Sage Journals: Discover world-class research

Abstract

In China, the absence of a strict referral system leads to overcrowding in tertiary hospitals and underutilization of primary care, challenging the implementation of a tiered diagnosis and treatment system. This study investigates how differential health insurance reimbursement rates can be leveraged to guide patient flow. We developed an agent-based simulation model, calibrated with real-world data on hypertensive inpatients from Beijing (2015-2017). The model simulated patient healthcare-seeking choices under various reimbursement scenarios for primary, secondary, and tertiary hospitals. Adjusting reimbursement rates significantly redirected patient flow and altered medical insurance fund expenditure. Increasing the reimbursement rate for primary hospitals proved particularly effective in channeling patients toward them. We identified an optimal balance—maximizing the proportion of hypertensive inpatients seeking care at primary hospitals (achieving 36.41%, compared to an observed baseline of 16.39%) while controlling total fund expenditure—when reimbursement rates were set at 95% for primary, 85% for secondary, and 70% for tertiary hospitals. Differential health insurance reimbursement serves as a potent policy lever to steer patient flow within a tiered healthcare system. The agent-based modeling approach provides a valuable tool for policymakers to simulate and identify optimal reimbursement strategies, thereby promoting efficient healthcare delivery and sustainable fund management.

Keywords

inpatients health care health expenditures reimbursement China

Introduction

Hierarchical Diagnosis and Treatment and Medical Insurance Payment

A defining feature of China’s healthcare landscape is the considerable freedom patients enjoy in choosing their healthcare providers. The absence of a strict mandatory referral system allows patients to directly access tertiary hospitals, bypassing primary care facilities. This autonomy, however, presents a double-edged sword: while empowering individual choice, it has resulted in severe overcrowding at top-tier hospitals and chronic underutilization of primary care institutions. Consequently, establishing and promoting an effective hierarchical diagnosis and treatment system has become a cornerstone of China’s ongoing healthcare reform, with its success largely hinging on the effective implementation of this tiered structure.

Recognizing this challenge, the State Council General Office issued the “Guidance on Promoting the Construction of Hierarchical Diagnosis and Treatment” in 2015. This policy proposed establishing a tiered system by starting with common and chronic diseases, aiming to guide patient flow through differentiated medical insurance payment policies across different levels of medical institutions.¹ To further cultivate a rational medical treatment order, the National 14th Five-Year Plan emphasized the implementation of differentiated medical insurance payment and pricing policies.² As an economic lever to modify residents’ healthcare-seeking behavior, medical insurance payment policy design plays a pivotal role in advancing hierarchical diagnosis and treatment in China.^3,4

Subsequent policy documents, including the 2017 “Guiding Opinions on Further Deepening the Reform of the Basic Medical Insurance Payment Method”⁵ and the 2020 “Opinions on Deepening the Reform of Medical Insurance System,”⁶ have called for more sophisticated management and a manageable, efficient payment mechanism. These developments impose stricter requirements on China’s medical insurance payment system reform. A central and challenging issue remains: how can medical insurance payment reforms effectively incentivize hierarchical diagnosis and treatment?

Medical Insurance Policies and Patient Choice of Medical Care

The unique characteristics of medical services distinguish patient demand and utilization behavior from that of ordinary goods.⁷ The study of healthcare demand behavior originated from utility theory, analyzing the maximization of utility function under income constraints. Grossman’s health human capital model and the inclusion of health insurance factors by Newhouse and Phelps’ are classic frameworks for analyzing factors influencing medical demand.⁸ Andersen further proposed that factors influencing medical demand include propensity characteristics, enabling resources, and need, providing a robust analytical framework for understanding patients’ healthcare choices.⁹ This model has been revised over the years, offering a basis for explaining the influence of various factors on healthcare choices and their interactions.

Among the numerous factors influencing residents’ healthcare behavior, key determinants include gender, income, age, and distance.^10
-13 Crucially, effective medical insurance policies can facilitate patient access to care.^14
-16 When the out-of-pocket payment decreases, the consumption of medical services increases significantly,¹⁷ a trend supported by empirical data from Taiwan.¹⁸

While it is well-established that medical insurance stimulates residents’ healthcare demand, a more nuanced question arises: does medical insurance policy specifically affect residents’ demand behavior for different levels of care?

Residents’ choice of medical institutions is fundamental to the hierarchical diagnosis and treatment system. Some studies have found that medical insurance policies can influence patient hospital choice by increasing the cost disparity between hospitals of different levels.¹⁹ Research from China indicates that reimbursement rates adjusting for different levels affect communitywide care choices.²⁰ Another Chinese study suggested that varying reimbursement rates influence the decision to seek treatment, with increases for primary hospitals boosting their patient share.²¹

Differential reimbursement rates create a tiered structure of out-of-pocket costs for patients. When seeking care for a comparable condition, the financial burden is significantly lower at primary hospitals compared to tertiary ones due to the higher reimbursement rate.^17,18,21 This cost disparity serves as a powerful economic incentive, making the policy-preferred option—visiting primary hospitals—also the most financially rational choice for patients, thereby channeling patient flow in a way that supports hierarchical diagnosis.

