Strategies for enhancing PHC accessibility through mobile and capsule clinics: a spatial location allocation study in China

Abstract

Background

Shortages in the Primary Health Care (PHC) workforce, especially in remote areas, pose a major challenge to achieving universal health coverage. Traditional strategies focused solely on hiring village doctors face limitations due to recruitment challenges in these regions. This study proposes a multi-strategy approach to PHC workforce allocation to address spatial imbalances and improve healthcare accessibility and equity.

Methods

We developed a Multi-Strategy PHC Workforce Location Allocation model that integrates three strategies: hiring village doctors, deploying mobile clinics, and establishing capsule clinics. The model aims to minimize patient impedance while considering constraints such as service coverage, capacity, and budget. Using Beijing’s suburban areas as a case study, we conducted empirical runs to compare two models. The first is a single-strategy model that relies solely on hiring village doctors. The second is a multi-strategy model that combines hiring village doctors with mobile and capsule clinics. Data were collected from 34 towns, 823 villages, and 894 medical facilities. We evaluated the models based on impedance, cost, accessibility, and Gini coefficients, which were calculated using methods such as the two-step floating catchment area approach.

Results

Compared with the single-strategy approach, the multi-strategy model consistently achieves higher coverage, lower impedance, better accessibility, and improved equity under the same budget. For example, with a 10% budget parameter, it covers 19.23% more villages in deep mountainous areas and improves accessibility in the plains, near mountainous, and deep mountainous areas by 1.59%, 4.80%, and 4.44%, respectively, outperforming the single-strategy model. As government investment increases, the improvement in impedance shows a trend of diminishing marginal returns.

Conclusions

Integrating hiring village doctors with mobile and capsule clinics offers a flexible and effective solution for addressing PHC workforce shortages in remote areas. This multi-strategy approach not only improves healthcare accessibility and fairness but also achieves higher financial efficiency. We recommend that governments adopt this comprehensive strategy, especially in resource-constrained regions, to achieve universal PHC service coverage.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-025-13083-z.

Introduction

In the 2018 Astana Declaration, the World Health Organization (WHO) emphasized that Primary Health Care (PHC) is the cornerstone for achieving “Health for All” [1]. Numerous studies have consistently demonstrated that PHC can deliver better health outcomes at lower costs [2, 3]. One of the most important aspects of improving PHC is how its workforce is allocated. PHC providers play a critical role in delivering care and ensuring resources are used effectively [1, 4, 5]. Consequently, countries worldwide are continuously striving to strengthen and expand their PHC workforce to achieve universal healthcare coverage [6]. For instance, in recent years, China has significantly increased its investment in PHC human resources [3], leading to a rapid growth in the number of PHC doctors per thousand people [7, 8]. Consequently, cities like Beijing have reached the WHO-recommended standard of one PHC doctor per thousand population. However, this rapid growth conceals a severe imbalance in the spatial distribution of PHC doctors [9–12]. Research indicates that in major cities like Beijing and Shanghai, PHC human resources are predominantly concentrated in urban areas, while rural regions face noticeable shortages of health human resources [13–16]. To address this challenge, the Chinese government has intensified its focus on the equitable distribution of the PHC workforce. They are striving to achieve comprehensive PHC doctor coverage at the village level. In this context, addressing the spatial imbalance of PHC human resources and identifying optimal locations to supplement the PHC workforce have become urgent challenges for China [17].

Existing studies have often relied on Location Allocation (LA) models to optimize the distribution of PHC human resources. These models focus on critical questions, such as “where to employ PHC village doctors.” For instance, Dogan et al. developed a multi-objective mixed-integer linear programming model to optimize the placement of PHC facilities and doctors, improving cancer screening participation rates and controlling waiting times in Istanbul [18]. Similarly, Mendoza-Gómez R. et al. applied the P-Median Model (PMM) within LA to identify optimal locations for PHC facilities in Mexico, with a particular focus on underserved remote areas [19, 20]. These studies offer valuable strategies for equitable resource distribution and for strengthening PHC services in rural and remote settings.

However, a persistent challenge remains: recruiting and retaining village doctors in rural and mountainous regions is difficult worldwide [21–25]. This limitation significantly undermines the feasibility of strategies that depend solely on hiring village doctors for these areas. As a result, PHC workforce LA models that rely exclusively on this approach often yield solutions that are difficult to implement and lack long-term sustainability.

In response to these challenges, technological advancements have enabled the development of alternative methods to address workforce shortages in remote rural areas. For example, mobile clinics and capsule clinics provide flexible and innovative solutions by delivering essential medical services where traditional workforce allocation methods face limitations. These approaches expand the reach of PHC services, offering a viable complement to traditional models. Mobile clinics allow doctors to travel to patients. Using mobile vehicles or even horseback, PHC providers bring diagnostic tools and treatment facilities directly to underserved areas. Relevant reviews have shown that mobile clinics can enhance the accessibility of PHC services at relatively low costs, making them particularly effective in improving health outcomes for rural populations [26]. This strategy not only expands PHC coverage but also ensures that essential healthcare reaches remote and hard-to-reach communities [27–30]. It has already been widely implemented in countries such as China, Syria, Iraq, and Ukraine to strengthen PHC accessibility [31]. Capsule clinics, on the other hand, use telemedicine to provide care [32–34]. These small, unmanned clinics are set up in villages and offer 24-hour health checkups, consultations, and medication services [35]. Related studies have also highlighted that the telemedicine services provided by capsule clinics demonstrate significant advantages in primary healthcare, chronic disease management, and other areas, with consistently high patient satisfaction rates [36, 37]. Pilot implementations have already emerged in countries such as China and the United States [38]. Together, mobile and capsule clinics provide new ways to improve access to healthcare in remote areas. In China, many regions have started pilot programs for mobile and capsule clinics. Mobile clinics are often staffed by doctors from township health centers. These doctors visit remote mountain villages once a week, traveling by vehicle to deliver basic healthcare services. In contrast, capsule clinics are fixed, unmanned facilities located within villages. They use telemedicine to connect villagers with doctors, who provide PHC services remotely. Medications are dispensed on the spot through automated machines in the clinics. During the pilot phase, both mobile and capsule clinics have proven to be effective solutions for improving PHC access. As a result, in 2023, the Chinese government introduced a policy supporting a combined approach that includes hiring village doctors, deploying mobile clinics, and establishing capsule clinics to address PHC accessibility issues [39]. However, balancing the allocation of the PHC workforce using this multi-strategy approach remains a complex challenge. The government must take into account not only the current distribution of the PHC workforce supply and demand, but also factors like budget limitations, the costs and service capacities of each strategy, and geographical considerations. To address this, we have developed an improved LA model designed to determine where to hire village doctors and where to deploy mobile clinics or capsule clinics. It has significant practical implications for optimizing PHC resource allocation and improving healthcare access in underserved areas.

