Dilemmas in the development of the general practitioner workforce in primary care: a qualitative study in Beijing, China

Abstract

Background

The general practitioner workforce (GPWF) is an important factor influencing primary care capacity. Many countries face GP shortages and substantial challenges in workforce development. However, the factors and mechanisms shaping GPWF development remain insufficiently understood.

Purpose

To explore key challenges affecting GPWF development and their interrelationships in the primary care system of Beijing, China.

Methods

Between October and December 2024, directors and GPs from 18 community health centres were recruited using purposive sampling. Data were collected through a focus group discussion and in-depth interviews; thematic analysis was conducted in line with Merriam and Tisdell’s approach. A causal loop diagram (CLD) was then developed to identify and visualise dynamic relationships among factors influencing the GPWF development.

Results

The participants comprised 10 directors and 8 GPs from urban and suburban areas, with a predominance of women and those holding senior professional titles. Participants had diverse educational and professional backgrounds. Three major themes were identified: (1) rising demand for GP services, (2) supply challenges in the GPWF, and (3) challenges to GPs’ professional sustainability. Across these themes, 12 subthemes and 10 key variables were identified. The CLD reflected participants’ perceptions of how these variables interacted and illustrated potential reinforcing interactions among them.

Conclusion

The development of the GPWF in Beijing is constrained by multiple interrelated challenges. Addressing these challenges entails coordinated policy actions to strengthen workforce planning and training, enhance job attractiveness, improve the alignment between training and practice, and reinforce professional support and continuing professional development, thereby promoting the sustainability of the GPWF.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study provides an in-depth qualitative analysis of key challenges in developing the general practitioner workforce (GPWF).

• The use of a causal loop diagram helps capture the systemic complexity of the GPWF and the interactions among multiple influencing factors.

• The findings offer policy-relevant insights into improving general practitioner (GP) recruitment, retention and workforce sustainability.

• The generalisability of the findings should be interpreted with caution, as the results are shaped by the specific institutional and policy context.

• The study reflects only the perspectives of GPs and community health centre directors and does not include views from other stakeholders.

Introduction

General practitioners (GPs) are the backbone of the primary care workforce, which influences primary care capacity and population health outcomes.^{1 2} In China, GPs mainly practise in community health centres (CHCs), where they provide both essential medical care and basic public health services. Although GPs are not yet established as mandatory first-contact providers, the ongoing health reform in China puts strong emphasis on building a first-contact primary care system with competent GPs and robust primary care teams.³ This reform aims to support health system transition, address the challenges of population ageing, improve health equity and align with the WHO’s sustainable development goals for health.⁴ Achieving these goals requires that the size and clinical competence of the general practitioner workforce (GPWF) meet growing healthcare demands and quality standards.

The government has implemented multiple initiatives to expand the GPWF, including standardised residency training (5-year undergraduate education and 3-year residency training), job-transfer training programmes (from other specialties to general practice), the Compulsory Service Programme (CSP), which provides tuition-free medical education in exchange for 6 years of service in rural or suburban areas, and the ‘3+2’ assistant GP training pathway (3-year junior college medical education plus 2-year GP training).⁵ These initiatives have substantially expanded the GPWF. Between 2015 and 2023, the number of registered GPs in China increased from 188 649 to 558 600, and the number of GPs per 10 000 population rose from 1.38 to 3.99.⁶ China has also set a policy target of reaching five GPs per 10 000 population by 2030.

Although the number of registered GPs has increased markedly, the GPWF actually engaged in primary care remains lower than the reported registration figure suggests, as some registered GPs who had completed job-transfer training programmes continue to work in hospitals rather than in primary care settings. There is still a gap between China and high-income countries in the number of GPs per 10 000 population, with corresponding figures of 8.1 in the UK, 11.5 in Switzerland and 14.6 in France in 2021.⁷ In addition, China’s primary care system still faces multifaceted workforce challenges, including an imbalanced distribution of the GPWF, lower educational attainment and professional titles among GPs than among other specialists and insufficient clinical training.⁸ These deficiencies constrain both the accessibility and quality of primary care.9,11

Addressing challenges in the primary care workforce is a pressing global health priority.¹² Despite widespread recognition of GP shortages as a major challenge, the international evidence base for comprehensive workforce development strategies remains underdeveloped. Existing research has predominantly focused on quantitative examinations of isolated workforce dimensions, including training programmes, attrition and professional satisfaction metrics.13,15 The GPWF development is shaped by complex and interdependent factors with multidimensional interactions that have not been fully elucidated. Qualitative research is well suited to exploring these complexities and providing in-depth insights into underlying mechanisms. Using this approach, this study examines the key challenges in the development of the GPWF in Beijing’s primary care context and explores their underlying causes and interrelationships to inform workforce planning and support the sustainability of the GPWF and primary care.

