Exploring long-term retention intention and influencing factors among rural-oriented medical students in primary care in Guangxi, China: a qualitative study

Abstract

Background

Retaining qualified health professionals in remote rural areas remains a persistent global challenge. China’s rural-oriented medical students programme(ROMSP) was established to address shortages of health professionals in township health centres. However, emerging evidence suggests considerable turnover intention among participants after completion of their mandatory service period This study aimed to explore long-term retention intentions and the factors influencing retention among rural-oriented medical students working in primary care in Guangxi, China.

Methods

This exploratory qualitative study conducted semi-structured interviews with 26 rural-oriented medical students who had worked in their designated township health centres for between one and three years. The interview guide was developed based on a multidimensional framework of person–environment fit theory. Thematic analysis was used to code, analyse, and interpret the findings according to the theory’s five dimensions.

Results

Long-term retention intentions among rural-oriented medical students in Guangxi appeared limited, with only 23.1% (6/26) of participants expressing willingness to remain in their positions beyond the mandatory service period. Guided by the five dimensions of person–environment fit theory, leadership support within person–person fit emerged as a prominent factor associated with differences between participants intending to stay and those intending to leave. The person–job fit encompassed the broadest range of factors, with work–family balance identified as a facilitator. The influence of person–vocation fit was reflected through its interactions with other dimensions, whereas person–organisation fit and person–group fit appeared to play comparatively smaller roles.

Conclusions

To enhance long-term retention among rural-oriented medical students in primary care, interventions should focus on strengthening leadership capacity among township health centre directors, promoting work–family balance, and improving career mobility mechanisms after the mandatory service period. This study provides empirical support for applying person–environment fit theory to contract-based health workforce programmes and offers transferable insights for other countries facing similar rural health workforce challenges.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12913-026-15642-4.

Introduction

Retaining high-quality health professionals in remote rural areas remains a longstanding global challenge [1–2]. In China, township health centres face persistent challenges in recruiting, attracting, and retaining health professionals, particularly in western rural regions [3]. For example, in 2024, the proportion of licensed (assistant) physicians among health technicians in Guangxi township health centres (30.08%) was lower than the national average (40.72%), and the density of licensed (assistant) physicians per 1,000 rural residents (1.15) was below the national level (1.24) [4–5].

Evidence from the WHO [6–7] and countries such as the United States [8–10], Japan [11–12], and Thailand [13–15] suggests that targeted medical training programmes can contribute to alleviating rural health workforce shortages. Building on these international experiences, China launched the ROMSP in 2010. The programme recruits students with rural backgrounds who sign service agreements with county health bureaus. Participants receive free tuition, accommodation, and monthly living allowances. After completing five years of medical education, participants undertake three years of standardised residency training followed by a mandatory three-year service period at designated township health centres [16–17]. By 2025, the programme had enrolled approximately 89,000 rural-oriented medical students, largely fulfilling its policy goal of “training students from rural areas and returning them to serve those counties " [18].

Guangxi implemented the national ROMSP in 2010. By 2024, the programme had assigned 4,400 qualified graduates to township health centres, with a contract fulfilment rate of 97.2%; among them, 950 had completed their mandatory service period [19–20]. However, surveys have reported substantial turnover intention and low job satisfaction among these graduates after completion of their service obligations in Guangxi [21]. Given that turnover intention is an important predictor of actual workforce departure [22–23], understanding long-term retention intentions and their influencing factors before completion of the mandatory service period is essential for improving retention and strengthening the primary care workforce.

This study is grounded in person–environment fit theory, which proposes that individuals’ vocational choices are influenced by the degree of congruence between personal characteristics and their work environments. A strong person–environment fit can facilitate employee attraction and retention, improve organisational effectiveness, and support sustained performance [24]. Although previous studies have applied this theory to explain job dissatisfaction and turnover intention among healthcare workers [25–26], limited research has examined rural-oriented medical students working in impoverished and remote rural areas.

Accordingly, this study aimed to: (1) explore long-term retention intentions among rural-oriented medical students working in township health centres in Guangxi; (2) identify and analyse factors influencing retention using person–environment fit theory as an analytical framework; and (3) develop evidence-based policy recommendations to optimise rural-oriented medical student training programmes and support retention of high-quality primary care professionals in underserved rural areas.

Methods

Study setting and operational definitions

This study was conducted in the Guangxi Zhuang Autonomous Region of China. To ensure conceptual clarity and operational clarity, the following key terms were defined for this study.