Globally, adjustments to medical insurance policies have been used to influence healthcare utilization. In the United States, commercial insurers have implemented “gatekeeper” systems through differentiated reimbursement packages.²² In France, patients can choose whether to participate in a social insurance-based “gatekeeper” system, with the government providing different reimbursement rates for different visit types.²³ Australia use different reimbursement rates for general practitioners and specialists to guide patient flow.

Agent-Based Modeling

Medical insurance policy creation and implementation are highly uncertain and complex processes involving various stakeholders, including patients, physicians, and hospitals. Traditional qualitative and quantitative research methods possess inherent limitations in capturing this complexity. In contrast, simulation methods, being experimental, process-oriented, and future-focused, offer a fresh perspective on health policy research and have recently gained prominence.

Agent-based modeling and simulation (ABMS) is a computational microsimulation technique that involves modeling a system’s individuals(agents) and developing the system’s overall behavior through their interactions.

In 2017, The Lancet published an expert review calling for a complexity modeling approaches to address global public health challenges, highlighting the important role of ABMS.²⁴ Currently, ABMS is widely used in public health to model infectious disease epidemics,^25
-27 chronic diseases,^28,29 health behaviors,^30
-32 healthcare delivery,^33,34 and health economics.³⁵

Several studies have explored the impact of insurance policies on patients’ healthcare choices using ABMS, including the influence of costs,³⁶ comparisons of outpatient reimbursement options,³⁷ the effects of accountable care organizations,³⁸ and the impact of widening cost gaps between hospital levels.¹⁹ Integrating the ABMS into medical insurance policy analysis can help identify impact patterns, forecast the macro and micro effects of potential policy adjustments, and thus facilitate more scientific, and effective policy formulation.

A study by Tan et al used ABMS to adjust reimbursement rate for different hospital levels separately and found the impact of differential medical insurance reimbursement on inpatient chronic disease residents to be limited.³⁹ This suggests that isolated adjustments to the reimbursement rate for a single hospital level have a limited effect on guiding patient distribution. Thus, a significant and underexplored research gap exists in determining how to systematically and combinatorially adjust rates across multiple hospital levels to identify the optimal “reimbursement combination” that best balances patient flow with medical insurance fund control.

This study selected hypertensive inpatients as the research population for the following reasons: First, hypertension is a common chronic condition and a key focus within the tiered diagnosis and treatment policy. Second, its disease severity has clear classification standards (ICD-10), facilitating the precise characterization of patient heterogeneity within the model. Finally, the relevant data possesses good accessibility, providing a reliable foundation for model construction and validation.

Compared to previous studies, this research introduces key innovations in both model design and research focus. While the ABMS studies^19,20 concentrated on the impact of cost disparities or a single type of insurance, and Tan³⁹ was confined to adjusting the reimbursement rates of hospital levels in isolation, this study aims to synchronously and coordinately adjust the medical insurance reimbursement rates for primary, secondary, and tertiary hospitals. It dynamically observes changes in patient flow and medical insurance funds under this complex scenario that more closely mirrors real-world policy, with the goal of identifying synergistic effects and optimal policy configurations that cannot be detected through single-policy adjustments.

Based on the above rationale, this study aims to construct a real-world data-based ABMS model to simulate the healthcare-seeking behavior of hypertensive inpatients and its impact on medical insurance fund expenditure under different medical insurance reimbursement rate combinations. We propose the following core hypothesis: Synchronously adjusting the medical insurance reimbursement rates across multiple hospital levels can effectively guide patient flow, and there exists an optimal reimbursement rate combination that maximizes the proportion of patients seeking care at primary hospitals while simultaneously ensuring the controllability and balance of medical insurance fund expenditure. The results of this study are intended to provide a quantitative decision-making basis for the refined reform of medical insurance payment methods.

Materials and Methods

Study Design and Data Sources

This was an agent-based simulation study utilizing real-world data from Beijing, China, covering the period from 2015 to 2017. The study population consisted of hypertensive inpatients enrolled in the Beijing urban residents’ basic medical insurance, screened according to the following criteria: ① coverage under the aforementioned insurance, ② inpatients status in primary, secondary, or tertiary hospitals in Beijing, ③ age over 35 years.

The simulation was initialized with 2165 agent residents, a number that represents a 1:10,000 scale-down of the actual Beijing resident population during the 2015 to 2017 period. A total of 65 hospitals were included based on the 2015 to 2017 Beijing Health and Family Planning Development Statistics Bulletin, comprising 40 primary, 17 secondary, and 8 tertiary hospitals.

Basic demographic data for Beijing residents were sourced from the Beijing Municipal Bureau of Statistics. The Information Center of the Beijing Municipal Health Commission provided retrospective, anonymized, individual-level discharge records for hypertensive inpatients from Level II and above hospitals in Beijing. Patient records with a primary diagnosis of hypertension (ICD-10 codes: I10xx02, I10xx03, I10xx04, I10xx05) were screened to extract information on admissions, total costs, out-of-pocket costs, and admissions by region and age. Data for primary care institutions were supplemented from 3 district-level social management centers and several community health service centers in Beijing.

