Analysis of factors influencing attrition among clinical traditional Chinese medical major graduates: based on a discrete choice experiment

Abstract

Background

Traditional Chinese Medicine (TCM) is essential for promoting health worldwide. Despite governmental support, TCM faces critical challenges, including workforce shortages and high attrition rates, with many clinical TCM graduates pursuing non-clinical careers. This study aims to explore the employment preferences of these graduates and to propose strategies for retaining TCM professionals, optimizing the input–output ratio of TCM education, and supporting the sustainable development of the TCM industry.

Methods

A stratified sampling method was used to select clinical TCM major graduates from 16 universities of Chinese medicine. A Discrete Choice Experiment (DCE) was conducted to investigate the factors influencing the attrition of clinical TCM professionals, including monthly income, work location, housing security, workload, work environment, and promotion opportunities. A mixed logit model was applied to analyze the DCE data.

Results

A total of 273 clinical TCM major graduates were included in the primary DCE analysis. The monthly income was more important than other attributes (RAI = 48.03%). Among non-economic factors, participants expressed the strongest willingness to decrease workload, being willing to forgo 3,370.312 yuan (approximately $500.3) in monthly income to reduce heavy workloads to moderate levels. Other significant factors included work environment (RAI = 13.64%), housing security (RAI = 9.47%), and promotion opportunities (RAI = 5.33%), with work location being the least important (RAI = 3.3%). Subgroup analysis showed that rural graduates were more willing than urban graduates to forgo monthly income for promotion opportunities. Graduates from the central region of China valued work environment and housing security more than those from the eastern and western regions. Postgraduates were more likely than undergraduates to give up monthly income for housing security.

Conclusion

In addition to economic factors, non-economic factors are also critical considerations for clinical TCM major graduates in their employment decisions. A combination of measures, including offering lower workloads, better work environments, housing security, and promotion opportunities, should be adopted to stabilize the employment environment for clinical TCM graduates.

Text box 1. Contributions to the literature
• Economic factors, such as monthly salary, are essential in job selection, but non-economic factors like workload and work environment also play a key role. The combination of the non-economic incentives strongly attracts TCM graduates to TCM-related careers.
• Regional and educational differences in graduate preferences offer new insights for workforce planning and policy development.
• Stable and favorable working conditions for TCM graduates are crucial to preventing talent loss, ensuring workforce stability, and supporting the sustainable development of the TCM industry.

Backgrounds

With a long history, Traditional Chinese Medicine (TCM) is a treasure of Chinese civilization, playing an irreplaceable role in maintaining human health as a traditional Chinese healing method [1]. With the widespread adoption of TCM in 183 countries and regions worldwide, its international recognition continues to grow [1]. According to the World Health Organization, 29 countries have passed laws and regulations to promote the development of traditional medicine [2]. The “Global Report on Traditional and Complementary Medicine 2019” and the “WHO Traditional Medicine Strategy: 2014–2023” further emphasize the importance of traditional medicine in achieving universal health coverage [2, 3], particularly in promoting the accessibility of safe, high-quality, and effective traditional medicine.

In this context, the Chinese government places great importance on the development of TCM and has issued laws, regulations, and policy measures to support its inheritance and innovative development [4,5,6,7]. For example, the implementation of policies, such as the “14th Five-Year Plan for Traditional Chinese Medicine Development” [4] and the “Healthy China 2030” [7] initiative, has further promoted the standardization and regulation of TCM. These policies not only facilitate the dissemination and application of TCM both domestically and internationally, but also provide institutional support for the training and reserve of TCM talent.