Rural

According to the State Council Provisions on the Statistical Division of Urban and Rural Areas (Guo Han [2008] No. 60), “rural” was defined as areas outside urban administrative districts. In this study, “rural” referred to the locations of township health centres and their corresponding service catchment areas.

Leadership

In this study, “leadership” referred specifically to the managerial behaviours and support provided by township health centre directors (or responsible vice directors) for rural-oriented medical students. This construct comprised four dimensions: recognition of value, job allocation, career development support, and humanistic care.

Bianzhi

In the Chinese public health system, “bianzhi” refers to the authorized staffing positions approved by government authorities, representing the officially sanctioned number of posts and personnel allocations for a given health institution.

Study design

This exploratory qualitative study used semi-structured interviews with 26 rural-oriented medical students who had worked in their designated township health centres for between one and three years. The study aimed to gain an in-depth understanding of participants’ long-term retention intentions and the factors influencing retention. The study was reported according to the Standards for Reporting Qualitative Research (SRQR).

Theoretical framework

Person–environment fit theory can be traced back to Parsons’ person–job fit theory proposed in 1909 [27] and was subsequently expanded by later scholars. In 2006, Jansen et al. [28] developed a multidimensional model of person–environment fit comprising five dimensions: person-vocation fit(PV), person-job fit༈PJ༉,person-organisation fit༈PO༉,person-group fit༈PG༉,person-person fit༈PP༉(Table 1). A key contribution of this model is the integration of these five dimensions with different career development stages—pre-entry, job search, vocational choice, socialisation, and long-term retention—highlighting that different dimensions may play distinct roles across stages [24] (Fig. 1). During the long-term retention stage, the combined effects of these five dimensions are considered important predictors of outcomes such as job satisfaction, organisational commitment, and turnover intention. This theoretical perspective aligns closely with the present study’s focus on long-term retention intentions among rural-oriented medical students.

Table 1.

Multidimensional model of person-environment fit theory

Dimension	Definition	Specific indicators
Person–Vocation Fit (PV)	Refers to the congruence between an individual and the characteristics of a chosen profession.	Alignment between personal interests, personality traits, and career aspirations and the characteristics, requirements, and social expectations of the profession, focuses on individual–industry congruence.
Person–Job Fit (PJ)	Refers to the compatibility between individual abilities and job demands, as well as between individual needs and job attributes.	Needs–supplies perspective: adequacy of remuneration and benefits, bianzhi, opportunities for promotion, training and continuing education, working conditions, and organisational support in meeting individual needs. Demands–abilities perspective: alignment between individual knowledge, skills, competencies, and job requirements.
Person–Organisation Fit (PO)	Refers to the compatibility between individuals and organisations share similar fundamental characteristics, or the extent to which each fulfils the needs of the other.	Consistency between personal values and the organisation’s values, mission, and organisational culture.
Person–Group Fit (PG)	Refers to the compatibility between individuals and their work teams in terms of skills, interpersonal relationships.	Complementarity of professional skills within the team; compatibility of communication style, work style, personality, and teamwork.
Person–Person Fit (PP)	Refers to the compatibility between an individual and other key individuals in the workplace.	Influence of leaders and patients on personal work experience, career development, access to resources, and emotional well-being.

Source: Jansen, KJ, and Brown, A K.Toward a multidimensional theory of person一environment fit[J]. Journal of Managerial Issues, 2006,18:193–212

Fig. 1.

A dynamic model of person-environment fit based on the vocational development process. Source: Jansen, KJ, and Brown, A K. Toward a multidimensional theory of person一environment fit [J]. Journal of Managerial Issues, 2006,18:193–212

Recruitment

Recruitment announcements were distributed through graduate liaison groups of three medical universities in Guangxi (Guangxi Medical University, Guangxi University of Chinese Medicine, and Youjiang Medical University for Nationalities). These groups included rural-oriented medical graduates from across Guangxi and provided access to the target population of this study.

The inclusion criteria were: having returned to their designated township health centres in Guangxi and being within the three-year mandatory service period; having completed at least one year of full-time service at the designated township health centre at the time of interview; and voluntarily agreeing to participate and providing written informed consent. The exclusion criteria were: having left the designated township health centre before completion of the mandatory service period; being on extended leave (e.g., maternity leave or long-term sick leave) during the study period; or being unable to complete the interview. These criteria ensured that participants had sufficient workplace experience to reflect on their work environment, job demands, organisational culture, and other relevant aspects.