Methods

Study context and design

This qualitative study was conducted in Beijing, the capital of China. The city has strong financial support, favourable policies, a skilled workforce and technological capacity, making it a potential model for primary care development nationwide. Beijing covers 16 400 km² and comprises 16 districts, with a resident population of 21.858 million at the end of 2023.

Primary care development in Beijing began in the 1990s and has resulted in an extensive network primarily composed of CHCs. According to statistics from the Beijing Health Services Management and Guidance Center (BHSMGC), the network includes 311 government-run CHCs and 6528 registered GPs, as well as 704 individuals with GP training qualifications who are not formally registered as GPs.

Data were collected using a focus group discussion and in-depth individual interviews to enhance the richness of the findings.¹⁶ Prior to developing the interview guide, we reviewed existing literature on the primary care workforce and referred to two WHO frameworks: ‘Human Resources for Health Action Framework’ and the ‘Outline Concept for Linking Health Workforce Requirements and Supply Projections’, which outline key factors related to health workforce supply, demand and planning.¹⁷ The semi-structured interview guide was further refined through discussions with two GPs, two CHC directors and members of the research team. Separate guides were used for the focus group discussion and individual interviews (see online supplemental appendix 1).

Study population and sampling

Purposive sampling was used to recruit participants from two groups: eligible GPs were required to be frontline GPs with at least 1 year of experience in general practice and good communication skills, and to be willing to participate. Eligible CHC directors were required to hold managerial responsibility for the GPWF and have at least 3 years of experience in the recruitment and management of GPs.

Participants were recruited with support from key informants. BHSMGC staff and CHC directors disseminated recruitment information through institutional communication channels. The recruitment notice described the study purpose, interview procedures and confidentiality arrangements. The final sample included 8 frontline GPs and 10 CHC directors from 18 CHCs across eight districts in Beijing: four urban districts (Dongcheng, Xicheng, Chaoyang and Fengtai) and four suburban districts (Tongzhou, Daxing, Huairou and Yanqing). None of the participants were previously acquainted with the research team.

Researchers’ characteristics and reflexivity

Seven authors were involved in the interview process. TS and LL, affiliated with BHSMGC, had an indirect hierarchical relationship with the participants and were therefore involved only in distributing recruitment information. MT and NJ also assisted with this process. YC, GJ and WW were researchers from Capital Medical University with experience in qualitative research and qualifications in general practice. YC, GJ and WW had no hierarchical relationship with the participants and were responsible for participant selection, conducting the interviews, taking field notes and anonymising the data.

To minimise hierarchical influence and social desirability bias, recruitment materials emphasised voluntary participation, directors were not informed of the final list of GP participants, and interviews were conducted in private locations chosen by participants. Participants were informed of the researchers’ roles and study purpose prior to interviews. Pilot interviews were conducted within the research team to standardise the interview process, specifically to avoid leading questions and to ensure sufficient depth of data collection.

Data collection

Participants received study materials before the interviews. In October 2024, a 3-hour focus group discussion was conducted with ten CHC directors. Subsequently, between November and December 2024, eight in-depth interviews were conducted with GPs, each lasting 30–80 min.

All data collection sessions were conducted face-to-face by YC, with GJ and/or WW taking field notes. Interviews were audio-recorded with participants’ consent, and no participants withdrew. Focus group recruitment was guided by data richness,¹⁸ while individual interviews continued until thematic saturation was reached.¹⁹

Data analysis

Audio recordings were transcribed verbatim in Chinese by YC and anonymised during transcription. Transcripts were coded by YC and TS, with discrepancies resolved through discussion within the research team. The data were then analysed using thematic analysis²⁰ to identify themes. The analysis followed the six-phase framework described by Merriam and Tisdell,²¹ as follows: (1) repeated reading of interview transcripts and identification of meaning units guided by the interview guide; (2) systematic coding using NVivo 20, with initial coding conducted by YC and reviewed by GJ; (3) comparison and categorisation of codes by YC and GJ; (4) synthesis of themes through category integration, led by YC with input from the research team; (5) refinement of subthemes; and (6) consensus validation of the final themes through team discussion.

Data collection and analysis were conducted iteratively. Preliminary analysis of the focus group data informed refinement of the GP interview guide, while ongoing analysis of interview transcripts guided subsequent data collection. The identified themes were translated into English by YC and reviewed by GJ. Transcripts were not returned to participants for comment.

The study adhered to the Consolidated criteria for reporting qualitative research (COREQ) checklist (see online supplemental appendix 2 for details).²²

CLD construction

To explore the interrelationships among constraints affecting the development of the GPWF, a causal loop diagram (CLD) was constructed based on the findings thematic analysis, following established approaches for translating qualitative data into system dynamics modelling.²³ The CLD was constructed as follows.