This study was reported in accordance with the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) guideline.⁴⁰

Modeling Approach and Justification

Unlike traditional econometric methods that often identify average effects from historical data, ABMS is particularly suited for modeling complex, dynamic systems involving heterogeneous individuals. Its advantage lies in capturing individual heterogeneity, simulating interactions between agents (eg, patients, hospitals), and conducting counterfactual policy experiments to observe emergent system-level outcomes, such as shifts in patient flow and insurance fund expenditure under novel reimbursement combinations.^41
-43

Study Assumptions

The following assumptions were made, informed by the ABMS literature and the practical context of patient visits.

H1: No demographic changes (eg, birth, death, migration) during the simulation period.

H2: Study limited to inpatients enrolled in the urban residents’ basic medical insurance scheme.

H3: Only inpatient behaviors are modeled; outpatient visits and referrals are excluded.

H4: All hypertension treatment costs are covered by the urban residents’ medical insurance.

H5: Hypertension severity is classified into 3 levels (ICD-10: I10xx03–I10xx05) based on diagnostic criteria.

These assumptions are interconnected and serve to define the scope and focus of our model. The stability of the study population (H1) and the specific insurance coverage (H2) together establish a well-defined cohort for examining policy effects. Focusing solely on inpatient care (H3) while assuming full insurance coverage (H4) allows us to isolate the role of reimbursement rates in patients’ decision-making. The standardized classification of hypertension severity (H5) introduces meaningful variation among patients, enabling us to examine how disease severity might influence choices when faced with different reimbursement rates. While these choices necessarily simplify the real-world context, they provide a coherent foundation for examining our core research question regarding how combinations of reimbursement rates might affect patient flow.

Model Design

The simulation system integrated 4 interconnected sub-models: patients, hospitals, diseases, and policies (Figure 1). The patient and hospital models were parametrized using empirical data, while the disease and policy models were constructed based on established literature and clinical guidelines.^44,45

Figure 1.

Schematic of the system design, including the following 4 models: patients, hospitals, diseases, and policies.

The core agent decision-making process is illustrated in Figure 2. A key behavioral mechanism is the distance tolerance coefficient (K), which modulates a patient’s sensitivity to travel distance based on disease severity. This design is grounded in the empirically observed tendency that patients with more severe conditions prioritize clinical capability over proximity.^46
-49 Consequently, in our model, K increases with hypertension severity (Level 1: K = 1; Level 2: K = 2; Level 3: K = 3), effectively expanding the acceptable travel range for more severe cases.

Figure 2.

Model structure representing the general flow of patients in the simulation. *Dcost: The average cost of medical care for hypertension.

Hospital choice was modeled using a preference index (PI), quantified as a decreasing function of the out-of-pocket cost ratio between high-level (X) and low-level (Y) hospitals:

P I = - 0.25 \times \frac{C_{X}}{C_{Y}} + 1.25

where $C_{X} = M_{X} \times (1 - P_{X})$ and $C_{Y} = M_{Y} \times (1 - P_{Y})$ represent the out-of-pocket costs at hospital tiers X and Y, respectively. The function was calibrated such that PI(1) = 1 and PI(3) = 0.5, meaning patients are half as likely to choose a high-level hospital when its out-of-pocket cost is 3 times that of a lower-level 1, a sensitivity parameter informed by prior discrete choice experiments in healthcare.⁵⁰

The model was implemented using AnyLogic.

Statistical Analysis

The data analysis in this study encompassed both descriptive statistics and model validation metrics. Descriptive statistics were used to summarize the characteristics of the patient population and hospital resources. For model validation, the primary statistical measures were the Relative Error (RE) and Root Mean Square Error (RMSE), calculated to quantify the discrepancy between simulated outputs and observed real-world data.

The formulas for these calculations are as follows:

R E = \frac{| y_{i j} - {\hat{y}}_{\begin{matrix} i j \end{matrix}} |}{{\hat{y}}_{\begin{matrix} i j \end{matrix}}}

RMSE = \sqrt{\frac{1}{n} \sum_{i = 1}^{n} {(y_{i j} - {\hat{y}}_{\begin{matrix} i j \end{matrix}})}^{2}}

Here, $y_{i j}$ and ${\hat{y}}_{\begin{matrix} i j \end{matrix}}$ denote the simulated and actual values for the i-th data point, respectively. A model is considered to have good predictive performance if over 70% of the compared variables exhibit an RE of less than 5%, and the RE for any single variable does not exceed 10%, with a complementary RMSE also below 10%.¹⁹

As shown in Table 1, all validation metrics fell within these acceptable ranges, confirming the model’s predictive accuracy.

Table 1.

Distribution of the Proportion of Hospital Visits and Tests by Level.

Hospital access rate	Real data	Simulation data	RE	RMSE
Primary hospital	16.39%	17.21%	0.041	0.014
Secondary and tertiary hospitals	83.61%	82.79%	0.009	0.014

Results

Overview of Key Findings

We conducted a total of 126 simulation runs, organized into 7 experimental sets. Each set involved adjusting reimbursement rates and was run 6 times with incremental 2% changes. To minimize the impact of random factors, the “random seed” option was selected under the “randomness” setting. Each specific alteration was tested 3 times, and the final outcome for a simulation experiment was determined by averaging these 3 results. Figures 3 and 5 illustrate changes in patients’ healthcare choices after adjusting reimbursement rates for different hospital levels separately or simultaneously, while Figures 4 and 6 show corresponding changes in the growth rate of medical insurance fund expenditure at each hospital level.