Purposive sampling was initially employed to ensure eligibility and achieve variation across geographical regions (northern, central, and southern Guangxi), resulting in the recruitment of 21 participants. However, this approach may have resulted in greater representation of individuals who were easier to contact through alumni networks. Snowball sampling was subsequently used, with initial participants invited to recommend eligible peers working in other township health centres, particularly those from different types of settings (e.g., township health centres near county seats vs. those in remote mountainous areas) and those approaching completion of their service period. This approach resulted in the recruitment of five additional participants and enhanced sample diversity in terms of geographic coverage and institutional characteristics.

Data collection

Data were collected between October and December 2025. The interview guide was developed based on person–environment fit theory, and all interviews followed the same semi-structured guide (Additional file 1). Individual interviews were conducted in Chinese via Tencent Meeting by a research team comprising one higher medical education administrator, one qualitative researcher, and two master’s students. The higher medical education administrator was involved in curriculum development for rural-oriented medical students but had never served as a course instructor, examiner, or evaluator for any participant, and had no direct supervisory or professional relationship with them. Before each interview, participants received a study introduction and informed consent form (Additional files 2–3), and an interview time was arranged. Interviews lasted 60–90 min, with a mean duration of approximately 75 min. With participants’ consent, interviews were audio-recorded through the videoconferencing platform and transcribed using speech-to-text software.Recruitment continued until code saturation was achieved, defined as the point at which no new codes emerged from the final two interviews.

Data analysis

Data analysis began after the first interview and continued concurrently with data collection. Transcripts were checked by two master’s students and reviewed by a researcher who participated in all interviews to ensure accuracy. All transcribed content was anonymised.

Thematic analysis was conducted using the multidimensional model of person–environment fit theory as an analytical framework and adopting a combined deductive–inductive coding approach. Two researchers independently coded the transcripts using MAXQDA 2020 (VERBI Software GmbH, Bismarckstrasse 10–12, 10625 Berlin, Germany) through two analytical stages. First, inductive coding was performed through line-by-line examination to generate initial codes. Second, codes were mapped onto the five theoretical dimensions, and unmatched codes were discussed among the research team.

Throughout the coding process, disagreements were discussed between researchers, and weekly team meetings were held to review emerging codes and interpretative issues. Themes were developed through iterative processes of interpretation, abstraction, and refinement until a coherent thematic structure was established. Final codes and themes were agreed upon by the research team, and representative quotations were translated into English for reporting.

Results

Participant demographic characteristics

A total of 26 participants were included in this study, including 12 males and 14 females. Table 2 presents the demographic characteristics of participants.

Table 2.

Participant demographics characteristics

Demographic variables	Number (n)	Percentage (%)
Gender
Male	12	46.2
Female	14	53.8
Ethnicity
Han	14	53.8
Zhuang	10	38.5
Others	2	7.7
Highest educational qualification
Bachelor’s degree	10	38.5
In-service master’s degree	16	61.5
Professional title
Junior	8	30.8
Intermediate	18	69.2
Marital status
Unmarried	9	34.6
Married	17	65.4
Department
General practice	12	46.2
Inpatient ward	4	15.4
Outpatient and emergency	3	11.5
Internal medicine/Obstetrics and gynaecology/Paediatrics	4	15.4
Chronic disease medicine	1	3.8
Administrative office	2	7.7
Administrative position
Vice director	4	15.4
Section-level	12	46.2
None	10	38.5
Region
Northern Guangxi	8	30.8
Central Guangxi	13	50.0
Southern Guangxi	5	19.2
Years of service
Year 2	9	34.6
Year 3	17	65.4
Annual income (RMB)
< 50,000	2	7.7
50,000–69,999	12	46.2
70,000–89,999	10	38.5
≥ 90,000	2	7.7

Note: Northern Guangxi includes Guilin, Liuzhou and Hechi; central Guangxi includes Baise, Nanning, Laibin, Guigang, Wuzhou and Hezhou; southern Guangxi includes Qinzhou, Fangchenggang, Chongzuo, Beihai and Yulin

Long-term retention intention

Among the 26 participants, six (23.1%) expressed an intention to remain in township health centres beyond the mandatory service period. Among these participants, five were from northern Guangxi. Twenty participants (76.9%) indicated an intention to leave their current township health centres after completing their mandatory service period. Their preferred destinations included community health service centres (2/20), county-level hospitals (8/20), and municipal-level hospitals (10/20). No participants reported intentions to leave the healthcare profession or discontinue employment to pursue further education (Fig. 2).

Fig. 2.