As mentioned earlier, PHC workforce LA models that focus only on hiring village doctors often fail to address regional disparities and human resource shortages. To overcome this, this paper proposes a multi-strategy approach for allocating PHC human resources. By combining the hiring of village doctors with mobile and capsule clinics, the approach improves the efficiency of the PHC workforce distribution. To test the model, we used Beijing as a case study. Beijing’s diverse topography creates significant disparities in PHC human resources. Suburban areas face severe imbalances, while remote mountainous regions continue to experience persistent shortages of village doctors [13, 14]. To address these challenges, Beijing has launched pilot programs combining all three strategies: hiring village doctors, mobile clinics, and capsule clinics. These pilots provide a solid real-world foundation for the empirical run in Beijing. In summary, this paper develops a multi-strategy PHC workforce LA model and presents an allocation plan based on Beijing’s case. The research offers valuable insights for developing countries looking to tackle PHC workforce shortages with a variety of strategies.

Methods

Model construction

Basic location allocation model

Location Allocation (LA) modeling is a research method used within Geographic Information Systems (GIS) to determine the optimal locations for public service facilities. It is commonly employed to address the optimization of medical resource allocation. From a computational perspective, the core of LA can be viewed as a linear programming problem, comprising key components such as the objective function, constraints, and optimization strategies. Traditional LA models, based on different objective functions, primarily include the minimization facility location model, the maximization coverage model, and the P-Median Model (PMM).

In practical planning, the PMM, which aims to minimize patient impedance, is widely applied. Impedance refers to the various barriers patients encounter when accessing medical facilities, including distance, time, traffic conditions, and other factors that may affect the convenience of seeking medical care. By reducing the total impedance experienced by the population within a region, the allocation of the PHC workforce can be optimized, thereby enhancing patient satisfaction. The objective function of the PMM is to minimize the impedance between patients and clinics. The model has two main constraints: hiring P additional doctors and ensuring all villages are covered. The optimization strategy focuses on hiring village doctors. Figure 1 shows the framework of the PMM.

Fig. 1.

Framework of the PMM and the multi-strategy PHC workforce LA model

The mathematical model is outlined in Eqs. (1, 2, 3 and 4). In the equations, i represents a village, and j denotes a clinic. D_i refers to the annual medical demand of village i, and d_ij represents the distance between village i and clinic j. The decision variable x_ij ∈ {0,1} indicates the supply-demand relationship between village i and clinic j. If clinic j serves village i, then x_ij= 1; otherwise, x_ij= 0. The decision variable y_j ∈ {0,1} represents whether candidate location j is selected. If j is selected, y_j= 1; otherwise, y_j= 0. P denotes the number of village doctors to be hired. Equation (1) is the objective function, which minimizes the weighted sum of distances between villages and clinics, weighted by the demand D_i, thereby minimizing impedance. Constraint (2) ensures that each village i is covered by exactly one clinic. Constraint (3) limits the number of village doctors hired to P. Constraint (4) stipulates that only selected clinics j can provide services.

1

s.t.

2

3

4

Construction of multi-strategy PHC workforce LA model

To meet the needs of the PHC workforce location allocation, we developed a Multi-Strategy PHC Workforce LA Model based on the PMM. As shown in Fig. 1, the model’s objective function is to minimize the impedance between villages and clinics. The model includes five constraints: (1) All villages are covered. We require that each village in the model be covered by at least one clinic. (2) All villagers’ healthcare demands are met. We require that the medical demands of each village can be fully met by one or more nearby clinics. (3) Patient load within the clinics’ capacity. To avoid medical congestion, we require that the demand assigned to each clinic does not exceed its capacity limit. (4) Clinics’ service radius constraint. Different levels of clinics have different service ranges. In the planning model, clinics can only serve villages within their service radius. (5) Budget constraint. In reality, government resource allocation is typically subject to budget limits. Therefore, we require that the total cost of all strategies in the model does not exceed the government’s budget cap.

We have developed three optimization strategies: hiring village doctors, deploying mobile clinics, and establishing capsule clinics. These strategies differ in terms of cost and capacity. For comparison, the single-strategy model employs only the hiring of village doctors, while the multi-strategy model integrates the hiring of village doctors with mobile and capsule clinics. In our analysis, we assume that these three approaches deliver relatively comparable service quality and diagnostic capabilities, allowing us to focus on differences in cost efficiency and resource allocation. The framework of the Multi-Strategy PHC Workforce LA Model is shown in Fig. 1.

The mathematical expressions of the model are detailed in Eqs. (5, 6, 7, 8, 9, 10, 11 and 12). In the equations, i represents a village, and j denotes a clinic. m is a candidate village for adding new clinics, and n ∈ {1: Hiring, 2: Mobile, 3: Capsule} represents the strategy for establishing new clinics. The decision variable y_mn ∈ {0,1} indicates whether candidate village m adopts strategy n to supplement the PHC workforce. Specifically, when y_m1 = 1, the model hires one village doctor in village m; when y_m2 = 1, it assigns mobile clinics to conduct rounds in village m; and when y_m3 = 1, it deploys capsule clinics in village m. If y_mn = 0, village m is not selected to supplement the PHC human resources. The decision variables x_ij and x_im are positive integers indicating the number of personnel assigned from village i to clinic j or candidate village m, respectively. d_ij and d_im denote the distance from i to j or m. r represents the hospital level, and d_r denotes the service radius of the hospital. D_i is the annual medical demand of village i. S_j is the service capacity of clinic j, while S_n is the service capacity of the three types of new clinics. P_n represents the annual cost of the three strategies. N_village is the total number of villages in the region. p is the proportion of village doctors to be supplemented, determining the budget based on the cost of supplementing p villages with village doctors.

5

s.t.

6

7

8

9

10

11

12

As shown in Fig. 1, Eq. (5) is the objective function, which minimizes the weighted sum of distances between villages and clinics, weighted by x_ij or x_im.

Equations (6, 7, 8, 9, 10, 11 and 12) constitute the model’s constraints. Equation (6) is the All Villages Are Covered constraint. It ensures that village i falls within the service radius of at least one existing clinic j or new clinic m, meaning that each village is covered by at least one clinic. Equation (7) restricts candidate location m to adopt only one of the three strategies: n ∈ {1: Hiring, 2: Mobile, 3: Capsule}, or not adopt any strategy.