Figure 3.

Changes to patient access options brought about by adjusting reimbursement rates for different levels of hospitals.

Figure 4.

Changes in the growth of the medical insurance fund brought by adjusting reimbursement rates for different levels of hospitals.

Figure 5.

Changes to patient access options brought by adjusting reimbursement rates for different levels of hospitals.

Figure 6.

Changes in the growth of the medical insurance fund brought about by adjusting reimbursement rates for different levels of hospitals. *Q1 through Q6 correspond to the reimbursement rates after each 2% increase, with Q1 being the initial rate of 85% for primary and secondary hospitals and 70% for tertiary hospitals.

Our simulation experiments demonstrated that synchronously adjusting medical insurance reimbursement rates across hospital tiers effectively redirected patient flow and altered the distribution of medical insurance fund expenditure. The most significant steering effect was achieved by increasing the reimbursement rate for primary hospitals, which consistently channeled a greater proportion of hypertensive inpatients toward primary care. An optimal balance—maximizing primary hospital utilization while maintaining control over total fund growth—was identified when reimbursement rates were set at 95% for primary, 85% for secondary, and 70% for tertiary hospitals.

The results in Figure 3 demonstrate that increasing the reimbursement rate for primary hospitals is the most potent single policy lever for redirecting patients from tertiary to primary care. For policymakers, this suggests that investing financial incentives at the grassroots level is a more direct and effective strategy to relieve overcrowding in top-tier hospitals than disincentivizing tertiary care alone.

Figure 4 highlights a critical financial insight: steering patient flow toward lower-cost primary and secondary hospitals is key to ensuring the financial sustainability of the medical insurance fund. This demonstrates that the policy goal of hierarchical diagnosis is not only about patient distribution but also an essential mechanism for long-term cost control.

The synergistic effect observed in Figure 5 underscores the importance of combinatorial policy design. Adjusting reimbursement rates at multiple levels simultaneously (eg, raising primary while lowering tertiary) creates a powerful dynamic that is more effective than any single adjustment. Policymakers should therefore consider systemic reimbursement packages rather than isolated rate changes.

As shown in Figure 6, the fiscal impact on the insurance fund varies significantly depending on the combination of reimbursement rates. This allows policymakers to calibrate and fine-tune multi-level policies, finding a balance that achieves the desired patient redistribution without placing undue stress on the fund.

Patient Flow in Response to Single-Level Reimbursement Adjustments

Adjusting the reimbursement rate for each hospital level individually revealed distinct patterns in patient steering.

As the reimbursement rate for primary hospitals increased, the proportion of patients choosing them rose steadily. This confirms that financial incentives are a potent tool for making primary care more attractive to patients. Increasing the reimbursement for secondary hospitals caused a dramatic shift, attracting patients primarily from tertiary hospitals rather than primary ones. This suggests that secondary hospitals serve as a viable and preferred alternative to tertiary facilities when the financial incentive is favorable. An increase in tertiary hospital reimbursement rates resulted in only a minor uptick in their patient share, but significantly increased the financial burden on the insurance fund, indicating poor cost-effectiveness.

The Synergistic Effect of Multi-Level Reimbursement Adjustments

Simultaneous adjustments revealed synergistic effects that were not apparent from single-level experiments. For instance, concurrently increasing primary reimbursement while decreasing tertiary reimbursement (eg, Primary: 95%, Tertiary: 70%) produced a strong effect, efficiently shifting patients from tertiary to primary care. In contrast, adjusting only secondary and tertiary rates together merely redistributed patients between these 2 levels without effectively promoting primary care utilization.

Identification of an Optimal Reimbursement Combination

Through 126 simulation experiments, we identified several reimbursement combinations that improved patient distribution. The most optimal combination, which achieved the highest flow to primary hospitals while ensuring sustainable fund expenditure, is presented in Table 2.

Table 2.

Optimal Reimbursement Combinations for Steering Hypertensive Inpatients.

Experimental groups	Adjusted hospital level(s)	Optimal reimbursement rate(%)	Reimbursement for other levels(%)	Optimal access rate to primary hospitals(%)
1*	Primary	95	85, 70	36.41
2	Secondary	87	85, 70	15.05
3	Tertiary	74	85, 85	15.86
4	Primary & secondary	91, 91	70	20.23
5	Primary & tertiary	95, 80	85	31.39
6	Secondary & tertiary	89, 74	85	15.37
7	Primary, secondary & tertiary	91, 91, 76	-	18.53

The optimal combination (Experiment 1) showing that a high reimbursement rate of 95% for primary hospitals, coupled with a descending rate for higher-level hospitals, successfully directs over 36% of inpatients to primary care—more than double the baseline rate of 16.39%.

Dynamics of the Medical Insurance Fund

The redistribution of patients had a direct impact on the medical insurance fund. Scenarios that increased the patient share at primary and secondary hospitals, which have lower average costs, helped alleviate the growth pressure on the fund. Conversely, policies that increased the patient share at tertiary hospitals accelerated fund expenditure. The optimal combination (95%, 85%, 70%) not only improved patient flow but also contributed to a more sustainable financial trajectory for the insurance fund.