Long-term retention intention of participants

Influencing factors

Guided by the five dimensions of person–environment fit theory, 15 themes were identified.

Person-Vocation Fit(PV)

Theme 1:Vocational interest

Among the 26 participants, seventeen reported that their initial motivation for studying medicine was to become clinicians, save lives, and achieve personal fulfilment, whereas nine chose the programme primarily because of financial constraints and the expectation of stable employment, which was consistent with the policy objectives of the programme. Notably, among the 20 participants intending to leave, seven reported that they had been assigned by their leaders to administrative departments or public health sections after returning to primary care, limiting opportunities to engage in clinical practice.

My clinical knowledge and skills have become rusty, and I feel very lost. My professional interest lies in clinical practice, but I have been reassigned by my leader to the hospital administration office. A significant amount of my time and energy is now spent on completing forms and writing reports, leading to a decline in my clinical knowledge and skills. I am feeling deeply uncertain about my career path. (Participant 6)

Theme 2:Personal traits

Among the 20 participants who intended to leave, two participants valued personal autonomy, prioritised work–life balance, and preferred a less demanding work environment. However, physician shortages in some township health centres resulted in long working hours, heavy workloads, and multiple role responsibilities. Consequently, these participants viewed community health service centres as a more attractive alternative because they offered lighter workloads and fewer or no night shifts.

I need to run outpatient clinics, manage inpatients, conduct skills training, deliver health education, lead teaching ward rounds, organise health check-ups for the elderly, regularly visit villages for follow-ups, and prepare materials for various inspections. I often start work at five or six in the morning and work overtime until midnight, leaving no time for personal life. Moving to a community health service centre would be easier and simpler. (Participant 10)

Furthermore, three of participants intending to leave expressed a strong preference for specialised clinical practice, enjoyed working in high-intensity and high-acuity environments, and derived professional fulfilment from managing acute and critically ill patients. However, general practice training emphasises broad clinical competence rather than specialty-specific expertise, and routine primary care mainly involves the management of common illnesses and chronic diseases, whereas patients requiring acute or critical care are typically referred to higher-level hospitals.

I prefer more challenging and stimulating emergency and acute care work, rather than sitting in the outpatient clinic every day prescribing medication for colds, fevers, or chronic diseases. (Participant 11)

Aside from obstetrics and gynaecology and paediatrics, I see everything. On the surface, I can handle a bit of everything, but I am not proficient in any single area. I hope to pursue in-depth learning in a specific field, such as endocrinology, which greatly interests me. (Participant 2)

Person–Job Fit (PJ)

Theme 3:Remuneration and benefits

Among the 20 participants intending to leave, 17 reported that perceived salary inadequacy contributed to their intention to leave. Among these participants, 14 described receiving mainly basic salaries with limited performance-based incentives despite substantial workloads, and perceived their income as considerably lower than that of peers in county-level hospitals or comparable positions. In contrast, all six participants intending to remain long-term expressed satisfaction with their remuneration and benefits. Their reported annual incomes ranged from RMB 70,000–80,000 to more than RMB 100,000, which they perceived as relatively favourable within their local context.

The workload is heavy and intense. Last year, there were six or seven months without performance bonuses, making it difficult even to cover basic living expenses.(Participant 20)

An annual income of RMB 80,000 is commensurate with my current workload, and the gap with county-level hospitals is not large. It is considered relatively favourable locally, and I can live comfortably.(Participant 21)

Theme 4:Bianzhi

All 26 participants had held an authorized staffing position (bianzhi), which provided a sense of job security and was perceived to facilitate long-term retention in remote township health centres.

County-level hospitals do not offer Bianzhi(authorized staffing positions) whereas my current position does. That makes me more willing to continue working here. (Participant 24)

Theme 5:Position and professional title

All six participants intending to remain long-term held section-level or higher positions and reported access to professional development opportunities. Among participants intending to leave, 14 reported limited promotion opportunities and unclear career development pathways within primary care institutions. Some participants perceived that promotion to management positions was influenced by personal relationships and that advancement through senior professional title examinations was highly competitive.

I think Liuzhou’s policy is very supportive. The Health Bureau offers targeted training and opportunities for internal promotion within the medical alliance(an integrated county-level healthcare network).Many people have already been promoted to vice director or even director of township health centres. With opportunities like these, I can continue to develop professionally. (Participant 21)

I have never had the opportunity to be appointed as a vice director. Here, what matters is not your professional skills or competence, but your relationship with the leadership. (Participant 20)

Theme 6:Training and continuing education

Nineteen participants reported limited opportunities for external training and continuing education, which they perceived as restricting clinical skill development and career advancement. Participants expressed a desire for further training in managing acute and critical conditions in resource-constrained settings, appropriate traditional Chinese medicine techniques, and artificial intelligence applications.