Equation (8) is the All Villagers’ Healthcare Demands Are Met constraint. It ensures that the medical demand D_i of village i can be fully satisfied by nearby existing clinic j or new clinic m, and that the distance to seek medical care is within the service radius of clinic j or m. Equation (9) further restricts that if a candidate location m is not selected (y_mn = 0), it cannot provide services to village i, meaning that x_im = 0 when y_mn = 0.

Equation (10) is the Patient Load Within Clinics’ Capacity constraint. It ensures that the medical demand assigned to clinic j and new clinic m does not exceed their capacity limits, S_j and S_n, respectively.

Equation (11) is the Clinics’ Service Radius Constraint. It ensures that, during the planning process, clinics of different levels can only serve villages within their respective service radii.

Equation (12) is the Budget Constraint. It ensures that the annual cost of supplementing the PHC workforce does not exceed the annual budget cap. The left side of the equation represents the actual cost, while the right side represents the budget cap. Assuming the government plans to supplement village doctors in 10% of the villages, the budget is set to 0.1 × N_village × P₁. Within the budget, it is permissible to combine multiple strategies to supplement the workforce.

For ease of comparison, we defined the single-strategy model to employ only the hiring village doctors strategy, setting n = 1. In the multi-strategy model, each village can choose any one of the three strategies: hiring village doctor, mobile clinics, or capsule clinics to supplement the workforce, with n ∈ {1: Hiring, 2: Mobile, 3: Capsule}.

Empirical run in Beijing

Research area

Beijing is located in the northern part of the North China Plain, with a total area of 16,410 square kilometers. The city is surrounded by mountains to the west, north, and northeast, while the southeast is predominantly flat. The suburban areas of Beijing can be divided into three types based on their terrain: plains, near mountainous areas, and deep mountainous areas, Fig. 2 presents the sample area map drawn based on GS(2023)2767 [40]. For this study, we selected Tongzhou District, Mentougou District, and Huairou District as representative samples of these three regions, respectively. Tongzhou District lies in the southeastern plains of Beijing, covering 907 square kilometers, with a permanent population of 1.843 million in 2022 [41]. Its rural areas include 14 towns and 423 villages. Mentougou District, located in the mountainous western part of Beijing, has flatter terrain in the east. It spans 1,455 square kilometers and had a permanent population of 396,000 in 2022 [41]. Its rural area includes 9 towns and 138 villages. Huairou District is in the remote northern mountains, where 89% of the land is mountainous. The district covers 2,122 square kilometers and had a population of 439,000 in 2022 [41], with 11 towns and 262 villages in its rural areas. In total, our study incorporates data from 34 towns, 823 villages, and 894 hospitals or clinics.

Fig. 2.

Relative location of the research area in China

Model setup and operation

We used Beijing as a case study to run our model, which includes three parts: input, solving, and output. The inputs are divided into four components: PHC demand data, PHC supply data, a distance matrix, and optimization strategy parameters. These inputs were collected through surveys and used to build the Multi-Strategy PHC Human Resource LA Model. The model dynamically adjusts decision variables based on the objective function and constraints to find optimal solutions.

To compare multi-strategy and single-strategy models, we set the single-strategy model to hire only village doctors, while the multi-strategy model could choose among three strategies. By comparing their outputs and indicators, we assessed the multi-strategy model’s applicability and proposed PHC workforce allocation plans for decision-makers.

ArcGIS was used for database management, while Python’s pulp library constructed and solved the linear programming model. Outputs were visualized with ArcGIS, and indicators were calculated using Python. Details of the model’s inputs, outputs, and data sources are in [Additional file 1], and the model’s operation schematic is in [Additional file 2].

Inputs and data sources

The model’s inputs consist of PHC demand data, PHC supply data, a distance matrix, and optimization strategy parameters. The PHC demand data were obtained through a questionnaire survey targeting township health centers in suburban Beijing. A baseline survey of village-level health human resources was conducted from December 15 to December 31, 2020, followed by a verification and supplementation phase from January 4 to June 30, 2021, to ensure data quality [40].

The PHC supply data were sourced from both the survey data and the 2020 “Compilation of Beijing Medical Statistical Data”. In China, healthcare institutions are categorized into three levels, with primary healthcare institutions including township health centers and their subordinate health stations and village clinics. Consequently, the supply data in our study covered five levels of healthcare facilities: village clinics, health stations, township health centers, secondary general hospitals, and tertiary general hospitals. The number of doctors in village clinics and health stations was derived from the survey data, whereas data for township health centers, secondary general hospitals, and tertiary general hospitals were obtained from the “Compilation of Beijing Medical Statistical Data”.

Village and hospital coordinates were collected using online mapping tools, and the travel distances between supply and demand points were calculated using the Baidu Maps distance calculation API. The optimization strategy parameters were primarily determined based on data from the “China Health Statistics Yearbook” and expert consultations. Detailed data sources and calculation results are provided in [Additional file 1]. The cost parameters for the optimization strategies were determined based on interviews with township health centers in sample areas and government reports. The detailed cost estimation is provided in [Additional file 3]. The final parameter values used in the model are summarized in Table 1.

Table 1.

Model input parameters

Hospital type	Level r	Service radius dr (km)	Service Capacity Sj, Sn (thousand persons/year)	Optimization Strategy n	Strategy Cost Pn (thousand RMB/year)
Tertiary general hospital	5	50.0	2.86 Docj	-	-
Secondary general hospital	4	25.0	2.86 Docj	-	-
Township health center	3	10.0	3.92 Docj	-	-
Health station	2	5.0	3.92 Docj	-	-
Village doctor clinic	1	2.5	3.53 Docj	1	48.00
Mobile clinics	1	2.5	1.18	2	21 + 0.59 distance
Capsule clinics	1	2.5	2.35	3	29.55

Outputs and calculation methods

The model outputs include workforce allocation plans, as well as the indicators Impedance, Cost, Accessibility, and Gini coefficient, with detailed calculations outlined as follows.

Impedance

Impedance minimization is the model’s optimization objective, and its optimal solution corresponds to the annual travel distance for medical visits by all villagers in the region. Considering the significant population differences among the three sample areas in Beijing, we use per capita impedance as the model’s output metric, defined as the annual average travel distance for medical visits per villager. This enhances comparability between regions, as specified in Eq. (13).

13

While impedance minimization is the model’s primary optimization objective, the cost, accessibility, and fairness of the plan are also important evaluation metrics. Therefore, we selected Cost, Accessibility, and Gini coefficients as evaluation indicators to further compare the solutions of the two models.

Cost

After the model obtains the optimal solution, it directly outputs the decision variable y_mn matrix results. Based on the y_mn matrix, we can derive the specific configuration plans for each candidate village m adopting strategy n. The relevant output results are visualized using ArcGIS. Additionally, the actual expenditure Cost of the plans is also output.