Despite the important role of TCM in preventing and controlling major chronic and infectious diseases, the development of TCM professionals still faces numerous challenges, including insufficient overall numbers and uneven distribution [4]. According to the latest data from the “China Health and Health Statistical Yearbook”, there were 4.29 million practicing physicians (including physician assistants) nationwide in 2021, of which 731,700 were TCM practicing physicians (including physician assistants), accounting for only 17% [8]. Furthermore, a survey by the National Health Commission revealed that over 90% of rural health clinics lack TCM-related professionals, and approximately 70% of township health centers also face a shortage of TCM professionals [9]. Due to constraints such as unfavorable working conditions and relatively low salaries, many graduates from TCM universities choose to pursue other vocations after graduation, exacerbating the shortage at the grassroots level [10]. Graduate reports from several TCM universities show a 2–3% decline in the proportion of graduates entering healthcare institutions in 2023 compared to 2022 [11, 12], with job stability also decreasing [11], as nearly 40% of graduates who leave their positions cite low salaries as the reason [13]. This situation not only indicates that significant national investment in TCM have not yielded the expected returns, but also exacerbates the uneven distribution of healthcare human resources, particularly in rural areas [14]. The growing talent gap and high attrition rates pose a serious threat to the inheritance and sustainable development of TCM [15]. Therefore, it is crucial to explore the underlying reasons for this problem.

Existing studies on the attrition of TCM graduates primarily focus on descriptive analyses of their current status and the reasons for their departure, with little attention to the employment preferences that influence their career decisions [16, 17]. This study aims to reveal the employment preferences of clinical TCM graduates and explore the reasons for the attrition of TCM professionals through a Discrete Choice Experiment (DCE). DCE is a widely used tool for measuring stated preferences and has been extensively applied in employment preference research within the healthcare sector [18], including studies on medical and pharmacy graduates [19,20,21,22,23,24]. While employment preference studies in other healthcare fields are more common, research focusing specifically on the preferences of TCM graduates is still limited. The findings of this study are expected to provide valuable insights that can inform efforts to optimize TCM training programs and enhance talent retention.

Materials and Methods

Sampling

This study employed a combination of convenience sampling and stratified cluster sampling to select samples from 16 universities of Chinese medicine across 16 provinces in China, covering three major regions: eastern (7 universities), central (4 universities), and western (5 universities) to ensure geographic representativeness. From April 22 to May 22, 2024, an anonymous online voluntary survey was conducted, attracting participation from 3,166 undergraduates and graduate students majoring in TCM. To gain deeper insights into the factors contributing to the attrition of graduates from TCM clinical programs, we incorporated screening questions before the main survey. Respondents were initially asked whether they had considered pursuing a career in TCM clinical practice after graduation. Those who answered 'no' were directed to proceed directly to the main part of the questionnaire.

The inclusion criteria for the sample consisted of enrolled TCM major graduates who did not express a willingness to pursue TCM-related jobs after graduation. Data collection was conducted using a professional online research platform (Wen-Juan-Xing), and the survey link was distributed to the target population via WeChat, a popular Chinese social media site. All respondents voluntarily participated in the survey after fully understanding the research content and signing an informed consent form.

The sample size calculation used the “thumb rule” proposed by Johnson and Orme[25], which is one of the commonly used methods in DCE studies. This method determined that the minimum sample size required for this study was 75 respondents. Considering the potential for a low questionnaire response rate and the need for subsequent subgroup analyses, the study aims to recruit a minimum of 100 respondents.

Selection of attributes and levels

The determination of attributes and levels is a key aspect of DCE studies. This study identified key attributes through a literature review and qualitative research. Firstly, 10 attributes related to the employment intentions of TCM major graduates were preliminarily identified based on the literature review, including monthly income [19,20,21,22,23, 26], housing security [23], work location [19,20,21,22,23, 26], work environment [19, 20, 23, 26], workload [19, 21, 26], on-the-job training and promotion opportunities [19, 20, 26], staffing [19,20,21, 23], social reputation/recognition [27,28,29], transportation [21], and educational opportunities for children [23]. Then, after multiple rounds of qualitative research, including in-depth interviews, focus group discussions, and expert consultations, the final attributes and their levels to be included in the DCE were determined.