First, statements containing causal expressions (eg, ‘leads to’, ‘increases’, ‘decreases’, or ‘because’) were extracted from the interview transcripts and coded as causal statements. Second, factors involved in causal relationships within and across themes were defined as key variables. Third, directional relationships between key variables were identified in causal statements and recorded as causal links (variable A → variable B). Fourth, causal links were integrated to identify feedback loops. Finally, the preliminary CLD was refined through multiple research team meetings, with each causal link traced back to the original data to ensure that the model remained grounded in the empirical material.

Results

As shown in table 1, the participants (n=18) were evenly distributed across urban and suburban districts (50.0% each), and most were female (61.1%). Half of the participants were aged 35–44 years, 77.8% held bachelor’s degrees and 72.2% held senior professional titles (chief physician or associate chief physician). CHC directors had substantially more years of work experience than the GPs (24.6±6.3 vs 11.6±5.2 years). Detailed participant information, including anonymised codes, is provided in online supplemental file 1. These codes were used consistently to attribute quotations in the Results section.

Table 1. Characteristics of participants.

Characteristic	Total (n=18)	Directors (n=10)	GPs (n=8)
Age (years)
25–34	2 (11.1%)	0 (0.0%)	2 (25.0%)
35–44	9 (50.0%)	4 (40.0%)	5 (62.5%)
45–55	7 (38.9%)	6 (60.0%)	1 (12.5%)
Gender
Male	7 (38.9%)	5 (50.0%)	2 (25.0%)
Female	11 (61.1%)	5 (50.0%)	6 (75.0%)
Highest degree
Bachelor’s degree	14 (77.8%)	7 (70.0%)	7 (87.5%)
Master’s degree	4 (22.2%)	3 (30.0%)	1 (12.5%)
Location
Urban	9 (50.0%)	5 (50.0%)	4 (50.0%)
Suburban	9 (50.0%)	5 (50.0%)	4 (50.0%)
Professional title
Attending Physician	5 (27.8%)	0 (0.0%)	5 (62.5%)
Associate Chief Physician	5 (27.8%)	3 (30.0%)	2 (25.0%)
Chief Physician	8 (44.4%)	7 (70.0%)	1 (12.5%)
Work experience (years), mean±SD	18.8±8.7	24.6±6.3	11.6±5.2

GP, general practitioner; SD, standard deviation.

Thematic analysis identified three core themes and 12 subthemes, reflecting key dilemmas in the development of the GPWF in primary care. The themes, subthemes and illustrative quotationsthemes, subthemes and illustrative quotations are presented in table 2.

Table 2. Themes and illustrative quotations.

Themes and subthemes	Descriptions	Illustrative quotations
Rising demand and impact on GP services
Policy-driven growth in demand	The implementation of national health policies promoting primary care has significantly increased the demand for GP services.	With the rollout of numerous hierarchical medical care policies, more patients have been encouraged to seek care at the community level, substantially increasing GPs’ outpatient workload. (GP-08)
Rising demand for geriatric and chronic care	Population ageing and the rising prevalence of chronic diseases have increased demand for geriatric and chronic care services in primary care.	The demand for GPs among community residents has been increasing due to population ageing and the growing prevalence of chronic diseases. (GP-07)
Increasing workload among GPs	Increasing service demands and expanding scope of work have increased the workload of GPs.	Our workload is already full, and this year chronic obstructive pulmonary disease (COPD) was added to the basic public health service. If we are to properly manage these patients, it will mean an even greater workload. (GP-04)
Inadequate service quality and efficiency	Current workforce capacity and system support are insufficient to ensure high-quality and efficient GP services.	At present, meeting the demand for basic medical services often requires extending clinic hours, while public health work is largely reduced to meeting quantitative targets, with little assurance of quality. (D-08)
Supply challenges in the GPWF
Insufficient GPWF supply	The supply of the GPWF is insufficient to meet the growing demand for primary care services.	Because there are not enough GPs with Bianzhi, we can only hire contract-based GPs when the budget allows. To keep services running, GPs have to rotate across different institutions through flexible scheduling. (D-10)
Low attractiveness of GP careers	Social prejudice, role ambiguity and low income were identified as factors reducing the attractiveness of GP careers.	Many residents, and even doctors from other specialties, tend to see GPs as less authoritative than specialists. These views may influence medical students’ career choices. (D-08)
Misalignment between postgraduate training and practice needs	The limited effectiveness of postgraduate GP training was attributed to a mismatch between the design and content of training programmes and the actual needs of primary care practice.	GPs who complete standardised residency training are usually well-trained, but many still struggle when they enter a CHC. Although they have strong medical knowledge and clinical skills, they often lack training in competencies specific to primary care practice. (GP-06)
Limited supply of CSP GPs	The supply of CSP GPs remained limited due to multiple factors, including insufficient training quotas and attrition after the compulsory service period.	Our institution has to wait several years to receive even one CSP GP. Many of them leave once their service commitment ends, leaving us with a persistent shortage of GPs. (D-10)
Barriers to GP recruitment and retention	Recruitment and retention challenges made it difficult for CHCs to maintain a stable workforce.	Overall, Bianzhi positions often receive few eligible applicants, while contract-based GPs are difficult to retain. When a GP leaves, substantial work must be handed over, and frequent staff turnover also disrupts the routine functioning of CHCs. (D-04)
Challenges to GPs’ professional sustainability
Vulnerability to burnout	GPs are highly vulnerable to burnout due to heavy workloads, administrative burdens and limited professional autonomy.	There’s just too much to do. When you try your best to handle everything at once but still can’t manage it all, you begin to doubt your competence as a GP and often feel quite helpless. (GP-08)
Barriers to CPD	GPs face difficulties in participating effectively in CPD programmes and achieving professional development.	There are plenty of training opportunities, but most are task-driven. Some trainings are really useful, but since we rarely see the relevant cases in primary care, those skills gradually fade over time. (GP-02)
Distrust and professional jeopardy	GPs face distrust from patients and are exposed to complaints and potential conflicts that threaten their professional safety.	Some residents don’t trust your abilities and tend to question your judgment. Complaints are quite common, so I’m constantly on edge, worrying that one wrong word might cause trouble. (GP-07)