Accessibility

Accessibility is a spatial metric that measures how easily rural residents can access surrounding medical resources. It is commonly used to evaluate the rationality of public resource allocations, such as PHC [42]. In this study, we quantified the accessibility of villagers’ medical services before and after optimization using the widely employed Two-step Floating Catchment Area (2SFCA) method [13, 43], as shown in Eqs. (14, 15 and 16).

First, the configuration plan provided by the optimal solution is entered into the Hospital GIS. Then, using Eq. (14), we calculate the Gaussian distance decay function g(d_ij,d_r) between village i and hospital j. Gaussian and other decay functions (e.g., kernel density) have been widely used to model how distance affects healthcare-seeking behavior. Following Zhuolin Tao et al., we adopt the Gaussian function for its gradual decline near zero travel distance and at the edge of the catchment [44], which better reflects actual patient choices. Here, i represents the village index, and j represents the hospital index after supplementation. d_ij is the distance between village i and hospital j. Let r denote the hospital level, and d_r represent the hospital’s service radius.

14

The 2SFCA is an evaluation method that separately calculates the supply-to-demand ratio within the service radius from both the supply and demand sides. In the first step, all villages i within the service radius of hospital j are identified, and the supply-to-demand ratio R_j for hospital j is calculated using Eq. (15). In the second step, all hospitals j that can serve village i are identified, and the accessibility A_i for village i is calculated using Eq. (16). A higher value of Ai indicates better accessibility. In these equations: Doc_j represents the number of doctors at hospital j; Pop_i is the population of village i; R_j is the supply-to-demand ratio of medical services at hospital j; A_i is the medical service accessibility at village i. The average Ai value within the district is defined as Accessibility.

15

16

Gini

The Gini coefficient is one of the most commonly used indicators of fairness [45, 46]. In this study, we use the Gini index to measure the fairness of medical impedance across different villages, as shown in Eq. (17). In the equation, X_i represents the cumulative percentage of the village population, and Y_i represents the cumulative percentage of per capita medical impedance in village i.

17

Results

Baseline description of the PHC workforce

The baseline conditions of PHC human resources in the sample area are presented in Table 2; Fig. 3. The accessibility shown in Fig. 3 refers to the weighted sum of the supply-demand ratios of medical resources surrounding each village. It is a positively oriented indicator, with green areas indicating better accessibility. Firstly, the number of PHC doctors per thousand population in Beijing’s plain area, near mountain area, and deep mountain area is 1.64, 5.58, and 3.73, respectively. All these figures exceed the WHO-recommended standard of one PHC doctor per thousand population. However, spatially, a large number of PHC doctors are concentrated in the plain areas near the main urban center. In contrast, some mountainous villages have only one village doctor, and a few villages even experience vacancies for village doctors. Secondly, the accessibility across the three districts shows a significant centripetal distribution (P < 0.01). Specifically, as depicted in Fig. 3, regions closer to the city center have higher accessibility, marked in green. Conversely, areas highlighted in red indicate a relative shortage of PHC human resources for residents. Lastly, impedance calculations for the baseline reveal that the annual per capita medical travel distance in the plain area is 16.96 km (P < 0.01), with a Gini coefficient of 0.368, indicating moderate fairness. In the near mountain area, the annual per capita impedance is 5.99 km (P < 0.05) with a Gini coefficient of 0.674, reflecting very poor fairness. In the deep mountain area, the annual per capita impedance is 4.76 km (P < 0.01) with a Gini coefficient of 0.493, indicating poor fairness.

Table 2.

Distribution of PHC human resources and description of PHC accessibility for villagers - baseline

Region	No. of villages	Population	Village clinics		Health stations		Township health centers		Accessibility	Impedance
		k persons	No. of clinics	No. of doctors	No. of stations	No. of doctors	No. of hospital	No. of doctors		km/(person·year)	Gini
Plain - Tongzhou	423	820	321	517	24	75	19	751	259.49**	16.96 **	0.368
Near mountainous - Mentougou	138	74	121	139	11	26	11	248	383.16**	5.99 *	0.674
Deep Mountain - Huairou	262	185	254	277	34	75	14	338	124.55**	4.76 **	0.493

Moran's I significance test ^*P<0.05，^**P<0.01

Fig. 3.

Spatial distribution map of PHC human resources and villager PHC accessibility– baseline

Configuration plan for the multi-strategy LA model

After constructing the multi-strategy LA model, we solved it by setting the government budget parameter p at 5%, 10%, and 15%, as shown in Table 3. The parameter p denotes the proportion of villages for which the government determines the budget based on the need to supplement village doctors. For instance, when p = 10%, the government’s budget is sufficient to support the addition of village doctors to 10% of the villages. Under the same budget constraint, the single-strategy model supplements PHC resources solely through the hiring of additional village doctors, while the multi-strategy model enables flexible selection among three strategies: supplementing village doctors, deploying mobile clinics, and establishing telemedicine stations.

Table 3.

Multi-Strategy PHC human resource LA model output configuration plan

Area	Manpower budget		Single-strategy model		Multi-strategy model
	p	Million	Hiring rural doctors	Cost (million)	Hiring rural doctors	Mobile clinics	Capsule clinics	Cost (million)
Plain	5%	1.015	21	1.008	8	0	21	0.624
	10%	2.030	42	2.016	21	0	34	1.226
	15%	3.046	63	3.024	34	0	47	1.720
Near mountainous	5%	0.331	6	0.288	3	2	3	0.288
	10%	0.662	13	0.624	7	4	6	0.623
	15%	0.994	20	0.960	10	5	11	0.953
Deep mountain	5%	0.629	13	0.624	10	2	3	1.005
	10%	1.258	26	1.248	17	3	11	2.013
	15%	1.886	32	1.536	24	9	10	3.021

Table 3 illustrates the comparative outcomes. Accessibility refers to the weighted sum of the supply-demand ratios of medical resources surrounding each village and is considered a positively oriented indicator. Healthcare impedance, on the other hand, refers to the average annual travel distance for villagers to seek medical care and is treated as a negatively oriented indicator. With the same budget level, the multi-strategy model covers more villages than the single-strategy model. For example, when p = 10% in Huairou District, the single-strategy plan covers 26 villages, while the multi-strategy plan reaches 31 villages—a 19.23% increase. Notably, the mobile clinics and telemedicine stations introduced in the multi-strategy model do not fully replace the hiring of village doctors; instead, they form complementary configurations. Taking Mentougou District as an example (with p = 10%), 41.18% of the sites adopt the traditional approach of hiring village doctors, 23.53% use mobile clinics, and 35.29% utilize telemedicine stations. These strategies are integrated to improve the efficiency of rural PHC resource allocation in a coordinated manner.