Informed consent was obtained from all participants before the qualitative research. During the in-depth interview phase, interviews with 16 TCM major graduates from Shanghai University of Traditional Chinese Medicine, Henan University of Chinese Medicine, and Guizhou University of Traditional Chinese Medicine were conducted to gain deeper insights into their feelings and preferences regarding career choices. Throughout the interviews, attributes that had minimal impact on graduates’ employment decisions were identified. 4 non-key attributes, “staffing”, “social reputation/recognition”, “transportation” and “educational opportunities for children” were eliminated during the process. These attributes were considered to have a limited direct influence on career choices or to be significantly affected by other factors, such as family background, making it difficult to present clear effects independently in the DCE.

During the focus group phase, definitions of the remaining attributes were optimized. Workload encompasses factors such as work shifts, on-call duties, rest periods, emergency arrangements, work-life balance, daily workload, and overtime situations. The work environment was further divided into macro and micro environments [21]. The macro environment includes the societal recognition of the profession, while the micro environment includes specific working conditions, such as cafeterias, shuttle services for commutation, childcare facilities, management style, doctor-patient relationships, staffing levels, and the availability of equipment and medications. Within the micro environment, the physical environment refers to the quality of workplace facilities, lighting, temperature, and noise levels, while the social environment pertains to the team atmosphere, colleague relationships, management style, and organizational culture [30].

To ensure that the study is precise and the attributes and levels are reasonable, 2 experts in the field of DCE and 6 experts working in related areas of TCM were invited. Based on the experts’ recommendations, the attributes “on-the-job training and career development opportunities” and “housing benefits” were retained and modified to “promotion opportunities” and “housing security”. Additionally, the monthly income levels were adjusted, ranging from 3,000 to 8,000 CNY, equivalent to approximately 445.3 to 1,187.5 USD (based on 2022 OECD data, 1 USD = 6.737 CNY). Table 1 provides a detailed list of all attributes and their levels.

Table 1 Attributes and levels

Experimental design and questionnaire development

This study followed standard methodologies to design the DCE in order to achieve unbiased and statistically valid responses [31]. Based on the work of Huber and Zwerina [23], the design method used the %MktRuns macro command in SAS 9.4 for efficient design, generating 18 sets of choice scenarios. To reduce the burden on respondents, the questionnaire was randomly divided into two versions, including 2 scenarios respectively, and a dominance task was included to test internal validity. In the dominance task, the levels in Scenario 1 were significantly superior to those in Scenario 2; if respondents did not choose Scenario 1, their responses would be considered questionable. Thus, each version of the questionnaire ultimately included 10 sets of choice scenarios, with respondents being randomly assigned to either version. A sample of the DCE choice set can be found in Appendix Table S1.

To assess the comprehensibility and acceptability of the questionnaire, a pre-test before the formal survey were conducted. The pre-test took place from March 26th to April 18th, 2024, and involved participants from 3 universities: Guangdong University of Chinese Medicine, Hubei University of Chinese Medicine, and Gansu University of Chinese Medicine, respectively located in eastern, central, and western China. Based on feedback from this pre-test, researchers modified the language and format of the questionnaire.

The study distributed and collected the questionnaire via an online research platform, which automatically categorized and analyzed the data, eliminating the need for manual entry. The study’s purpose and informed consent are explained at the beginning of the questionnaire, which assure participants of no risks, free participation, and confidentiality of their information. Upon agreeing, respondents could proceed to complete the questionnaire.

Data analysis

All analyses were conducted by Stata 15.1 (StataCorp). In each simulated scenario, the utility function for job choice consists of a deterministic part V_ni and an unobservable part ε_ni. The utility function for individual n choosing job i is defined as follows:

U_ni = β₁ Monthly Income + β₂ Work Location _{Second-tier City} + β₃ Work Location _{Third-tier City} + β₄ Housing Security _{Housing Benefit} + β₅ Housing Security _{Provided Housing} + β₆ Workload _Moderate + β₇ Workload _{Relatively Light} + β₈ Work Environment _Normal + β₉ Work Environment _Excellent + β₁₀ Promotion Opportunities _{Moderate Likelihood} + β₁₁ Promotion Opportunities _{High Likelihood} + ε_ni.