Bianzhi, government-funded positions with full benefits; CHC, community health centre; CPD, continuing professional development; CSP, compulsory service programme; D, director; GP, general practitioner; GPWF, general practitioner workforce.

Rising demand and impact on GP services

Policy-driven growth in demand

To reduce overcrowding in tertiary hospitals, the government has implemented a range of policies aimed at encouraging residents to seek care at primary care institutions. Participants reported that policies such as increased drug reimbursement rates, expanded medication coverage, waived registration fees for older adults, long-term prescriptions and the establishment of specialty disease clinics (eg, hypertension and diabetes clinics) had made CHCs more attractive to residents. A GP explained, “We can now dispense all the medications patients need, sometimes with an even wider range than that available in hospitals. This helps encourage residents to seek care at CHCs” (GP-02).

Rising demand for geriatric and chronic care

Both directors and GPs pointed out that population ageing and the increasing burden of chronic diseases continue to increase the workload of GPs. One participant noted: “There are more and more older adults who are willing to seek care at CHCs. If minor illnesses and chronic conditions can be managed close to home, the cost of seeking care will be lower and older people can avoid the burden of travelling” (GP-02). Participants generally believed that this trend would continue in the future, with GPs likely to face growing demand for healthcare services related to older adults. One GP noted: “CHCs are now expected to provide an expanding range of services for older adults, including geriatric care, palliative care, and long-term home care” (GP-04).

Increasing workload among GPs

GPs described specific situations illustrating their excessive workload. An urban GP described: “I often see 60–80 patients daily—sometimes over 100, without a lunch break, and I also remain on duty during holidays” (GP-07). Several GPs also noted that, after completing their regular daytime shifts, they were often required to staff extended outpatient clinics and to undertake night shifts and emergency duties.

GPs reported that they were responsible for most of the work related to Family Doctor Contract Services (FDCS) (a healthcare delivery model implemented through CHCs since 2016), while also handling team task allocation and performance evaluation. Most GPs indicated that the assessment requirements had significantly increased their workload. One GP noted, “The assessment indicators for the FDCS service are overwhelming” (GP-05).

Excessive workload was also attributed to public health services and mandated administrative tasks. These included “health education, chronic disease management, community health supervision, routine geriatric assessments, emergency response coordination, research questionnaire distribution, policy evaluation surveys, and healthcare policy promotion” (D-07, GP-07).

Inadequate service quality and efficiency

With the advancement of national health policies and the growing health needs of the population, the current GPWF struggles to deliver high-quality services. As GP-06 stated, “I love working in general practice, but there are too many tasks. As a result, I cannot fully focus on patients’ overall health problems or conduct adequate follow-up care.” Workforce shortages and insufficient capacity building were described as closely intertwined challenges. GP-07 explained, “These health management services for older people should indeed be provided by primary care, but there are not enough GPs, and we lack targeted training to improve our ability to manage ageing-related health problems.”

In addition, the lack of adequate information system support further increased the workforce burden. As GP-01 noted, “Because the health information system is not well developed, we have to manually enter physical examination and FDCS data, which greatly increases workforce demand.”

Supply challenges in the GPWF

Insufficient GPWF supply

Participants described the shortage of GPs. “Last year, outpatient visits increased by about 50%, but staffing levels have barely changed. In fact, we are already so short of GPs that even losing one GP would make it difficult to keep our current services running” (D-07). “The community health service stations cannot meet the requirement of having two GPs at each station. Although beds have been allocated, they cannot be put into operation because of the GP shortage” (D-10).