We further analyzed the output indicators of the two models, as shown in Table 4. First, the multi-strategy model demonstrates superior optimization efficiency. Under different budget parameters and across all regions, its accessibility outcomes are consistently better than those of the single-strategy model. For instance, when p = 10%, the multi-strategy model improves accessibility in the plains, near mountainous, and deep mountainous areas by 1.59%, 4.80%, and 4.44%, respectively—higher than the corresponding 1.37%, 4.08%, and 3.56% improvements achieved by the single-strategy model. Notably, even with only 5% investment, accessibility in the near mountainous and deep mountainous areas increases by 3.81% and 3.31%, indicating a stronger effect in underserved regions.

Table 4.

Comparison of output results between the two models

Area	Model	Baseline	Single-strategy model			Multi-strategy model
	p	-	5%	10%	15%	5%	10%	15%
Plain	Accessibility	259.49	262.56	263.05	263.16	263.05	263.62	265.28
	Impedance	16.96	16.16	15.69	15.32	16.14	15.66	15.29
	Gini	0.368	0.327	0.315	0.285	0.326	0.300	0.285
	Cost	-	1.008	2.016	3.024	1.005	2.013	3.021
Near mountainous	Accessibility	383.16	396.00	398.80	398.84	397.76	401.55	409.80
	Impedance	5.99	5.09	4.98	4.98	5.04	4.98	4.98
	Gini	0.674	0.440	0.380	0.378	0.392	0.378	0.378
	Cost	-	0.288	0.624	0.960	0.288	0.623	0.953
Deep mountain	Accessibility	124.55	128.53	128.98	129.98	128.67	130.08	135.22
	Impedance	4.76	3.85	3.80	3.80	3.83	3.80	3.80
	Gini	0.493	0.299	0.291	0.291	0.299	0.291	0.291
	Cost	-	0.624	1.248	1.536	0.624	1.226	1.720

Moreover, the impedance per unit investment in the multi-strategy model is consistently equal to or lower than that of the single-strategy model. In plain district, for example, with p = 15%, the single-strategy model reduces villagers’ average impedance by 542 m per million RMB, while the multi-strategy model achieves a reduction of 552 m, boosting investment efficiency by 1.93%. This highlights that a combined approach integrating village doctors, mobile clinics, and telemedicine yields higher returns and better accessibility than investing solely in village doctors.

In terms of equity, the multi-strategy model also significantly improves the fairness of healthcare resource distribution. Under the same budget levels, its Gini coefficients are consistently lower than those of the single-strategy model. At p = 10%, the Gini coefficient of impedance in the plains drops from 0.368 to 0.300, in the near mountainous area from 0.674 to 0.378, and in the deep mountainous area from 0.493 to 0.291. This evidence shows the model’s advantage in promoting equitable healthcare access.

Furthermore, the relationship between investment and impedance reduction shows diminishing marginal returns. With a small budget, accessibility can be greatly improved. For instance, in the near mountainous district under the multi-strategy model, p = 5% reduces average annual impedance from 5.99 km to 5.04 km (a 15.86% drop), and the Gini coefficient falls from 0.674 to 0.392, at a cost of just 288,000 RMB. However, as p increases, the marginal benefits in the plains decline gradually, while in mountainous areas, the improvements come to a complete halt after p = 10%. These findings suggest that targeted, needs-based resource allocation can significantly reduce impedance, while excessive or misaligned investment leads to diminishing returns, particularly in mountainous regions.

As mentioned earlier, the multi-strategy model demonstrates relatively better accessibility. However, when constructing the multi-strategy model, it assumes that mobile and capsule clinics provide comparable service quality to traditional clinics. To further assess the potential trade-offs between coverage expansion and service quality, we present a comparative analysis of accessibility indices across three dimensions: (a) accessibility under the single-strategy model, (b-1) aggregated accessibility under the multi-strategy model, and (b-2) traditional clinic accessibility under the multi-strategy model (Table 5).

Table 5.

Comparing overall and traditional clinic accessibility between the two models

Area	Baseline accessibility	p	(a) Accessibility under the single-strategy model	(b-1) Aggregated accessibility under the multi-strategy model	(b-2) Traditional clinic accessibility under the multi-strategy model
Plain	259.49	5%	262.56	263.05	261.31
		10%	263.05	263.62	261.39
		15%	263.16	265.28	261.43
Near mountainous	383.16	5%	396.00	397.76	389.41
		10%	398.80	401.55	389.59
		15%	398.84	409.80	389.65
Deep mountain	124.55	5%	128.53	128.67	126.40
		10%	128.98	130.08	127.02
		15%	129.98	135.22	127.42

Under different budget coefficients, the accessibility of the multi-strategy model is slightly higher than that of the single-strategy model. This improvement comes from the fact that, in the multi-strategy model, the opportunities originally intended for hiring village doctors to provide PHC services are partially reallocated to mobile and capsule clinics. For example, in the optimized results of the multi-strategy model with p = 5%, the traditional clinic accessibility in Plain, Near mountainous, and Deep Mountain areas is 261.31, 389.41, and 126.40, respectively. This indicates that the strategy of hiring village doctors in the multi-strategy model improves the baseline accessibility by 1.82, 6.25, and 1.85, respectively. In comparison, the single-strategy model improves it by 3.07, 12.84, and 3.98, respectively. The differences—1.25, 6.59, and 2.13—represent the opportunity costs incurred by substituting village doctors with mobile and capsule clinics. The corresponding return is the accessibility gain brought by mobile and capsule clinics in the multi-strategy model, which is 3.56, 14.60, and 4.12, respectively. Due to the difficulty in quantifying service quality across the three strategies, these potential trade-offs between coverage expansion and service quality require local policymakers to carefully weigh the pros and cons based on local conditions.

Spatial visualization of the model’s configuration plan

We selected the model output plans with p = 10% for spatial visualization, as shown in Fig. 4. Taking Fig. 4a as an example, we sequentially display the results for the plains area from baseline to the single-strategy model and the multi-strategy model. In the figure, the points represent villages and clinics. When a village’s medical demand exceeds the local clinic’s supply, villagers seek medical care across regions. The lines in the figure illustrate the flow of patients between regions, with darker colors indicating a higher number of cross-regional patients. The product of the number of non-local PHC visits and the spatial distance represents the total impedance of these patients seeking care outside their local area. The map at the top represents the baseline situation. The central map in Fig. 4a corresponds to the single-strategy plan, where the locations of hiring village doctors are marked with red ⊗. The map at the bottom corresponds to the multi-strategy plan, where the locations of different strategies are marked with different colored ⊗.