I think training opportunities are very limited. The training I need—such as communication skills and clinical decision-making about medication use—is simply not available. (Participant 5)

Theme 7:Working conditions

Participants reported that diagnostic equipment, medication supplies, and information technology generally met routine needs; however, gaps remained between available resources and clinical requirements, limiting opportunities to apply their skills. Medication shortages and limited drug availability were frequently identified as major constraints on clinical practice, with 23 participants highlighting these issues.

Medications are frequently out of stock. For example, for the past two months, the only antibiotics available have been basic drugs such as amoxicillin capsules and first-generation cephalosporins, making it difficult to provide the most appropriate treatment. (Participant 14)

Theme 8:Work–family balance

With increasing age and family responsibilities, 15 participants expressed a preference for workplaces closer to home to facilitate caregiving responsibilities and maintain family relationships. Among participants intending to remain, five reported that their work location met their family needs, whereas 13 participants intending to leave described geographical distance from home as a challenge. One participant intending to remain highlighted adequate rest time, manageable workload, and lower stress as facilitating work–family balance.

I’m from this area. The health centre is close to my home, so I can visit my family frequently and easily care for my elderly parents and children. I feel that I belong here. Otherwise, I would probably leave as well. (Participant 21)

Theme 9:Competence

Participants generally perceived that their knowledge, skills, and abilities were adequate for their roles; however, limitations in equipment and medication supplies restricted their ability to apply these competencies. Eight participants felt that their clinical knowledge had not kept pace with evolving treatment protocols and technological developments after working in township health centres. They also identified gaps in leadership, management, and communication skills.

The range of conditions that can be managed in primary care is limited. Because I rarely have the opportunity to use many of my clinical skills, I feel that some of them have become rusty. Training opportunities are also limited, so I have not kept up with new medications and advances in clinical practice. (Participant 19)

Throughout my training, relatively little attention was paid to communication skills, such as discussing medication adherence, communicating with terminally ill patients and their families, and interacting with supervisors and colleagues. (Participant 5)

Person–Organisation Fit (PO)

Theme 10:Value orientation

Eighteen participants perceived alignment between their personal values and those of their township health centres, including commitment to patient-centred care, improving healthcare quality, and serving the health needs of local communities. Among participants intending to leave, two perceived that organisational priorities emphasised revenue generation, including increased testing, hospital admissions, and prescriptions, which conflicted with their professional values.

My original intention was to provide effective treatment while reducing patients’ financial burden. However, my leader believes that increasing admissions and hospital revenue is what matters most. I feel very conflicted and distressed. (Participant 20)

Theme 11:Organisational culture

Twenty participants described their township health centres as having an organisational culture characterised by accessibility, collaboration, and innovation. Among participants intending to leave, three described organisational barriers including inefficient management, excessive meetings, limited opportunities for innovation, slow adoption of new technologies or services, and inflexible service models.

It is very comfortable working here. The team emphasises collaboration, everyone cares about patients, and we strive to improve the quality of diagnosis and treatment despite limited resources. (Participant 8)

We mainly provide routine services. The leadership appears reluctant to develop new services, such as ambulatory blood pressure monitoring or pulmonary function testing. (Participant 5)

Person–Group Fit (PG)

Theme 12:Skill complementarity

Most participants perceived that they contributed substantially to their teams, such as standardising diagnosis and treatment processes, promoting rational medication use, providing skills training, introducing new technologies and medications, improving medical documentation, optimising clinical management, and facilitating physician–nurse collaboration. These contributions helped improve healthcare quality, enhanced service delivery, reduced unnecessary referrals, and increased patient recognition. Among participants intending to leave, four felt they had limited opportunities to make similar contributions within their teams.

After returning to work at the township health centre, I restructured the duty roster and standardised protocols for emergency resuscitation and trauma management. I also promoted the appropriate use of antibiotics and standardised medication management for chronic conditions such as hypertension and diabetes. In addition, I improved medical record documentation and fostered a more collaborative working environment among the staff. In the county-wide healthcare quality assessment, our health centre rose from near the bottom of the rankings in previous years to second place last year. (Participant 18)

Theme 13:Team relationships

Most participants described positive relationships characterised by communication, collaboration, and mutual support among colleagues. However, five participants reported strained team relationships characterised by interpersonal conflicts, perceived exclusion, limited interaction, and inadequate communication.