“One person holding multiple roles” (D-04) was prevalent across CHCs. GPs reported: “There are too few doctors, and we constantly work overtime” (GP-07). Others added: “Tertiary hospitals have trainees and refresher physicians, but no one shares our workload” (GP-04). The GP shortage is further exacerbated by temporary staff redeployments for various reasons. One GP explained, “We need to train at higher-level hospitals for one to six months to qualify for professional title promotion” (GP-04).

Low attractiveness of GP careers

Participants emphasised that the low attractiveness of the GP profession has resulted in a limited number of postgraduate trainees and residents in general practice, leaving a very small pool of medical graduates available for recruitment. Participants believed that social prejudice against GPs may influence medical students’ career choices. One director explained, “Many people still don’t really understand what GPs do, and very few medical students choose to pursue further training in general practice. Our CHC hasn’t recruited a single fresh graduate since 2011” (D-07).

In addition, participants highlighted economic factors. They reported that the income of GPs is substantially lower than that of physicians working in tertiary hospitals, which significantly discourages medical graduates from choosing general practice as a career.

Misalignment between postgraduate training and practice needs

Participants indicated that although standardised GP residency training can improve trainees’ clinical competence, the training content was often disconnected from the realities of primary care practice. For example, one participant stated: “Most trainees have no prior experience working in primary care, and the residency programme lacks targeted training in the core competencies required in CHC settings. As a result, many of them struggle to adapt smoothly to primary care posts after graduation” (D-06). A GP who had recently completed residency training also reflected: “When I came back to the community after training in a tertiary hospital, the psychological gap was huge. I honestly didn’t know how to demonstrate my professional value” (GP-02).

A proportion of registered GPs who had completed job-transfer training did not go on to work in GP positions. As one participant explained, “The training cycle is too long, and resources are often wasted because some people leave soon after completing the training” (D-05). Another added, “Finishing the training doesn’t necessarily mean doctors will take up or remain in GP positions. For example, surgeons who undergo job-transfer training are often not fully able to perform GP duties and therefore remain in their original positions” (D-07).

Limited supply of CSP-trained GPs

CHC directors generally acknowledged that the CSP is an important pathway for strengthening the primary care GPWF, particularly in remote suburban areas. One director explained: “Working conditions in mountainous areas are tough, and it is very difficult to attract and retain qualified GPs. We basically have to rely on CSP quotas to ensure that residents’ basic healthcare needs are met” (D-10).

Participants also identified several challenges facing the CSP, including insufficient quotas for suburban areas and high attrition after completion of the service obligation. One participant recruited through the CSP commented: “Once my service period ends, I will weigh things up—career development opportunities, job location, distance from my family, and salary—before deciding whether to stay here” (GP-01).

Barriers to GP recruitment and retention

In most CHCs, GPs are recruited centrally by district health authorities and employed through Bianzhi posts (government-funded positions with full benefits). However, Bianzhi quotas are fixed and based on the size of the population served, which constrains the number of GPs that CHCs can recruit. Participants noted: “Bianzhi allocation does not account for service scope or regional variation, and it has not been revised for many years, so it no longer meets current workforce needs” (D-10).

In addition, the strict eligibility requirements for Bianzhi posts further limit the available pool of candidates. As one manager explained, “Because of the Beijing household registration restrictions, the requirements for Bianzhi posts are quite high. Non-local applicants must have a master’s degree and be under 30. We haven’t recruited a single qualified GP in the past five years” (D-01).

For applicants with Beijing household registration, the educational requirement can be lowered to a bachelor’s degree in medicine, but they are required to complete 3 years of residency training before taking up the position. Nevertheless, attrition occurs during this period. As directors noted, “Some resign during training, and others fail the licensing exam, so the position just stays vacant” (D-03). Another director added, “Even after completing training, they may still leave to pursue postgraduate studies or move to higher-paying hospitals” (D-08).

Due to limited Bianzhi quotas, CHCs often recruit contract-based GPs to supplement their workforce. However, despite undertaking responsibilities comparable to those of Bianzhi staff, contract-based GPs receive significantly lower salaries. This disparity contributes to high turnover among contract-based GPs, making it difficult to effectively alleviate the shortage of GPs.

Challenges to GPs’ professional sustainability

Vulnerability to burnout

Most GPs reported working under high pressure. The sustained heavy workload and the complexity of their tasks have dampened their enthusiasm for work. One GP said, “Many times, I work in the outpatient clinic during the day and handle emergency calls (120) at night. Especially during the pandemic, I could hardly get proper rest almost every day. Both physically and mentally, I was under tremendous pressure, and at times I even questioned the meaning of my work” (GP-01).

When faced with assigned administrative tasks, GPs commonly reported a lack of “professional autonomy” (GP-06). For example, GP-08 noted, “Who wants to do work outside their own specialty? But because GPs have the most direct contact with residents, they think we’re the most suitable to take on these administrative duties.” In addition, GPs provided 24-hour consultation services for contracted residents. Although many GPs acknowledged that this helped maintain doctor-patient relationships, most also expressed a sense of helplessness: “Sometimes I get calls in the middle of the night, and they’re not even emergencies—it’s just exhausting” (GP-08).

