA) was performed to extract the factors, and varimax rotation (VR) was used to improve the interpretability of the solution[44, 45]. SPSS software (version 23.0) was used for this process. For the second sample, AMOS software (Version 24.0) was used for CFA and SEM. In this part, we performed SEM following the two-step approach recommended by Anderson and Gerbing. First, CFA was carried out for each factor to test whether these factors had a significant factor loading index and to analyze the reliability and validity of the questionnaire. Second, based on the hypothesized path model, the SEM parameters were estimated by the maximum likelihood method. In addition, the model was assessed by the following model fit indexes, with the values in parentheses indicating the cutoffs for acceptable fit[47, 48]: (1) the chi-square value ( ); (2) the chi-square degrees of freedom ( /DF<5); (3) the root mean square error of approximation (RMSEA ≤ 0.08); (4) the comparative fit index (CFI ≥ 0.90); (5) the Tucker-Lewis index (TLI ≥ 0.90); and (6) the incremental fit index (IFI≥ 0.90). All statistical tests were two-sided with the level of significance set at 0.05. Finally, the bootstrap method was used to test the potential mediator effects, and we calculated the total, direct and indirect effects.
Exploratory Factor Analysis and Research Hypotheses
Exploratory Factor Analysis
The EFA results showed that the Kaiser-Meyer-Olkin (KMO) value was 0.966, which was higher than 0.6, indicating the appropriateness of conducting EFA. Further, the result of Bartlett's test was significant ( =33912.366, P<0.001), indicating that the relationship among the items was strong and that the data were suitable for EFA.
Seven factors with eigenvalues greater than one were extracted from the first EFA, and their cumulative contribution rate reached 64.874%. However, six items were gradually deleted because the factor loadings were less than 0.45 or higher than 0.4 on two or more factors simultaneously: Social Position of Medical Personnel, Government Policy Support, Number of Patients, Skill of Existing Personnel, Economy of the City Where the PMHC Is Located, and Organizational Culture. Finally, seven factors were extracted from the remaining 46 items, and their cumulative contribution rate reached 66.896%. Detailed results of the final EFA are shown in the attachment.
Factors Analysis and Research Hypotheses
Sense of Gain (SG)
There were 3 items in this factor: Professional Pride, Fulfill Personal Value and Job-Related Well-being. First, the medical profession has the vital responsibility of providing medical and health services and safeguarding people's health, so the sense of professional pride is high and is closely related to the work quality, job satisfaction and intention to leave of medical personnel[53-55]. Second, in China, the upfront investment of medical personnel, in terms of education and training time, is often longer than that of workers in other industries, and they need considerable knowledge, training and practice to be competent or obtain professional titles. It is thus particularly important for medical personnel to demonstrate their personal value in their work. Finally, Job-Related Well-being is the emotional response of employees at work. Negative emotions can lead to stress, depression and anxiety, while positive emotions can help people thrive in the face of difficulties[57, 58]. Moreover, Job-Related Well-being has been identified as a key area for attracting and retaining employees[59-61], and it also played a decisive role in employment willingness , so we proposed that the manifestation of Sense of Gain determined whether medical personnel were amenable to seeking employment at a PMHCI. Therefore, Sense of Gain was used as a dependent variable to explore the relationship between various factors and the willingness of medical personnel seeking employment at PMHCIs.
Remuneration and Development (RD)
This factor included 11 items, including Wage, Working Bonus, Social Insurance and Accumulation Fund, Position and Professional Title Promotion, etc. These items reflect the concerns about the Remuneration and Development that can be obtained by seeking employment at PMHCIs. Wages and individual development have always been common issues of concern to career groups. For medical personnel, the investment of time and money and the difficulty of obtaining a professional title make them sensitive to remuneration and development issues. In addition, previous studies[27, 63, 64] have reported that remuneration and development are also important factors that affect the work satisfaction and enthusiasm of community health workers. Therefore, we recommended that the sense of gain of medical personnel seeking employment at PMHCIs can be increased through appropriate remuneration and development opportunities.