I became a leader at a young age, and they all ganged up against me. They neither understand nor support the introduction of new treatments and medications. They delay submitting medical records, and there are many cliques. They feel that I have increased their workload, and many things are difficult to implement. (Participant 17)

Person–Person Fit (PP)

Theme 14:Leader–subordinate relationship

All six participants intending to remain long-term in primary care described their leaders as approachable, democratic, and open-minded, with leaders who accepted suggestions, allocated positions according to individual development needs, delegated responsibilities, and provided promotion and professional development opportunities. All six participants highlighted leadership recognition as an important factor supporting retention. Among participants intending to leave, eight perceived insufficient leadership recognition, inappropriate job placement, limited responsiveness to suggestions, or even harboured prejudice and friction. These participants commonly ranked insufficient leadership recognition as the primary reason for their intention to leave. Ten participants reported limited communication with their leaders and perceived insufficient humanistic care and career development support. This “indifferent” leadership style similarly increased participants’ intention to leave.

I did not feel valued after returning. Suggestions to introduce data monitoring instruments or medications were rarely adopted. Over time, I stopped making suggestions. My leader once told me, ‘A small hospital does not need highly educated personnel.’ This made me feel that I had no reason to stay. (Participant 20)

I have almost no communication with my leader. They rarely pay attention to me, so I feel there is no point in trying to build a relationship. (Participant 14)

Theme 15:Patient–physician relationship

Most participants reported establishing positive relationships with patients, and patient trust enhanced their sense of professional fulfilment. Six participants reported that their young age and limited professional seniority affected patients’ perceptions of their competence, sometimes resulting in reduced cooperation and adherence. Participants described encountering complaints and occasional verbal conflicts when patients’ conditions did not improve, when treatment decisions were misunderstood, or when expectations could not be met.

Sometimes patients are not here for a consultation but just to get medication. However, the hospital does not have the medications they need. When their objective is not achieved, they inexplicably lose their temper and scold us. (Participant 3)

Integrated model of influencing factors

By synthesising findings across the five dimensions of person–environment fit, this study developed an integrated model illustrating factors influencing long-term retention intentions among rural-oriented medical students (Fig. 3). In this model, factors were analytically classified into five categories according to their perceived roles in shaping retention intentions: critical factors (leadership support), facilitating factors (work–family balance), contextual factors (value orientation and organisational culture), stage-dependent factors (Bianzhi, person-vocation fit, position and professional title, skill complementarity, remuneration satisfaction, utilisation of competence, and development opportunities), and secondary factors (training opportunities, working conditions, team relationships, and patient–physician relationships) (Fig. 3).

Fig. 3.

An integrated model of factors influencing long-term retention intention among rural-oriented medical students based on person–environment fit theory. Note: PV = person–vocation fit; PJ = person–job fit; PO = person–organisation fit; PG = person–group fit; PP = person–person fit. Factors in the figure are organised chronologically according to vocational stage (early service → mid-service → late service) and are presented separately for the “retention” and “turnover” directions. Leadership support was identified as a critical factor throughout the entire process; work–family balance served as a facilitating factor. “Full utilisation of competence” and " lack of opportunities for competence utilisation " involved intersections across PJ, PO, and PP. Secondary factors are listed separately in the figure, given their relatively limited influence

The five dimensions demonstrated distinct patterns and different levels of relevance within the model. The PJ dimension included the widest range of factors, spanning participants’ experiences from initial employment through later stages of service. However, factors within the PJ dimension appeared to differ in importance depending on participants’ career stages and the challenges they encountered. The PP dimension included leadership support and patient–physician relationships. Leadership support emerged as particularly salient, suggesting that interpersonal dynamics may influence retention beyond formal organisational arrangements. The PO dimension was reflected through value orientation and organisational culture. When personal values conflicted with organisational culture, participants described feelings of reduced belonging and motivation. Such value incongruence appeared unlikely to independently determine turnover intention but could intensify departure considerations when combined with mismatches in other dimensions. The PG dimension, encompassing team relationships and skill complementarity, appeared to function mainly as a secondary factor or stage-dependent factor. Its relatively limited influence may reflect the relatively stable team structures commonly found in township health centres in China. The PV dimension, including vocational interest and personal traits, appeared to influence retention through its interactions with PJ and PP, reflecting the dynamic interrelationships among different forms of fit.