Due to the need to complete tasks such as follow-up visits and scheduling physical examinations, as well as to ensure compliance with health insurance and other policy requirements during consultations, GPs reported having limited time and energy for patient care. As GP-07 stated, “We don’t have time to provide targeted health management for patients. Sometimes we even end up prescribing medication almost mechanically, which makes me feel like I’m not a competent GP” (GP-07). This inability to provide individualised care made it difficult for GPs to realise the ideal model of general practice, leading to a sense of frustration and a diminished professional identity. As GP-01 stated, “Under the current working model, patient-centred care has become little more than an empty slogan.”

Barriers to continuing professional development

GPs had access to various continuing professional development (CPD) pathways, including “online and offline lectures, specialised training, academic conferences, and placements in tertiary hospitals” (GP-04, GP-07). However, most participants felt that these opportunities were insufficient for meaningful professional development. Participation was often limited by conflicts with clinical duties and personal time. In addition, much of the training content was poorly aligned with CHC settings, limiting its value for improving GPs’ clinical competence.

One GP explained, “I completed the training, but I’ve never encountered such cases in my outpatient work. In fact, after moving to primary care, I feel like my ability to manage many conditions has actually declined” (GP-02). GP-06 also questioned the value of some training: “Now we have opportunities to take part in research and teaching training, which is great. But what worries me is that even if we acquire these skills, CHCs don’t have the resources to support research.”

Directors also discussed GPs’ motivation to engage in CPD. As one noted, “The CPD credit requirements are rather formalistic, and there is no effective assessment mechanism” (D-05). Others added that “publications are not required for promotion in primary care, and senior professional titles are recognised only within primary care settings” (D-02, D-06). Directors believed that these factors might reduce GPs’ motivation to participate in CPD.

Distrust and professional risks

Participants generally believed that residents’ distrust of GPs’ competence and high service expectations contributed to doctor-patient conflict. Many residents tended to prefer large hospitals. As one GP explained, “A patient’s blood pressure was poorly controlled, so I adjusted his medication. But he insisted that the treatment plan from a tertiary hospital three years earlier was better—this clearly reflects prejudice against GPs working in primary care” (GP-03).

High patient expectations and frequent complaints were also reported to place substantial psychological pressure on GPs. As GP-06 noted, “I’m always worried about receiving complaints, and every complaint takes a lot of time and energy to deal with.” In some cases, unmet demands escalated into conflict or violence. GP-02 recalled: “A patient once threw a small stool at me just because I didn’t prescribe the medication he wanted.” The same GP further noted that CHCs often lacked adequate security measures: “There’s no security screening or security staff in the CHC, so it’s hard to get help quickly when conflicts happen.”

Findings-based CLD

Figure 1 presents a CLD illustrating the challenges in the development of the GPWF in primary care. Ten key variables were identified: high workload, insufficient GPWF supply, reduced quality of care, barriers to CPD, recruitment and retention difficulties, difficulties in GP training, social stereotypes, low job attractiveness, inadequate salary and benefits and job burnout. In this study, job burnout was defined on the basis of participants’ descriptions of emotional exhaustion, depersonalisation and reduced personal accomplishment. Table 3 presents the causal relationships between key variables, along with illustrative quotations.

Figure 1. CLD of GPWF development in primary care. Arrows in the diagram denote causal relationships. CLD, causal loop diagram; CPD, continuing professional development; GP, general practitioner; GPWF, general practitioner workforce.

Table 3. Causal relationships between key variables with illustrative quotations.