This study employed both the Conditional Logit Model (CLM) and the Mixed Logit Model (MXL) for data analysis. The Conditional Logit Model is often used in DCE due to its simplicity and widespread acceptance [32]; however, it is limited in its ability to address preference heterogeneity among individuals [33]. In contrast, the Mixed Logit Model is not limited by the distribution type of random utility, offering high flexibility and the ability to account for preference heterogeneity across different individuals [34]. It has been widely used for analyzing data from DCE [35]. By comparing the Log Likelihood (LL), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC), the Goodness of Fit of these two models were evaluated. The model with the better result would be selected as the final main effects model result.

All attribute levels were encoded as dummy variables. Except for monthly income, which was treated as a continuous variable, all other attribute levels were treated as categorical variables. Based on the main effects model results, the relative importance of each attribute was calculated by taking the range of coefficients for that attribute and dividing it by the total sum of the ranges for all attributes. Additionally, the researchers calculated the marginal rate of substitution between monthly income and other non-economic factors [36], interpreting the results as the amount of monthly income graduates were willing to forgo to improve working conditions. To better explain the preference differences among respondents with varying sociodemographic characteristics, the respondents were grouped by gender, urban or rural origin, school distribution, education level, and annual household income. Subgroup analyses were conducted applying the Mixed Logit Model and the willingness to discard for different subgroups was calculated. Lastly, the scenario analysis was performed to assess how the probability of job choice varied under different attribute levels.

Result

Participants’ characteristic

A total of 3,166 responses were collected in this study. Among these, 321 respondents indicated that they did not intend to pursue TCM-related clinical jobs after graduation, while 2,683 expressed a willingness to do so. Additionally, 162 graduates refused to participate in the survey. From the 321 respondents who did not express a willingness to pursue TCM-related careers, 48 responses were excluded due to failing an internal validity check.

After the exclusion, 273 valid samples were collected for analysis. The average age of the analyzed sample (n = 273) was 24.4 years (SD = 3.1). Most respondents were female (66.3%), and a majority (58.6%) had their household registration in rural areas, including suburban regions. The distribution of respondents by school location was as follows: 34.4% from the eastern region, 14.7% from the central region, and 50.9% from the western region. One-third of the respondents (35.5%) had postgraduate degrees, and 59.3% came from families with an annual income over 50,000 CNY (Table 2).

Table 2 Demographic characteristics

Main effects model results

This study employed both a Conditional Logit model and a Mixed Logit model. After comparing the Goodness-of-Fit between the two models (see Appendix Table S3), the Mixed Logit model, which demonstrated better fit, was selected as the main effects model. Additionally, after comparing the full sample with the sample excluding the 48 respondents who did not pass the internal validity check (Appendix Table S2) through a sensitivity analysis, significant differences were found between the two sets of results. To ensure the accuracy of the estimates, respondents who failed the internal validity check were excluded as the sample size was sufficient. As a result, 273 respondents were included in the final analysis.

The statistical analysis results are presented in Table 3. Among the 6 attributes included in the study, at least one level of each attribute had a statistically significant impact (p < 0.05), indicating that all the attributes examined in this study are key factors influencing graduates’ job preferences. The standard deviations of the estimated coefficients for work environment, workload, and work location were statistically significant (p < 0.05), suggesting preference heterogeneity among graduates for these attributes. In contrast, the preferences for housing security and promotion opportunities exhibited homogeneity (p > 0.05).

Table 3 Mixed logit estimates and WTD (n = 273)

The results of the Mixed Logit model indicate that among the non-economic factors (work location, housing security, workload, work environment, promotion opportunities), lighter workload provided the highest utility for graduates (β = 1.635, p < 0.001), followed by excellent work environment (β = 1.103, p < 0.001). Graduates also preferred jobs offering accommodation (β = 0.766, p < 0.001) and those with a higher likelihood of promotion within three years (β = 0.431, p < 0.001). In terms of work location, graduates’ preference for first-tier cities was not significant; instead, they were more inclined to choose third-tier cities over second-tier cities (β = -0.266, p = 0.01).