Discussion

The agent-based simulation model developed in this study effectively captured the dynamics of patient flow and medical insurance fund expenditure across hospital levels under varying reimbursement rates, aligning with findings from prior research in China indicating that reimbursement rates significantly influence medical demand behavior.⁵¹ The most significant finding is that an optimal configuration exists (95% for primary, 85% for secondary, and 70% for tertiary hospitals), which successfully balances the dual objectives of guiding patients to primary care and ensuring the sustainability of the medical insurance fund. Consistent with the literature, increasing the reimbursement rates generally increased patients volume and visits at the respective hospital level.⁵² Although chronic disease patients often prefer high-level hospitals, cost remains a factor in their decision-making process.^53,54 The implementation of differentiated reimbursement rate demonstrates a positive impact, with an appropriate differential proving more effective. To harness the benefits of medical insurance adjustments, a differentiated reimbursement policy based on disease type, offering higher rates for chronic diseases treated in primary hospitals, could effectively alleviate patient’ financial burdens.

Our simulation results confirm that differential reimbursement rates significantly impact medical insurance fund flows. A critical observation is that when rates were adjusted for multiple hospital levels simultaneously, trends in average costs and patient out-of-pocket expenses were not always aligned. This implies that while a hierarchical system might meet access needs, it does not automatically guarantee lower overall medical costs. Although increasing the proportion of patients visiting primary and secondary hospitals can reduce individual out-of-pocket expenses, it may simultaneously strain the medical insurance fund if not carefully calibrated. Therefore, the effect of any single medical insurance policy is limited in motivating fully rational patient care-seeking.⁵⁵ Achieving hierarchical diagnosis and treatment requires synergistic efforts integrating medical insurance, medical service delivery, and pharmaceutical policies.

Currently, medical insurance reimbursement rates in Beijing are relatively high, shifting focus toward addressing systemic weaknesses, such as the quality of primary care and medication availability. When designing a hierarchical system, it is crucial to consider the interests and establish coordination mechanisms across insurance, care delivery, and pharmaceuticals to genuinely guide patients toward primary healthcare. Indeed, insurance policy adjustments represent just 1 lever for promoting hierarchical diagnosis and treatment.

Improving the service capacity of primary hospitals is fundamental to change residents’ medical care choices. Previous studies consistently identify hospital quality as a top selection criteria for patients,^56
-58 which explains the default preference for high-level hospitals. As economies develop and incomes rise, health consciousness increases, leading individuals to seek better medical technology, often perceived as concentrated in higher-level facilities. From an institutional perspective, hospitals at different tiers have distinct functions and resource endowments; high-level hospitals possess more abundant resources and advanced technologies, better meeting the complex needs of chronic diseases patients.

Several studies from China suggest that implementing family doctor contract services and establishing medical alliance can improve patient perceptions of primary care quality and help address quality deficits.^59
-62 Therefore, for primary hospitals to enhance their service capacity, they must actively leverage favorable policies like family doctor contracts and medical alliances.⁴⁵ Providing technical and expert support from higher-level hospitals to primary hospitals within the same region can raise standards. Simultaneously, the participation of higher-level hospital specialists in family doctor contract service can improve service quality and build resident trust. This improves the patient’s access environment, encouraging voluntary choice of primary hospitals and fostering a more rational care-seeking pattern. In this regard, China’s National Health Commission proposed supportive mechanisms for primary hospitals in 2023, such as mobile medical services and specialist secondments.⁶³ Future efforts could effectively enhance primary care quality by deploying medical resources through mobile services and specialist stations.

Strengths and Limitations

The simulation approach used in this study bridges, to a certain extent, the gap between traditional qualitative and quantitative research methods. Our model accounts for the heterogeneity and complexity of patients’ healthcare choice behavior, enabling better simulation and prediction of the impacts of medical insurance policy adjustments on both patient choices and fund expenditures.

Several limitations of this study should be considered. First, the model was designed and calibrated specifically for hypertensive inpatients in Beijing; thus, the exact optimal reimbursement rates may not be directly transferable to other regions with different economic levels, medical resource distributions, or patient preferences. Second, the model considered only inpatient behavior, excluding the potentially influential role of outpatient visits and referrals. Third, we focused on a single chronic disease; the steering effect of reimbursement rates might differ for acute conditions or diseases with different clinical pathways. Fourth, given the agent-based simulation methodology employed—distinct from empirical studies relying on random sampling—no traditional a priori sample size estimation or power analysis was conducted.

Despite these limitations, the ABMS framework itself is highly adaptable. The model can be recalibrated and extended to other contexts by incorporating new regional data, adjusting core parameters (eg, distance tolerance, cost structures), and introducing different disease modules. Future research could apply this framework to other chronic diseases or integrate outpatient behavior to provide a more comprehensive tool for health policy design.

Conclusions

Leveraging real-world data, this research establishes a simulation framework to demonstrate that the systematic, combinatorial adjustment of reimbursement rates across hospital tiers is a potent mechanism for optimizing healthcare delivery. The findings confirm that such coordinated policy levers can simultaneously redirect patient flow toward primary care and ensure the financial sustainability of the medical insurance fund. The identified optimal rates offer concrete guidance for chronic disease payment reform, while the adaptable modeling framework provides a generalizable tool for forecasting policy impacts in diverse contexts. Prospective model enhancements, including the integration of parameters for health literacy and drug policies, promise to further increase its analytical robustness and utility for evidence-based health policy design.