Fig. 4.

Spatial visualization of the configuration plan and flow of non-local PHC visits by patients

Figure 4 illustrates that the optimized locations for PHC workforce allocation primarily focus on town junctions and remote areas. By supplementing the workforce in these regions, extreme commuting in remote villages can be reduced, thereby lowering villagers’ medical impedance and enhancing medical equity. Taking Mentougou as an example, Fig. 4b shows that both models establish PHC clinics in Zhaitang Town. This addition prevents villagers in the southeastern part of the town from having to travel several kilometers to seek medical care in the Mentougou urban area. Furthermore, the establishment of clinics in Qingshui Town reduces local villagers’ reliance on township health centers and neighboring village clinics, effectively lowering medical impedance.

Regarding the selection of optimization strategies, each of the three strategies is suited to specific areas. Firstly, hiring village doctors remains indispensable for villages with large populations, significant PHC supply gaps, and scarce surrounding medical resources, as these villages require dedicated village doctors to meet their substantial medical demands. Secondly, for villages with smaller populations but located near township health centers, mobile clinics offer a flexible and cost-effective solution for supplementing the PHC workforce. Taking Mentougou as an example, Fig. 4b shows that two villages in Zhaitang Town and two villages in Bohai Town have relatively small PHC supply gaps and are located close to township health centers, making them ideal candidates for mobile clinics. Lastly, in villages with smaller PHC supply gaps but relatively remote geographical locations, capsule clinics serve as an efficient alternative to hiring village doctors. For instance, several villages in the eastern part of Tongzhou’s Xiji Town and four villages in Mentougou’s Qingshui Town, as shown in Fig. 4a and b respectively, are geographically remote but have insufficient population demand to support new village-level clinics. In these cases, deploying capsule clinics enhances the efficiency of the PHC workforce allocation.

Discussion

In the context of PHC workforce shortages in remote areas, this study proposes a multi-strategy PHC workforce allocation solution that combines hiring village doctors, deploying mobile clinics, and establishing capsule clinics. We developed a multi-strategy PHC Workforce LA model and conducted empirical runs using Beijing as a case study. The results show that integrating village doctors, mobile clinics, and capsule clinics can significantly improve the efficiency and accessibility of primary healthcare resource allocation in rural areas. Across different budget levels, this model performs well in enhancing accessibility and reducing impedance, highlighting its low dependence on healthcare personnel, which makes it well-suited to address the diverse needs of rural healthcare systems. Notably, in resource-scarce mountainous regions, even small-scale investments (e.g., p = 5%) can yield substantial improvements, suggesting that the multi-strategy model holds unique advantages in addressing healthcare accessibility issues in underserved areas. These improvements stem not only from the additive effects of multiple service strategies but also from the synergy among them.

However, the improvement in healthcare accessibility does not continue indefinitely. The findings reveal a diminishing marginal return on government investment in reducing travel impedance, particularly in sparsely populated regions where the supply–demand gap is relatively small. In our case study of mountainous areas in Beijing, as p increases, the marginal benefits in the plains decline gradually, while in mountainous areas, the improvements come to a complete halt after p = 10%. This is due to a relatively small coverage gap in rural medical infrastructure: since around 2019, Beijing has largely eliminated ‘blank’ villages through large-scale investments. As a result, many areas have already met the model’s coverage constraints, such as “all villages are covered” and “all healthcare demands are met,” limiting further reductions in impedance. In contrast, the plains, which have denser populations and larger service gaps, still show potential for improved accessibility with increased investment. These findings suggest a shift in resource allocation priorities—from broad expansion to more targeted strategies based on actual supply–demand gaps, rather than adopting a one-size-fits-all or indiscriminate expansion approach.

It is worth noting that while the multi-strategy model demonstrates higher overall accessibility, these gains are achieved by substituting traditional village doctor services with mobile and capsule clinics—incurring notable opportunity costs. Specifically, reallocating resources from hiring village doctors to deploying mobile and capsule clinics leads to a decline in traditional clinic accessibility within the multi-strategy model. For example, in the case of a 5% budget coefficient, traditional clinic accessibility in the Plains, Near Mountainous, and Deep Mountain areas is significantly lower under the multi-strategy model than under the single-strategy model, with respective gaps of 1.25, 6.59, and 2.13. These differences represent the opportunity costs of replacing village doctors with non-traditional service modalities. Although such substitution results in net accessibility gains through mobile and capsule clinics, this trade-off raises concerns about potential compromises in service quality—an aspect not directly captured by current accessibility metrics. Given the uncertain clinical effectiveness and varying levels of patient trust in these alternative modalities, policymakers should approach resource reallocation with caution, balancing the quantitative expansion of coverage against the qualitative integrity of care.

In summary, this study demonstrates the potential of integrating digital health tools with traditional workforce strategies. The “Hiring + Mobile + Capsule” strategy has proven highly efficient in utilizing government budgets, enabling more precise allocation of PHC personnel in rural areas. Compared to other PHC resource LA studies, our research presents a novel multi-strategy approach for PHC workforce allocation. We introduced two digital health tools: capsule clinics and mobile clinics, which allow for more efficient allocation of PHC resources. This aspect has not been addressed in previous studies [47, 48]. Additionally, our research delves into the administrative village level, providing more detailed and precise solutions for PHC workforce allocation compared to studies at the township level [47–49]. In summary, this study introduces capsule clinics and mobile clinics to enhance PHC delivery in rural areas, offering methodological insights for optimizing PHC workforce distribution. By focusing on the administrative village level, it provides detailed and accurate solutions for PHC workforce allocation, demonstrating significant theoretical and practical value.

Three types of optimization strategy adaptability analysis

The study also indicates that different workforce supplementation strategies are suitable under varying conditions. Based on the model run results in Beijing’s suburban areas, we summarize the applicability of the three strategies—hiring village doctors, deploying mobile clinics, and establishing capsule clinics—in different regions. First, the hiring village doctors shows clear advantages in villages with larger populations, significant PHC supply gaps, and a scarcity of surrounding medical resources. These areas typically face high medical demands, but existing medical facilities cannot meet them. By increasing the number of village doctors, medical service gaps can be directly filled, enhancing the accessibility of medical services for villagers. Second, the mobile clinics strategy is suitable for villages with smaller populations and closer proximity to township health centers. These areas have lower and fluctuating medical demands. The flexibility of mobile clinics allows for dynamic adjustment of service frequency based on demand, thereby effectively utilizing resources and reducing unnecessary investment and construction costs. Finally, the capsule clinics strategy demonstrates high efficiency in villages with smaller PHC supply gaps but relatively remote geographic locations. Compared to building new village health centers, deploying capsule clinics incurs lower costs and shorter construction periods, enabling rapid fulfillment of local residents’ medical needs. Additionally, by leveraging telemedicine, capsule clinics circumvent the issue of village doctor shortages, thereby enhancing the feasibility of policy implementation. In summary, these three optimization strategies exhibit differentiated conditions for their use, and these abstract characteristics will help PHC policymakers make initial decisions on supplementary strategies [50].