No.	Description of causal relationship	Illustrative quotations
C1	High workload leads to job burnout.	The workload has been very heavy, especially since the COVID-19 pandemic. At times, it feels as if we are completely drained. (GP-01)
C2	High workload leads to reduced quality of care.	When seeing patients, we also have to check whether they are due for follow-up or need to complete any questionnaires. These administrative tasks make it harder to focus on clinical care, sometimes leading to less thorough consultations and poorer continuity of care. (GP-07)
C3	High workload leads to barriers to CPD.	We’re so busy that we hardly have time for training, and opportunities for research and teaching in primary care are limited. It’s hard for GPs to make progress in these areas. (GP-07)
C4	High workload leads to recruitment and retention difficulties.	CHCs now have to provide emergency services and extended outpatient hours. Many contract-based GPs, in particular, have left for private hospitals because of the heavy workload. (GP-04)
C5	Low job attractiveness leads to difficulties in GP training.	Few graduates choose general practice residency training or postgraduate study in general practice. Most are assigned to the specialty through reassignment, and even after completing training, they do not necessarily go on to work as GPs. (D-08)
C6	Difficulties in GP training lead to insufficient GPWF supply.	The root problem is that too few GPs are trained from the outset. Many applicants have backgrounds in traditional Chinese medicine or other specialties, making shortages in the GPWF almost inevitable. (D-02)
C7	Insufficient GPWF supply leads to high workload.	We simply don’t have enough doctors, so we have to rely on overtime and rotating shifts to meet service demands. As a result, GPs’ workload keeps increasing. (D-05)
C8	Recruitment and retention difficulties lead to insufficient GPWF supply.	We haven’t been able to recruit anyone for a long time, and the workload just keeps increasing, so we’re definitely understaffed. (D-03) Suburban areas just aren’t attractive, and it’s hard to retain people, so shortages are especially serious. (D-01)
C9	Reduced quality of care leads to social stereotypes.	Service quality still needs improvement. Many residents have a limited understanding of GPs to begin with, and if their experience during a visit is not good, it only reinforces the belief that GPs, and primary care more broadly, are there only to prescribe medications. (GP-07）
C10	Barriers to CPD lead to reduced quality of care.	Training is seriously disconnected from practice—when guidelines are updated, there’s no training on how to apply them, and chronic disease management training is often just theoretical. This directly undermines service quality and does little to improve residents’ health. (GP-02)
C11	Barriers to CPD lead to recruitment and retention difficulties.	Especially in suburban areas, opportunities for professional development are very limited. It is hard to make progress, whether in clinical skills or title promotion, so capable GPs often try to move to higher-level hospitals. (D-10)
C12	Inadequate salary and benefits lead to recruitment and retention difficulties.	Although performance-based pay is now the same for contract-based GPs and those in Bianzhi posts, gaps in base salary and benefits remain, leading to high turnover among contract-based GPs. (D-06)
C13	Inadequate salary and benefits lead to low job attractiveness.	The salary and benefits cannot compete with those of specialists in tertiary hospitals, which is one reason why GP careers are less attractive. (D-01)
C14	Social stereotypes lead to recruitment and retention difficulties.	This situation can undermine GPs’ confidence. I sometimes question whether I chose the right specialty and have even considered leaving the profession. (GP-02)
C15	Social stereotypes lead to low job attractiveness.	There is still bias against GPs in society. Medical students often think that general practice offers limited career prospects and therefore do not consider it their first career choice. (GP-05)
C16	Job burnout leads to reduced quality of care.	Sometimes, when we’re extremely tired, we don’t feel like talking much with patients, and we lack the motivation to follow up on their health proactively. (GP-08)

Bianzhi, government-funded positions with full benefits; C, causal relationship; CHC, community health centre; CPD, continuing professional development; D, director; GP, general practitioner; GPWF, general practitioner workforce.

The CLD suggests two interacting reinforcing loops perceived by participants (Loop A and Loop B), which together capture the self-reinforcing nature of the dilemmas facing the primary care GPWF. Increased service demands contribute to a high workload. However, workforce supply has not kept pace with this demand, and adjustments to staffing and remuneration policies have lagged. Consequently, high workload leads to job burnout, barriers to CPD and reduced quality of care. These outcomes, in turn, reduce the attractiveness of GP careers, reinforce social stereotypes and worsen difficulties in training, recruitment and retention. Collectively, these factors exacerbate workforce shortages, which further increase the workload and perpetuate the cycle.

Discussion

This study employed thematic analysis combined with a CLD to examine the multiple challenges facing the development of the GPWF in primary care in Beijing and to explore the mechanisms linking these challenges. The findings indicate that, in the context of rapidly expanding service demand, the supply of GPs is insufficient. High workload, low job attractiveness, reduced quality of care and limited professional development interact to form a self-reinforcing cycle that constrains GPWF development. These findings deepen the understanding of GP shortages and provide a basis for identifying key intervention points and optimising workforce strategies.

From the demand side, the growth in primary care utilisation is both a major driver of GPWF demand and a direct source of increased workload. Policies aimed at strengthening primary care have successfully increased demand for its services. However, even in Beijing—where primary care infrastructure is relatively strong and GPs are comparatively well-trained—significant workforce shortages persist, suggesting limitations in current workforce planning and development. In practice, the long training cycle of health professionals, reliance on lagged statistical data for workforce planning²⁴ and the need for coordination across multiple sectors (eg, health, human resources and finance) all contribute to delayed policy responses. As a result, adjustments in GP training, remuneration and support systems often lag behind service expansion, leading to a sustained increase in workload that exceeds workforce capacity. Similar patterns have been observed in countries such as the UK and Canada, where imbalances between service expansion and workforce support have been associated with declining care quality and patient satisfaction.^{25 26} These findings highlight the importance of maintaining alignment between demand growth and workforce capacity in primary care reform.