Internal Organization Development (IOD)
This factor includes 12 items, including Department Setting, Software and Hardware Facilities, Human Resource Allocation, Organizational Management System, Culture and Working Environment, etc. Several studies[65-68] have documented that an organization’s internal development is fundamental to the turnover tendency, happiness, job satisfaction and burnout of medical personnel. Besides, this internal development reflects the diagnosis and treatment capabilities of the institution, and affects patients' identification with PMHCIs and willingness to seek treatment[69, 70]. We proposed that, on the one hand, Internal Organization Development is related to medical personnel’s work arrangements, work pressure, and workload and to the Remuneration and Development, which affects the Sense of Gain of medical personnel seeking employment at PMHCIs as well as patients’ willingness to seek medical treatment in PMHCIs.
Condition of the City Where the PMHCI Is Located (CCPL)
There are 7 items in this factor, including the city’s Development, Environment, Transportation, Culture and Customs, Economy, and Distance from Home and the Reputation of PMHCI Where the PMHCI Is Located. These items reflect the fact that medical personnel not only pay attention to an organization’s internal development but are also concerned about the external environment of the PMHCI. Besides, medical personnel’s wages, individual development, family, and doctor-patient relations are affected by the condition of the city where the PMHCI is located, including economy, culture and customs, which certainly affect their sense of gain as well.
Job Responsibilities (JR)
This factor includes five items: Work Intensity, Stress, Hours, Workload, and Post of Duty. These items reflect medical personnel’s concern about their specific work content and job responsibilities. Several studies[71, 72] have documented that job characteristics are important predictors of job-related happiness. In addition, the arrangement of the work is determined by an organization’s internal development. Therefore, we proposed that Internal Organization Development affects Job Responsibilities and that Job Responsibilities affect the Sense of Gain.
Family Support (FS)
This factor includes 4 variables, Spouse, Children, Parents and House, which represent the family factors considered when medical personnel seek employment at a PMHCI. Studies have pointed out that family support can buffer employees' job stress and prevent negative work-related outcomes such as job burnout, and family members have also been shown to provide both instrumental and affective support, which positively affect employee’s work life. Medical personnel are no exception; a meta-analysis showed that the conflict between work and family has a strong impact on the high turnover rate of medical personnel, and reducing this conflict can improve their happiness. In addition, communication research has shown that the family as a socialization agent conveys both extrinsic and intrinsic work values for developing a professional identity. Therefore, we propose that Family Support will have a positive impact on the professional identity and work enthusiasm of medical personnel and will inevitably be affected by the Condition of the City Where the PMHCI Is Located via the economy, culture, etc.
Patient Factor (PF)
This factor includes local patients' Trust and Respect in Physicians, the Doctor-Patient Relationship and the Moral Character of the Patients. Globally, the concept of medical service has changed from being doctor-centric to patient-centric, which reduces physician dominance, advocates greater patient control, and encourages more mutual participation. However, the lack of coordination and conflict between doctors and patients has aroused widespread concern in society and academia, and it is becoming a serious dilemma facing the medical industry and even society as a whole[79-82]. As such, it is having a substantial impact on the working conditions and psychological pressure experienced by medical personnel[83, 84], and it is likely to lead to work fatigue. Studies[86, 87] have pointed out that the doctor-patient relationship is also an important factor affecting the resignation or career choices of medical personnel. Therefore, we propose that the character, trust in PMHCIs, and respect for medical personnel can reduce work pressure and increase enthusiasm and job-related well-being. In addition, the Condition of the City Where the PMHC Is Located factor, as represented by the city’s economy, culture, etc., can also affect patients’ attitude toward PMHCIs.
Based on the above discussion, the following research hypotheses were proposed:
Hypothesis 1 (H1). Internal Organization Development has a positive effect on Sense of Gain.
Hypothesis 2 (H2). Condition of the City Where the PMHCI Is Located has a positive effect on Sense of Gain.
Hypothesis 3 (H3). Remuneration and Development have a positive effect on Sense of Gain.
Hypothesis 4 (H4). Family Support has a positive effect on Sense of Gain.
Hypothesis 5 (H5). Patient Factor has a positive effect on Sense of Gain.
Hypothesis 6 (H6). Job Responsibilities have a positive effect on Sense of Gain.