Discussion on the scalability of the “Hiring + Mobile + Capsule” strategy

Comprehensive analysis indicates that the “Hiring + Mobile + Capsule” strategy can be used complementarily under different conditions, achieving a “flexible and stable” PHC workforce allocation model [36, 37]. In practice, cases from various countries beyond China demonstrate the scalability of mobile and capsule clinics. In regions facing extreme shortages of medical resources, these clinics often serve as the most accessible—or even the only—available healthcare option due to their low cost and flexibility. For example, in crisis-affected countries such as Syria, Iraq, and Ukraine, mobile clinics provided by the WHO delivered essential medical services [31]. Similarly, public health nurse-led mobile teams reached underserved rural areas in Haiti [51], while a web-based telemedicine solution supported tuberculosis care in remote communities in Papua New Guinea with limited clinical infrastructure [52]. These examples illustrate the adaptability of mobile and telemedicine-enabled clinics in meeting urgent healthcare needs under severe constraints. In relatively well-resourced areas, mobile and capsule clinics play a complementary role within the broader healthcare system, offering specialized or flexible services. This adaptability also makes them suitable for highly privatized healthcare systems like that of the United States. For instance, during the COVID-19 pandemic, mobile clinics helped alleviate primary healthcare workforce shortages in rural southern Minnesota [53]. In parallel, capsule clinics located within pharmacies provided direct-to-consumer telemedicine services, enhancing access to both routine and urgent care within a market-based context [38]. These cases collectively highlight the potential for broader application of mobile and capsule clinic models. However, their successful implementation hinges on alignment with local healthcare system structures and financing mechanisms. Tailoring the model to fit local contexts is essential for long-term sustainability.

In terms of weaknesses, compared to traditional clinics, mobile and capsule clinics may have limitations in equipment and technological capabilities, particularly in handling complex diseases. Studies have pointed out that while mobile clinics have certain advantages in improving the accessibility of PHC, their service scope and quality still fall short compared to traditional clinics [26, 54, 55]. Similarly, although telemedicine and capsule clinics show high patient satisfaction, their service scope and quality remain limited, and they cannot effectively replace traditional in-person clinics [35, 56–58]. Interviews with township health center administrators indicated that although the physicians at mobile and capsule clinics are generally more qualified than village doctors at village clinics, patients often rate their perceived treatment efficacy and satisfaction lower than those for village doctors. This is mainly due to objective limitations such as insufficient diagnostic equipment, off-site consultations, and incomplete drug formularies. Another study conducted at the same field site also suggests that patients’ perceived usefulness, willingness to use, and satisfaction with these new types of clinics still need to be improved [59]. In addition to concerns about medical quality, interviews also revealed that high personnel mobility can affect the continuity of medical services. Achieving effective coordination between township health centers and clinics with limited resources remains a pressing management challenge. This issue is further supported by other studies and reports [26, 60]. In conclusion, the limitations of multi-strategy approaches in areas like healthcare quality will hinder their wider adoption. Balancing the accessibility of PHC with high-quality medical services will be a central consideration for PHC management authorities.

Regarding opportunities, advancements in medical and communication technologies enable mobile and capsule clinics to integrate with emerging technologies such as telemedicine and AI-assisted diagnostics, further improving the accuracy and efficiency of remote medical consultations. Recent studies have shown that GPT-4 achieved a diagnostic accuracy of 87.9% across 557 complex adult case reports [61], and some large language models have already demonstrated performance on par with attending physicians, indicating their potential to assist or even replace village doctors in providing PHC services [62]. Furthermore, AI-generated responses have been evaluated as more accurate and empathetic than those of human doctors, gaining satisfaction from both patients [63] and clinicians [64]. At the product level, China is also exploring the development of Autonomous Mobile Clinics (AMCs). This initiative integrates autonomous driving, telemedicine, and AI-driven healthcare to upgrade capsule clinics into self-driving mobile units equipped with AI doctors. It aims to alleviate the physician burnout often associated with mobile clinics while addressing the underutilization of diagnostic and treatment equipment in capsule clinics. This innovation could significantly reshape PHC delivery strategies, building upon current mobile and capsule clinic models [65, 66].

However, several challenges must be addressed to ensure the success of mobile and capsule clinics. First, regulatory and institutional integration remains complex. These new models must comply with existing healthcare regulations, meet medical standards, and be eligible for reimbursement. Yet, current legal frameworks in many countries do not fully support them. In China, policy delays during pilot phases led to reimbursement issues, which hindered adoption. Health insurance policies heavily influence residents’ medical choices. Similar problems have occurred globally. In the U.S., telehealth surged during COVID-19, but uneven reimbursement across states created disparities in access [67]. Similarly, the NHS in the UK faced challenges due to poorly defined telemedicine reimbursement policies, which created uncertainty for both healthcare providers and patients [68]. Second, ensuring financial sustainability is critical [22]. In India, Jignesh Patel et al. emphasized that the affordability of optimized strategies is key to the long-term viability of PHC clinics [69]. Conversely, the closure of HealthSpot in the U.S. illustrates the risks of unsustainable business models [36]. In China, mobile and capsule clinics are cost-effective under the current PHC system due to lower labor and infrastructure costs, but they rely heavily on government funding. When applying these experiences to other countries, it is crucial to reconsider the financial sustainability of PHC optimization strategies, especially with the need for private sector involvement alongside government funding. An example is the Family Van mobile clinic in the U.S., which improved its competitiveness through cost control and return-on-investment calculations. It also ensured financial sustainability by combining government funding (grants, public health initiatives, and community health programs) with private sector funds (philanthropy, insurance reimbursements, and business partnerships). This diversified funding approach helps maintain long-term viability [70]. Such models could be adopted in other countries to make mobile healthcare services more sustainable and adaptable to different healthcare systems. Finally, patient acceptance and cultural barriers pose significant challenges. In rural areas, traditional care models are deeply rooted, and unfamiliar formats may be met with skepticism. The digital divide further limits access among rural residents, the elderly, low-income groups, and ethnic minorities [71–76]. These barriers reduce utilization and hinder widespread adoption of multi-strategy models.