Furthermore, “utilisation of competence” and “lack of opportunities for competence utilisation " emerged inductively from the interview data. Although closely related to the “competence” theme within the PJ dimension, these factors also reflected organisational value alignment (PO) and leadership support for vocational development (PP). Therefore, these factors were conceptualised as cross-dimensional constructs (PJ/PO/PP), representing interconnected demands related to job placement, value alignment, and leadership support.

This integrated model illustrates the hierarchical roles and interrelationships among factors shaping long-term retention intentions and provides a theoretical basis for developing targeted policy interventions.

Discussion

This study explored long-term retention intention and its influencing factors among rural-oriented medical students in Guangxi. The findings indicated that only a minority of participants (23.1%) intended to remain in their designated township health centres after completing the mandatory service period, whereas most participants (76.9%) preferred transferring to county-level or urban healthcare institutions; none reported an intention to leave healthcare altogether. This pattern of mobility, characterised by movement towards higher-level healthcare institutions, is consistent with the global trend of rural health professionals concentrating in urban areas [29].

This study applied the multidimensional person–environment fit model to systematically identify 15 factors influencing long-term retention intention among rural-oriented medical students in Guangxi. These factors covered five dimensions: PV, PJ, PO, PG, PP, reflecting the model’s emphasis on the integrated effects of multidimensional person–environment fit in supporting long-term retention [24].

Among the five dimensions of person–environment fit, leadership-related factors within the person–person fit dimension emerged as the most salient factors distinguishing participants with retention and turnover intentions. Differences between the two groups were particularly evident in perceptions of leadership recognition, exceeding the difference observed for remuneration satisfaction. Although 85% (17/20) of participants intending to leave cited salary concerns, consistent with findings from lower-middle-income countries such as the Philippines and Nepal [30–31], the present study suggests that remuneration perceptions were closely intertwined with leadership practices. When leaders recognised the contributions of rural-oriented medical students, provided appropriate clinical roles, and supported career development, participants perceived greater opportunities for professional growth and competence utilisation, which appeared to reduce the influence of financial concerns. Conversely, when participants perceived limited recognition, inappropriate job placement, and restricted promotion opportunities, remuneration concerns became more prominent reasons for considering departure. This finding supports the WHO’s assertion that effective retention strategies in remote rural areas require strong management and leadership capacity, particularly at the institutional level [32]. Similar findings have been reported among rural health workers in Vietnam [33] and sub-Saharan Africa [34], where managerial recognition and supportive leadership were identified as important factors affecting attraction and retention. However, previous studies have generally examined managerial support as one component of the broader work environment [34–35]. The present study extends this understanding by highlighting leadership support as a central mechanism within contract-bound health workforce programmes, where mandatory service obligations limit individual mobility. Therefore, improving leadership practices within township health centers may represent a higher-priority intervention than financial incentives alone. Furthermore, because leaders shape organisational culture and professional values, supportive leadership may strengthen belonging and motivation, whereas perceived value incongruence and limited recognition may accelerate attrition. This finding is consistent with evidence from the United States showing that institutional leadership support significantly shapes rural physicians’ perceptions of workplace culture [36].

This study also identified work–family balance as a facilitating factor influencing long-term retention intention. As participants aged and family responsibilities increased, geographic proximity between workplaces and homes, as well as the ability to care for family members, became important considerations in retention decisions. Evidence from developed countries, including Canada, the United Kingdom, Australia, and the United States [37–41], indicates that family-related factors—such as family safety and wellbeing, children’s educational opportunities, and partners’ employment prospects—substantially influence physicians’ decisions to remain in rural areas. Similarly, studies from lower-middle-income countries have highlighted the importance of family proximity and support in sustaining rural health workforce retention [42–43]. Therefore, future rural-oriented medical student programmes should incorporate family-oriented policies, including consideration of family-related factors such as a spouse’s workplace and children’s education.

In contrast, factors such as training opportunities, working conditions, team relationships, and patient–physician relationships appeared less salient in participants’ immediate retention decisions. Although previous studies have identified these factors as important predictors of rural health worker turnover [44–45], participants in this study, who were bound by service contracts, placed greater emphasis on internal institutional factors such as leadership behaviour, career development opportunities, and job security. Nevertheless, over the longer term, limited training opportunities, constrained working conditions, and persistent difficulties in team and patient relationships may gradually undermine professional development and job satisfaction, representing potential longer-term retention risks.