From the supply side, one of the central challenges is the low attractiveness of GP careers. In many contexts, general practice is perceived as offering limited opportunities for academic advancement and career progression.^{27 28} In this study, unclear professional roles and income disparities between GPs and specialists further reduced the appeal of the profession among medical graduates. To address this issue, China has introduced measures to increase early exposure to general practice in medical education and to promote positive role models.²⁹ In addition, more competitive compensation and incentive structures are essential to influence career choices among medical students and GP trainees.^{30 31} Recent policy initiatives in China have sought to strengthen the primary care system by directing health insurance resources towards primary care institutions and improving payment mechanisms.³² While these efforts may enhance the attractiveness of GP positions and support workforce recruitment and retention, their effective implementation requires sustained coordination across relevant sectors.

In addition, the limited efficiency of translating GP training into actual workforce supply further exacerbates shortages. This study identified a clear disconnect between training and employment, with participants reporting difficulties in transitioning into primary care roles after completing training. This observation is consistent with previous studies showing that only a proportion of trained physicians ultimately register and practise as GPs.^{33 34} International evidence suggests that maintaining continuity between training and employment can facilitate more stable entry into the workforce.^{35 36} Therefore, training content should be better aligned with the realities of primary care practice, and post-training job placement and follow-up support should be strengthened. Furthermore, as the CSP plays a critical role in supporting rural primary care, expanding its scale and enhancing career development and retention pathways for participants will be important for its long-term effectiveness.

Finally, the CLD developed in this study provides an analytical framework for understanding GP shortages as a dynamic and self-reinforcing process. In the CLD, factors such as workload, service quality, recruitment and retention, and CPD play key roles within multiple feedback loops and represent priority areas for intervention to address the supply–demand imbalance. These findings suggest that isolated policy measures are unlikely to effectively address workforce challenges; instead, a coordinated policy approach is needed to break the negative cycle linking high workload, low job attractiveness and poor retention.

Based on these findings, several policy implications can be drawn. First, dynamic monitoring of workload, workforce mobility and service demand is needed to improve the responsiveness and precision of workforce planning and reduce reliance on lagged data. Second, stronger intersectoral coordination across health, education, human resources and finance is essential to shorten the policy implementation chain linking training expansion, funding and workforce deployment. Third, the professional role of GPs should be redefined in the context of a non-mandatory gatekeeping system, with enhanced CPD design and competency development to strengthen their role in health management, integrated care and continuity of care. Fourth, alleviating excessive workload requires stronger system-level support, including reducing low-value administrative tasks, optimising incentives, strengthening team-based care and leveraging digital tools to streamline administrative workflows and enhance clinical care.837,39

Methodological considerations

This study used a qualitative design to explore the experiences and institutional contexts influencing GPWF development. The focus group discussion provided insights into managerial perspectives on policy implementation, workforce allocation and organisational operations, while individual interviews offered detailed accounts of frontline GPs’ experiences, including workload, burnout and career development. The integration of thematic analysis with CLD construction allowed the study to move beyond identifying key issues and to illustrate the relationships among key variables, thereby enhancing understanding of the mechanisms underlying GP shortages. However, the CLD findings should be interpreted as providing insight into potential mechanisms and relational pathways rather than as definitive evidence of causal relationships.

These methodological choices may also have influenced data generation and interpretation. In the focus group setting, participants with higher status or stronger communication styles may have shaped the direction of discussion, potentially limiting the expression of divergent views. In individual interviews, responses may have been influenced by interviewer–participant interactions and questioning approaches.⁴⁰ In addition, participants were primarily recruited through institutional channels, which may have resulted in the overrepresentation of individuals with stable organisational relationships and a willingness to participate, while underrepresenting those who had left the GPWF or held more critical perspectives.

Limitations

First, the study primarily adopted a supply-side perspective, focusing on GPs and CHC directors, and did not include patients, community residents, policymakers or physicians who have left the GPWF. As a result, the understanding of demand-side dynamics and attrition processes remains incomplete. Second, Beijing represents a relatively well-resourced and policy-advanced setting, and its institutional context may differ from that of less-developed regions; therefore, caution is required when generalising the findings. Third, the CLD is based on qualitative data and is intended to illustrate potential feedback mechanisms rather than to establish causal relationships. The direction, magnitude and relative importance of these relationships require further validation through studies involving broader samples and quantitative approaches.

Conclusion

The development of the GPWF in Beijing faces multiple interrelated challenges. Growing service demand, insufficient workforce supply, low job attractiveness, misalignment between training and practice, and inadequate support systems interact to create a self-reinforcing cycle of workforce shortages. To prevent this cycle from intensifying, future policies should prioritise coordination across sectors, establish dynamic workforce planning mechanisms responsive to changing service demands, develop clear career pathways to enhance workforce stability, and strengthen support systems to reduce the workload. At the same time, targeted efforts to improve GP competencies and enhance societal recognition are essential for strengthening professional identity and ensuring the sustainable development of the GPWF.

Data availability statement

The anonymised coding framework is available upon reasonable request.