The relative importance analysis in Fig. 1 shows that monthly income (48.03%) is the most prioritized factor for graduates when seeking employment, with its importance nearly equal to the sum of all other attributes combined. Workload ranks second in importance (20.23%), followed by work environment (13.64%). In contrast, work location is considered the least important attribute (3.3%). Housing security (9.47%) and promotion opportunities (5.33%) hold slightly more importance than work location.

Fig. 1

The relative importance of attributes

Willingness to discard

Table 3 presents graduates’ willingness to forego a portion of their monthly income in exchange for changes in job attributes. Compared to heavy workload, graduates are willing to accept a monthly income reduction of 3,370.31 CNY ($500.30) for lighter workloads, or 1,946.90 CNY ($289) for moderate workloads. To gain an excellent work environment rather than a poor one, graduates are willing to give up 2,272.46 CNY ($337.30) of their monthly income. In terms of housing security, graduates are prepared to forgo 1,578.20 CNY ($234.30) in monthly income if housing is provided, compared to having no provided housing or housing benefits. Additionally, if there is a strong possibility of a promotion within three years, graduates are willing to give up 887.45 CNY ($131.70). When comparing work locations, graduates expect compensation of 549.07 CNY ($81.50) if the job is located in a second-tier city rather than a third-tier city.

Subgroup analysis

The results of the subgroup analysis are shown in Fig. 2. Respondents with different sociodemographic characteristics exhibited distinct preference patterns, and most attributes had statistically significant effects on graduates’ job preferences. Based on the WTD (Willingness to Discard) results, lighter workloads were more attractive to female graduates, graduates from urban areas, graduates in central regions, undergraduates, and graduates from higher-income families. Notable differences emerged between genders: male graduates were more willing to forego a larger portion of their monthly income for housing benefits, whereas female graduates favored excellent work environments and were willing to give up more monthly income for such conditions. Between rural and urban graduates, the former were more inclined to sacrifice a larger portion of their monthly income for jobs that offered a higher likelihood of promotion within three years and housing benefits, while the latter placed higher value on workload and environment. Regionally, graduates from central region showed a greater willingness to forego monthly income for excellent work environments and housing benefits compared to graduates from the eastern and western regions. Regarding education level, graduate graduates placed more emphasis on housing benefits, while undergraduates were more willing to sacrifice monthly income for improved work environments. Additionally, family income affected graduates’ preferences: graduates from high-income families preferred jobs with fast promotions and good work environments, whereas those from low-income families were more inclined to choose positions offering housing benefits.

Fig. 2

The WTD of Subgroups

Scenario analysis

Figure 3 illustrates respondents’ preferences for TCM jobs under various scenarios. Compared to the baseline job conditions—such as a monthly income of 3,000 CNY, no housing benefits, high workload, a poor work environment, and low promotion opportunities within three years in a third-tier city—the study found the following:

Fig. 3

Graduates’ Probability of Choosing the Simulated Job Scenarios

If the likelihood of promotion within three years significantly increased, the probability of respondents choosing the new job rose to 60.6%.

If the new job offered housing benefits, the likelihood increased to 68.26%.

If workload was reduced to moderate or light, the probability of choosing the new job climbed to 72.01% and 83.69%, respectively.

If the work environment improved to an excellent level, the likelihood of selecting the new job increased to 75.08%.

Notably, the most attractive incentive among individual factors was an increase in monthly income to 8,000 CNY, which appealed to the majority of respondents, with 91.88% expressing a preference for this option. A combination of non-economic incentives also significantly boosted respondents’ willingness to choose the new job. For instance, if the job offered “a high likelihood of promotion within three years + housing security + an excellent work environment + light workload”, nearly all respondents (98.08%) would opt for the new position.