Supplemental Material

sj-pdf-1-inq-10.1177_00469580261422679 – Supplemental material for Channeling Patient Flow in China: An Agent-Based Model of Insurance Reimbursement and Hierarchical Care

Supplemental material, sj-pdf-1-inq-10.1177_00469580261422679 for Channeling Patient Flow in China: An Agent-Based Model of Insurance Reimbursement and Hierarchical Care by Ning Zhao, Jia Yang, Mei Gu, Jin Li, Danhui Li, Wensheng Ju and Moning Guo in INQUIRY: The Journal of Health Care Organization, Provision, and Financing

Footnotes

Acknowledgements

We thank all the participants, experts and researchers who participated in this study. We thank Home for Researchers editorial team () for language editing service.

List of Abbreviation

AMBS: Agent-based modeling and simulation

ORCID iD

Ning Zhao

Ethical Considerations

The studies involving human participants are reviewed and approved by the Ethics Committee of Capital Medical University (NO.Z2020SY117).

Consent to Participate

Verbal informed consent was obtained from all participants. All methods were performed in accordance with the relevant guidelines and regulations.

Author Contributions

NZ and JY contributed to conception and design of the study, and contributed to the manuscript revision. WJ and MnG collected data and organized the database. NZ wrote the first draft of the manuscript. MG, DL and JL performed the statistical analysis. All authors contributed to reading the manuscript and approved the submitted version.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Natural Science Foundation of China (Grant Nos. 71603175 and 72274129).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The original contributions presented in the study are included in the article material, the data for this study are part of the overall project and so are not publicly available. Access to the datasets of this study can be directed to the corresponding authors.

Supplemental Material

Supplemental material for this article is available online.

References

Central People’s Government of the People’s Republic of China, Guidance on Promoting the Construction of Hierarchical Diagnosis and Treatment. 2015. Accessed June 7, 2023. http://www.gov.cn/zhengce/content/2015-09/11/content_10158.htm

Central People’s Government of the People’s Republic of China, The Fourteenth Five-year Plan for National Economic and Social Development of the People’s Republic of China. 2021. Accessed June 7, 2023. http://www.gov.cn/zhengce/content/2021-09/29/content_5639967.htm

Wang

Shan

Downstream referral decisions and payments mechanism in a hierarchical healthcare system. Syst Eng Theory Pract. 2019;39(08):2126-2137.

Zhang

Zhou

, et al. Analysis of related policy of differential payment of health insurance in China. Med Soc. 2017;30(02):1-4. doi:10.13723/j.yxysh.2017.02.001

Central People’s Government of the People’s Republic of China, Guiding Opinions on Further Deepening the Reform of the Basic Medical Insurance Payment Method. 2017. Accessed June 7, 2023. http://www.gov.cn/zhengce/content/2017-06/28/content_5206315.htm

Central People’s Government of the People’s Republic of China, Opinions on Deepening the Reform of Medical Insurance System. 2020. Accessed June 7, 2023. http://www.gov.cn/zhengce/2020-03/05/content_5487407.htm

Arrow

KJ.

Uncertainty and the welfare economics of medical care. 1963. Bull World Health Organ. 2004;82(2):141-149.

Grossman

On the concept of health capital and the demand for health. J Polit Econ. 1972;80(2):223-255.

Andersen

RM.

National health surveys and the behavioral model of health services use. Med Care. 2008;46(7):647-653. doi:10.1097/MLR.0b013e31817a835d

10.

Chen

Song

Zhou

Wang

Chen

Analysis of influencing factors of economic burden and medical service utilization of diabetic patients in China. PLoS One. 2020; 15(10):e0239844. doi:10.1371/journal.pone.0239844

11.

Chen

Liu

Study on the equity of medical services utilization for elderly enrolled in different basic social medical insurance systems in an underdeveloped city of Southwest China. Int J Equity Health. 2018;17(1):54. doi:10.1186/s12939-018-0765-5

12.

Meng

Fang

Liu

Yuan

Consolidating the social health insurance schemes in China: towards an equitable and efficient health system. Lancet. 2015;386(10002):1484-1492. doi:10.1016/S0140-6736(15)00342-6

13.

Lasser

Himmelstein

Woolhandler

Access to care, health status, and health disparities in the United States and Canada: results of a cross-national population-based survey. Am J Public Health. 2006;96(7):1300-1307. doi:10.2105/AJPH.2004.059402

14.

Yang

Study on healthcare seeking behavior of patients in different levels of hospital in Beijing under the promotion of hierarchical medical system. Chin Hosp. 2018;22(03):1-4. doi:10.19660/j.issn.1671-0592.2018.03.01

15.

Sudano

Baker

DW.

Intermittent lack of health insurance coverage and use of preventive services. Am J Public Health. 2003;93(1):130-137. doi:10.2105/ajph.93.1.130

16.

O’Connor

. The impact of insurance coverage on consumer utilization of health services. Int J Bank Mark. 2015;33(3):276-297.

17.