Hypothesis 7 (H7). Internal Organization Development has a positive effect on the Patient Factor.
Hypothesis 8 (H8). Condition of the City Where the PwwwMHCI Is Located has a positive effect on the Patient Factor.
Hypothesis 9 (H9). Internal Organization Development has a positive effect on Remuneration and Development.
Hypothesis 10 (H10). Condition of the City Where the PMHCI Is Located has a positive effect on Remuneration and Development.
Hypothesis 11 (H11). Internal Organization Development has a positive effect on Job Responsibilities.
Hypothesis 12 (H12). Condition of the City Where the PMHCI Is Located has a positive effect on Family Support.
Confirmatory Factor Analysis
CFA was used to test the reliability and validity of our measurement instrument. In the process, nine variables were removed because the standardized factor loadings were lower than 0.7. These variables were Organizational Management System, Working Environment, Reputation of the PMHCI, Availability of Drugs, Degree of Emphasis on Clinical, Institution Size, Level of Knowledge, Distance from Home and Post of Duty.
Finally, as shown in Table 1, all the Cronbach’s alpha and composite reliability (CR) values were above 0.8, indicating acceptable reliability for all constructs. Further, the average variance extracted (AVE) value of each construct was above 0.5, and the standardized factor loading of each item was above 0.7, indicating good convergent validity. In addition, as shown in Table 2, the discriminant validity is verified because the square roots of the AVEs of each construct is higher than its correlation . Therefore, we concluded that the remaining items have sufficiently good reliability and validity to test the structural model of our proposed hypotheses.
Since the data for all constructs were collected using the same measurement instrument, we tested the possibility of common method bias. First, the values of the correlation coefficients in Table 2 were all lower than 0.9, indicating that there were no pairs with very strong correlations. Second, the Harman single factor test was conducted by PCA, and the results showed that the first extracted factor in the unrotated solution accounted for 41.02% of the variance, which was less than 50%. Finally, controlling for the effects of an unmeasured latent methods factor, common method bias was tested for. As shown in Table 3, after adding the common method factor, the variance of the fit index was very small, and even the RMSEA value decreased. Therefore, common method bias did not seem to affect the result.
Structural Model Analysis
Table 4 presents the fit of the structural equation model. As can be observed, the hypothesized model fits the data well. The t values of each path were computed to test the hypothesized relationships in our research model in AMOS, and the results are shown in Figure 1. In the hypothesized model, Internal Organization Development (β=0.154; P<0.001), the Patient Factor (β=0.547; P<0.001), Remuneration and Development (β=0.129; P=0.004), and Family Support (β=0.081; P=0.018) had a significantly positive effect on the Sense of Gain of medical personnel seeking employment at PMHCIs. Therefore, H1, H3, H4 and H5 were supported, whereas the hypotheses regarding Job Responsibilities (β=0.055; P=0.053) and Condition of the City Where the PMHCI Is Located (β=0.022; P=0.652) were not supported. In addition, both Internal Organization Development (β=0.377; P<0.001, β=0.344; P<0.001) and Condition of the City Where the PMHCI Is Located (β=0.460; P<0.001, β=0.497; P<0.001) had significantly positive effects on Patient Factor and Remuneration and Development, so H7, H8, H9 and H10 were all supported. Finally, Internal Organization Development (β=0.283; P<0.001) had a significantly positive effect on Job Responsibilities, and Condition of the City Where the PMHCI Is Located (β=0.523; P<0.001) had a significantly positive effect on Family Support, indicating that H11 and H12 were supported. The results provide a useful theoretical perspective for taking corresponding measures to guide medical personnel toward work in PMHCIs.
In addition, to test the mediating role of the Patient Factor, Family Support, Job Responsibilities, and Remuneration and Development, we applied the bootstrapping technique in AMOS[92, 93]. A 95% confidence interval of the indirect effects was obtained with 5000 bootstrap resamples. As shown in Table X, the indirect effects of Internal Organization Development and Condition of the City Where the PMHCI Is Located on Sense of Gain are the significant. The Patient Factor, Family Support, and Remuneration and Development significantly mediated the relationship between the internal and external environment of the institution and Sense of Gain, whereas the mediating effect of Job Responsibilities was not significant.