Policy recommendations

First, given the proven efficiency of the multi-strategy model, policymakers may try integrating mobile and capsule clinics as supplements to traditional PHC workforce deployment to enhance rural healthcare accessibility. By leveraging the unique strengths of each service modality, governments can achieve broader coverage without overextending limited resources. Spatial visualization results further suggest that policymakers can make preliminary decisions based on the abstract applicability of the three strategies. Specifically, hiring village doctors is most suitable for villages with large populations, significant PHC shortages, and limited access to surrounding medical resources. Mobile clinics are ideal for smaller villages near township centers with low and fluctuating medical demand. Capsule clinics offer a cost-effective solution for geographically remote areas with moderate service needs. Importantly, resource inputs must be tailored to actual supply–demand gaps. Investments beyond a certain threshold, especially in mountainous areas with near-universal coverage, yield diminishing returns. Thus, instead of pursuing uniform expansion, governments should shift toward precise supplementation strategies that balance efficiency and service quality. Caution is warranted when reallocating resources from traditional clinics, as non-traditional modalities may impact perceived care quality. Ultimately, adopting differentiated supplementation pathways aligned with local needs will maximize the impact of PHC investments.

Second, ensuring the financial sustainability of mobile and capsule clinics is key to scaling up the multi-strategy PHC model. While China’s PHC is primarily government-funded, many countries rely on insurance or public-private partnerships. Diversifying funding sources spreads risk and strengthens resilience. To this end, governments should actively promote blended financing strategies, combining public investment, private insurance reimbursements, corporate sponsorships, and philanthropic donations. Crucially, achieving sustainable multi-source financing requires aligning the interests of different stakeholders to create mutual benefits. For example, insurers may benefit from reduced long-term healthcare costs through preventive services; corporations (such as pharmaceutical, tech, or logistics companies) can gain branding opportunities, data insights, or entry into new markets; while NGOs and foundations can fulfill their social mission by improving access to care. Such win-win collaborations are already in practice. The U.S. “Family Van” program combines public funding, insurance reimbursements, and donations to sustain mobile services [70]. The FQHC Telehealth Consortium in Massachusetts blended HRSA funding, Medicaid support, and philanthropic grants to rapidly scale services during COVID-19 [77]. These examples highlight the value of multi-source financing in expanding coverage and ensuring long-term viability. Governments can further incentivize collaboration through co-financing, tax relief, or performance-based payments, building a more resilient PHC delivery system.

Third, promoting public acceptance and technological integration is essential to the success of mobile and capsule clinics, particularly in rural and underserved regions. Governments should support community-based education and outreach to raise awareness about the safety, convenience, and quality of these healthcare models. Partnering with local leaders and healthcare advocates can foster trust and acceptance, while training healthcare providers to clearly communicate the benefits of mobile services, including telemedicine, can further enhance patient confidence and utilization [72, 75, 76]. At the same time, bridging the digital divide remains critical. Policymakers must ensure marginalized groups, such as the elderly, low-income individuals, and rural residents, have access to the necessary technology and infrastructure [71, 73, 74]. In parallel, governments should encourage collaboration between healthcare providers and tech companies specializing in telemedicine, AI, and diagnostics to enhance service accuracy and efficiency. Establishing standardized telehealth protocols—covering consultations, diagnostics, and patient management—will help ensure consistent, high-quality service across platforms. These efforts, taken together, will strengthen both the public’s willingness to adopt new care models and the technical capacity of mobile and capsule clinics to meet medical standards.

Limitation

This study has several limitations. First, due to the relatively short pilot period of mobile and capsule clinics, the actual operational efficiency of mobile clinics and capsule clinics, as well as their impact on patients’ health outcomes, requires further long-term follow-up assessments. Second, our analysis is primarily based on data from Beijing; therefore, the applicability of our conclusions to other regions remains to be verified. The specific details of policy implementation and the coordination of primary healthcare resources need an in-depth discussion that considers the unique circumstances of different regions. Additionally, this study assumes that the service quality and diagnostic capabilities of village doctors, mobile clinics, and capsule clinics are relatively homogeneous. However, actual differences may exist, which should be addressed in future research. Finally, this study mainly focuses on the accessibility of PHC human resources. Patients’ preferences for medical care and their influence on healthcare service selection were not considered, and the model could be further refined in future research.

Conclusions

This study proposes a multi-strategy spatial allocation method for the PHC workforce that combines hiring village doctors, deploying mobile clinics, and establishing capsule clinics. Through empirical runs conducted in Beijing, we validated the effectiveness of this multi-strategy LA model in addressing resource distribution imbalances and enhancing the accessibility of medical services. The findings indicate that multi-strategy allocation significantly reduces the average annual travel distance (Impedance) for villagers seeking medical care, thereby improving both the accessibility and equity of medical services. Under the same budget, the multi-strategy model demonstrated higher financial efficiency and rationality compared to the single-strategy models. Based on these results, this study recommends that the government adopt a comprehensive approach in PHC resource allocation by integrating the strategies of hiring village doctors, deploying mobile clinics, and establishing capsule clinics, especially in remote and resource-constrained areas.

This research contributes to academia by presenting a novel multi-strategy approach to PHC workforce allocation, incorporating digital health tools such as mobile and capsule clinics alongside traditional strategies. The study provides more precise, administrative village-level solutions, improving on previous research that focused primarily on township-level models. For policymakers, the findings offer valuable insights into the potential benefits of integrating mobile and capsule clinics to address workforce shortages and optimize healthcare resource distribution in rural areas. Finally, from a societal perspective, this study emphasizes how adopting a multi-strategy approach can significantly improve healthcare accessibility, particularly for rural and underserved populations, and reduce healthcare disparities. These contributions offer a comprehensive framework for addressing global challenges in PHC resource allocation, especially in regions facing similar workforce shortages.

Abbreviations

WHO

World Health Organization

PHC

Primary Health Care

LA

Location Allocation

GIS

Geographic Information System

PMM

P Median Model

2SFCA

Two-step Floating Catchment Area

COVID-19

Coronavirus Disease 2019

NHS

National Health Service

AMCs

Ambulatory Medical Clinics

AI

Artificial Intelligence

Authors’ contributions

HL was responsible for writing the manuscript and performing the data analysis. CM contributed to the study’s conception and design, revised the manuscript and oversaw the overall content as the corresponding author. YY, WL, and YW assisted with the surveys.

Funding

This work was supported by the R&D Program of Beijing Municipal Education Commission (SZ202210025011).

Data availability

The datasets used during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Capital Medical University with the approval number"Z2022SY075”. Additionally, this study received an exemption from further ethical review by the Ethics Committee of Capital Medical University due to the use of publicly available official data.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.