Discussion

This study is the first DCE study targeting clinical TCM major graduates, aiming to assess their decision-making behavior in various scenarios and uncover the key factors influencing the attrition of TCM clinical graduates. The study identified six attributes closely related to their job preferences. Mixed Logit model analysis revealed that monthly income, work location, housing security, workload, work environment, and promotion opportunities all significantly impact graduates’ employment choices. Among these, monthly income was the most valued factor, with its importance far outweighing other non-economic factors. Of all incentives, increasing monthly income from 3,000 to 8,000 CNY had the greatest impact on job selection, aligning with findings from studies on graduates in other health-related professions in China [19,20,21, 26, 30]. However, scenario analysis indicated that a combination of improvements in non-economic factors can produce similar effects to the economic incentive. Among these non-economic factors, workload and work environment had the most significant influence on job choices, while the impact of other attributes was relatively minor.

Workload: a significant factor in the employment choices of clinical TCM graduates

Workload is the most important non-economic factor for these graduates, with a preference for positions with lighter workloads. They are willing to forgo approximately 3,370 CNY per month in exchange for a reduced workload. This trend is particularly pronounced among female graduates, those from urban areas, and those from higher-income families. This may reflect a shift in the values of younger generations [37, 38], with many graduates seeking to prioritize personal life and family over long hours of intense work. The high workload in the health-related professions often leads to burnout and health issues. TCM clinical graduates, especially women, are more focused on their physical and mental well-being, opting for lighter workloads to reduce stress and maintain a healthy work-life balance. However, in certain unique social contexts, the findings differ [21]. Tian et al. [30] found that during the COVID-19 pandemic, when preventive medicine major graduates faced a job crisis, workload became one of the least important factors. The urgency and pressure caused by the pandemic may have pushed graduates to prioritize job stability and safety. Overall, workload plays a critical role in whether TCM clinical graduates continue in their chosen field. Therefore, to attract and retain top TCM talent, policymakers should consider offering flexible work arrangements, balanced workloads, and policies that promote work-life balance. These measures can enhance satisfaction and support sustainable long-term career development.

Work environment: a notable factor in the employment choices of clinical TCM graduates

The work environment is another crucial non-economic factor, ranking second among all non-economic attributes. Scenario analysis revealed that improving the work environment from poor to average alone could attract nearly 72% of graduates. This finding aligns with preference studies of graduates from other health-related professions [19,20,21, 26]. A comfortable work environment not only enhances job satisfaction and personal well-being but also improves service quality and reduces the occurrence of medical errors and doctor-patient conflicts [37]. A study in Kenya also demonstrated that improving hospital infrastructure and the availability of resources can increase physician retention rates [39]. Additionally, in this study, the work environment encompasses relationships with superiors and colleagues. Previous research has shown that good colleague relationships positively influence the recruitment and retention of medical graduates [20, 26]. Therefore, to reduce the attrition of TCM graduates, the government should prioritize improving the work environment. This includes providing training and promotion opportunities, establishing supportive mentorship programs, improving workplace facilities and resources, and fostering positive teamwork and communication. Such measures can help attract and retain TCM professionals, boost motivation, and enhance both job commitment and satisfaction.

Housing security: moderate impact on the employment choices of clinical TCM graduates

Housing security is also an important factor in the employment decisions of TCM graduates. This study, which shows that graduates are more inclined to choose jobs that offer housing benefits, is consistent with research on medical graduates in several other countries [39,40,41]. In recent years, with the rapid development of the housing market in Chinese cities, home ownership has become a heavy burden for many families [24]. As a result, graduates tend to have stable housing to meet their needs for stability and financial security. Notably, male graduates place greater importance on housing security than female graduates. Subgroup analysis reveals that men are willing to forgo 2,051 CNY in monthly income for a job that offers housing, compared to 1,327 CNY for women. This may be linked to traditional Chinese cultural norms, where housing is viewed as a prerequisite for marriage, and men often bear more of the responsibility for purchasing a house [42]. Thus, men tend to prioritize jobs with housing benefits when considering employment options. In summary, providing housing support can effectively reduce the attrition rate of TCM graduates. Governments can address the housing needs of TCM graduates through various means, such as partnership programs, housing subsidies, or offering affordable rental housing.

Promotion opportunity: limited influence on the employment choices of clinical TCM graduates

Promotion opportunities are also a key consideration for TCM graduates when choosing jobs [26, 39,40,41]. This finding also aligns with other studies, which indicate that graduates are willing to sacrifice approximately 887 CNY of monthly income for jobs that offer greater promotion potential. This demonstrates their desire for growth and advancement opportunities to achieve their career goals. In the TCM field, the tradition of mentorship is particularly important. Learning from and working under a renowned master not only allows graduates to gain extensive clinical experience and refine their medical skills [43], but it also represents an intangible form of wealth. Compared to a singular focus on career advancement, TCM graduates may place greater value on this holistic growth and learning, which is often more appealing and impactful in the long term. Thus, to reduce graduate attrition, it is crucial not only to offer solid promotion opportunities but also to develop and promote mentorship-based education in TCM [44]. The government could support a talent development scheme that combines mentorship with university education, providing regular training and advanced study programs. Encouraging graduates to participate in research and academic exchanges, offering opportunities for promotion alongside the chance to learn under esteemed mentors, would stimulate motivation, foster a sense of belonging, and ensure the sustainability of career development for TCM graduates.

Work location: relatively low influence on the employment choices of clinical TCM graduates

Work location has the least influence on the employment preferences of TCM graduates. The study found that, compared to first-tier cities, graduates are more inclined to choose jobs in third-tier cities. However, research by Liu et al. [23] indicates that for PhD graduates major in Social Medicine and Healthcare Management, work location is the most important factor, with a preference for working in first-tier cities. This difference may stem from varying professional and educational development needs. For PhD graduates in Social Medicine and Healthcare Management, first-tier cities offer better career prospects and educational opportunities. In contrast, under the policy guidance [45], TCM, due to its unique strengths, is increasingly utilized in community healthcare services, particularly in chronic disease management and preventive care. This has elevated the role of TCM in community health services, creating more positions for TCM graduates at the community-level healthcare centers. Additionally, third-tier cities generally have lower economic and work-related pressures compared to first-tier cities, making them more attractive to graduates [46]. To encourage more graduates to work at the community level, the government and healthcare authorities should increase support for community healthcare institutions, improve working conditions, and offer better salaries and development opportunities.

Strengths and Limitations

This research introduces an innovative approach to explore the reasons behind the attrition of clinical Traditional Chinese Medicine graduates by applying a DCE. First, the study has a broad sample coverage, employing stratified sampling to include clinical graduates from most TCM universities across China. This ensures the sample’s representativeness and provides a more accurate reflection of TCM graduates’ job preferences. More importantly, this is the first study to apply the DCE method to analyze the issue of TCM graduate attrition, filling a gap in this area of research and providing a novel framework and empirical support for addressing the training and attrition of healthcare professionals.

However, there are certain limitations to this study. First, the DCE method is typically based on hypothetical scenarios, which may not fully cover the complexities graduates face in real-world work environments, potentially affecting the generalizability of the results. Second, the study did not include graduates from junior colleges and specialized vocational institutions, meaning the results may not fully represent the employment preferences of those with different educational backgrounds. Additionally, the attributes and levels used in the DCE are limited, so the study may not encompass all factors influencing employment choices, which could impact the comprehensiveness and validity of the findings. Future research could enhance the policy implications of such studies by including more diverse samples from various institutions and optimizing attribute design to more comprehensively uncover the deeper reasons for TCM graduate attrition. This would provide empirical evidence for relevant authorities to support the retention and development of TCM professionals.

Conclusion

To address the aforementioned issues, it is recommended to implement a combination of measures to improve the employment environment for clinical TCM graduates. These include offering positions with lighter workloads, better working conditions, housing security, and promotion opportunities. Moreover, employment opportunities and potential in first-tier cities should also be emphasized to attract and retain TCM graduates. By holistically considering both economic and non-economic factors above, the employment stability of TCM graduates can be ensured, which will facilitate the preservation of TCM knowledge and skills and lay a solid foundation for the future development of the TCM industry.