Manning

Newhouse

Duan

Keeler

Leibowitz

Marquis

MS.

Health insurance and the demand for medical care: evidence from a randomized experiment. Am Econ Rev. 1987;77(3):251-277.

18.

Cheng

Chiang

TL.

The effect of universal health insurance on health care utilization in Taiwan. Results from a natural experiment. JAMA. 1997;278(2):89-93. doi:10.1001/jama.278.2.89

19.

Cheng

, et al. Agent-based modeling and simulation research into Residents Healthcare Choice. Syst Eng. 2009;27(09):96-101.

20.

Chen

The effects of the basic medical insurance for urban workers on the patients’ hospital choice behavior in Shenyang: a study based on multi-agent modelling simulation. Northeastern University, 2012.

21.

Yang

Comparison of the influence of differentiated medical insurance reimbursement on the choice of medical treatment for hospitalized patients with different chronic diseases in Beijing. Chin Hosp. 2023;27(01):11-14. doi:10.19660/j.issn.1671-0592.2023.01.03

22.

Forrest

CB.

Primary care in the United States: primary care gatekeeping and referrals: effective filter or failed experiment?

BMJ. 2003;326(7391):692-695. doi:10.1136/bmj.326.7391.692

23.

Chevreul

Durand-Zaleski

Bahrami

Hernández-Quevedo

Mladovsky

France: health system review. Health Syst Transit. 2010;12(6):1-291, xxi.

24.

Rutter

Savona

Glonti

, et al. The need for a complex systems model of evidence for public health. Lancet. 2017; 390(10112):2602-2604. doi:10.1016/S0140-6736(17)31267-9

25.

Tuite

Gallant

Randell

Bourgeois

Greer

AL.

Stochastic agent-based modeling of tuberculosis in Canadian Indigenous communities. BMC Public Health. 2017;17(1):73. doi:10.1186/s12889-016-3996-7

26.

DePasse

Smith

Raviotta

, et al. Does choice of Influenza vaccine type change disease burden and cost-effectiveness in the United States? An agent-based modeling study. Am J Epidemiol. 2017;185(9):822-831. doi:10.1093/aje/kww229

27.

Wares

Lawson

Shemin

D’Agata

. Evaluating infection prevention strategies in Out-Patient dialysis units using agent-based modeling. PLoS One. 2016;11(5):e0153820. doi:10.1371/journal.pone.0153820

28.

Giabbanelli

Crutzen

Using agent-based models to develop public policy about food behaviours: future directions and recommendations. Comput Math Methods Med. 2017;2017:1-12. doi:10.1155/2017/5742629

29.

Hennessy

Ornstein

Economos

, et al. Designing an agent-based model for childhood obesity interventions: a case study of ChildObesity180. Prev Chronic Dis. 2016;13:E04. doi:10.5888/pcd13.150414

30.

Gorman

Mezic

Gruenewald

PJ.

Agent-based modeling of drinking behavior: a preliminary model and potential applications to theory and practice. Am J Public Health. 2006;96(11):2055-2060. doi:10.2105/AJPH.2005.063289

31.

Yang

Diez Roux

Auchincloss

Rodriguez

Brown

DG.

A spatial agent-based model for the simulation of adults’ daily walking within a city. Am J Prev Med. 2011; 40(3):353-361. doi:10.1016/j.amepre.2010.11.017

32.

Barbrook-Johnson

Badham

Gilbert

Uses of Agent-based modeling for Health Communication: the TELL ME Case Study. Health Commun. 2017;32(8):939-944. doi:10.1080/10410236.2016.1196414

33.

Guclu

Kumar

Galloway

, et al. An agent-based model for addressing the impact of a disaster on access to Primary Care Services. Disaster Med Public Health Prep. 2016;10(3):386-393. doi:10.1017/dmp.2016.44

34.

Ernecoff

Keane

Albert

SM.

Health behavior change in advance care planning: an agent-based model. BMC Public Health. 2016;16:193. doi:10.1186/s12889-016-2872-9

35.

Megiddo

Colson

Chisholm

Dua

Nandi

Laxminarayan

Health and economic benefits of public financing of epilepsy treatment in India: an agent-based simulation model. Epilepsia. 2016;57(3):464-474. doi:10.1111/epi.13294

36.

Alibrahim

An agent-based simulation model of patient choice of health care providers in accountable care organizations. Health Care Manag Sci. 2018;21(1):131-143. doi:10.1007/s10729-016-9383-1

37.

Einzinger

Popper

Breitenecker

, et al. The GAP-DRG Model: simulation of outpatient care for comparison of different reimbursement schemes. Proc. 2013 Winter Simul. Conf. Simul. Mak. Decis. A complex world. Piscataway, IEEE Press; 2013:pp. 2299-2308.

38.

Liu

An agent-based simulation model to study accountable care organizations. Health Care Manag Sci. 2016;19(1):89-101. doi:10.1007/s10729-014-9279-x

39.

Tan

Chen

Zhang

, et al. Effects of the basic medical insurance policy for residents on hypertensive inpatients’ medical services seeking behaviors and medical insurance fund flow in Beijing based on multi-agent modeling simulation. Chin Heal Econ. 2019;38(02):17-20.

40.