Based on the findings of this study, several strategies may be considered to further optimise the rural-oriented medical student programme from three perspectives. First, during undergraduate medical education, greater emphasis could be placed on rural health service training, including increased rural health-related coursework and structured rotations in township health centres. These experiences may strengthen professional identity, enhance understanding of primary care practice, and support future retention. In parallel with ongoing integrated healthcare system reforms, a “county-level management with township-based employment” model could be explored, whereby rural-oriented medical students establish employment relationships with county-level hospitals that coordinate their deployment to township health centres. After completing the mandatory service period, opportunities for internal mobility within the county could be provided to improve career development prospects. Second, health administrative departments at different levels could prioritise leadership capacity development among township health centre directors. Selected directors could participate in targeted training programmes to strengthen their understanding of the programme objectives and the professional contributions of rural-oriented medical students. Retention and professional development of rural-oriented medical students could be incorporated into leadership evaluation systems, encouraging directors to assign appropriate clinical roles, support career advancement, and cultivate organisational environments conducive to retention. Third, allocation of service institutions could incorporate a “hometown-oriented placement” principle. For those serving away from their hometowns, health authorities in both regions could explore mobility mechanisms, such as reciprocal adjustment of admission quotas or inter-institutional transfers, allowing rural-oriented medical students to relocate closer to their hometowns or spouses when feasible. Additional supportive policies, including access to children’s education, could be considered to facilitate better integration of family and professional responsibilities.

This study has several limitations. First, the qualitative design, relatively small sample size, and purposive and snowball sampling approaches may have resulted in potential selection bias. Therefore, the findings aim to provide theoretical and analytical insights rather than statistical generalisability. Second, all participants were interviewed during their mandatory service period, and a gap may exist between retention intention and actual future employment behaviour. Future longitudinal studies are needed to examine whether expressed intentions translate into subsequent retention outcomes. Third, discussion of sensitive topics such as turnover intention may have introduced social desirability bias, despite efforts to minimise this risk through clear communication of research independence and data anonymisation procedures. Fourth, the study was conducted exclusively in Guangxi, and caution should be exercised when transferring the findings to other geographical and institutional contexts. Fifth, online interviews may have limited the researchers’ ability to capture non-verbal communication cues, potentially reducing the depth of interpretation for sensitive topics. Sixth, translation of interviews and quotations into English may have resulted in partial loss of contextual meaning. Despite these limitations, the in-depth interviews with 26 participants provide valuable insights into the multidimensional factors shaping long-term retention intention among rural-oriented medical students and offer a foundation for future research.

Conclusions

This study demonstrates that long-term retention intention among rural-oriented medical students is shaped by multiple dimensions of person–environment fit, with leadership support emerging as the most salient factor distinguishing those intending to remain from those intending to leave. Improving the retention of rural-oriented medical students requires more than contractual commitments or financial incentives; it requires supportive leadership, opportunities for career development, family-sensitive policies, and flexible post-service mobility mechanisms. These findings provide evidence to inform the optimisation of China’s rural-oriented medical student programme and offer transferable insights for countries implementing contract-based rural health workforce strategies.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ROMSs

Rural Oriented Medical Students

ROMSP

Rural oriented medical students programme

PE

fit Person-Environment Fit

PV

Person Vocation Fit

PJ

Person Job Fit

PO

Person Organization Fit

PG

Group Fit

PP

Person Person Fit

SRQR

Standards for Reporting Qualitative Research

Author contributions

Xin Zhang designed the study, collected, processed, and analyzed the data, and drafted and finalized the manuscript. Yue Zhao participated in the study design and data collection. Xiaodan Li, Siying Liang, Mingzhen Lin, Boyi He, Tao Cen, and Yanhu Liu were actively involved in data collection and analysis, as well as in the revision of the paper. Yanli Zuo supervised the study design and provided suggestions on the discussion and analytical framework. All authors read and approved the final manuscript.

Funding

This study was supported by the National Natural Science Foundation of China (Grant No. 72364004), the Guangxi Natural Science Foundation (Grant No. 2026GXNSFAA00641212) and the Project of Guangxi Philosophy and Social Science Research (Grant No.23FGL038). The Research Center for Health and Economic-Social Development, Guangxi University Key Research Base of Humanities and Social Sciences(Grant No.2026RWB02).

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to confidentiality agreements with participants, but are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval was granted by the Medical Ethics Committee of Guangxi Medical University (Grant No.2023-KY0101). All participants were informed of the study details and provided consent by checking the corresponding options on the informed consent form before the interview. This study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.