Husereau

Drummond

Augustovski

Consolidated Health Economic Evaluation Reporting Standards (CHEERS 2022) explanation and elaboration: a report of the ISPOR CHEERS II Good Practices Task Force. Value Health. 2022; 25:1060. doi:10.1016/j.jval.2022.03.002

41.

Bonabeau

Agent-based modeling: methods and techniques for simulating human systems. Proc Natl Acad Sci USA. 2002;99(l_3):7280-7287. doi:10.1073/pnas.082080899

42.

Marshall

Burgos-Liz

IJzerman

, et al. Applying dynamic simulation modeling methods in health care delivery research—the SIMULATE checklist: report of the ISPOR simulation modeling emerging good practices task force. Value Health. 2015;18(1):5-16. doi:10.1016/j.jval.2014.12.001

43.

Zhang

Giabbanelli

Arah

Zimmerman

FJ.

Impact of different policies on unhealthy dietary behaviors in an urban adult population: an agent-based simulation model. Am J Public Health. 2014;104(7):1217-1222. doi:10.2105/AJPH.2014.301934

44.

Wang

Chen

, et al. 2019 Chinese Hypertension League guidelines on home blood pressure monitoring. J Clin Hypertens. 2020;22(3):378-383. doi:10.1111/jch.13779

45.

Krumholz

Yip

, et al. Quality of primary health care in China: challenges and recommendations. Lancet. 2020; 395(10239):1802-1812. doi:10.1016/S0140-6736(20)30122-7

46.

Liu

Berman

Yip

, et al. Health care in China: the role of non-government providers. Health Policy. 2006;77(2):212-220. doi:10.1016/j.healthpol.2005.07.002

47.

Wei

. Influence of physical attractiveness and gender on patient preferences in digital doctor consultations: experimental study. J Med Internet Res. 2024;26:e46551. doi:10.2196/46551

48.

Andersen

RM.

Revisiting the behavioral model and access to medical care: does it matter?

J Health Soc Behav. 1995;36(1):1-10.

49.

Qian

Lucas

Chen

Zhang

Determinants of the use of different types of health care provider in urban China: a tracer illness study of URTI. Health Policy. 2010;98(2-3):227-235. doi:10.1016/j.healthpol.2010.06.014

50.

Ryan

Farrar

Using conjoint analysis to elicit preferences for health care. BMJ. 2000;320(7248):1530-1533. doi:10.1136/bmj.320.7248.1530

51.

Whether health insurance policy can promote the realization of the hierarchical medical system: an empirical analysis based on medical demand behavior. Comp Econ Soc Syst. 2018;(01):28-35.

52.

Gan

You

Reimbursement gap, patient behavior and medical expenses——A three-stage dynamic game analyses. Syst Eng Theory Pract. 2014;34(11):2974-2983. doi:CNKI:SUN:XTLL.0.2014-11-026

53.

Peng

Xiong

, et al. Study on medical service utilization preference of urban elderly chronic patients based on discrete choice experiment. Chinese Journal of Health Policy. 2019;12(09):54-59. doi:CNKI:SUN:ZGWZ.0.2019-09-011

54.

Qian

Pong

Yin

Nagarajan

Meng

Determinants of health care demand in poor, rural China: the case of Gansu Province. Health Policy Plan. 2009;24(5):324-334. doi:10.1093/heapol/czp016

55.

Gao

Wang

Does differentiated reimbursement rate help to promote hierarchical medical services: evidence from inpatient claim data. Insur Stud. 2018;(07):89-103. doi:10.13497/j.cnki.is.2018.07.008

56.

Varkevisser

van der Geest

Schut

FT.

Do patients choose hospitals with high quality ratings? Empirical evidence from the market for angioplasty in the Netherlands. J Health Econ. 2012;31(2):371-378. doi:10.1016/j.jhealeco.2012.02.001

57.

Beckert

Christensen

Collyer

Choice of NHS-funded hospital services in England. Econ J. 2012;122(560):400-417. doi:10.1111/j.1468-0297.2012.02496.x

58.

McConnell

Lindrooth

Wholey

Maddox

Bloom

Modern Management Practices and hospital admissions. Health Econ. 2016;25(4):470-485. doi:10.1002/hec.3171

59.

Feng

Cheng

Luo

Xiao

Zhang

Effect of family doctor contract services on patient perceived quality of primary care in southern China. BMC Fam Pract. 2020;21(1):218. doi:10.1186/s12875-020-01287-7

60.

Zhong

Mei

Liang

Kuang

Effect of family practice contract services on the quality of primary care in Guangzhou, China: a cross-sectional study using PCAT-AE. BMJ Open. 2018;8(11):e021317. doi:10.1136/bmjopen-2017-021317

61.

Dai

Exploration and practice of primary health care during the construction of county-level Medical Alliance. Chin Gen Pr. 2021;24(01):23-29.

62.

Cai

Wen

Tang

, et al. Exploration and consideration of the Medical Alliance modes. Iran J Public Health. 2018; 47(8):1160-1165.

63.

China’s National Health Commission. Guidance on the work of mobile medical service and stationed services in the county. 2023. Accessed June 7, 2023. http://www.nhc.gov.cn/jws/s7874/202303/bba864e9c6874c36a654a443816ab701.shtml

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB