Introduction

Expensive and difficult access to health care has been a common public health concern in most countries.1–3 The reform of China’s health care system also faces significant challenges in reducing individuals’ financial access to health care, improving their access to health services, and promoting their overall health status.4,5 The Chinese government, since 2009, launched a new round of health reform to alleviate the problem of expensive and difficult access to health care and to improve individuals’ healthcare experience.6,7 The reform’s main initiative is to increase the supply of healthcare resources and establish a basic universal healthcare system to improve the accessibility to health services, thereby improving people’s overall health level. Presently, the Chinese government has established a basic social medical insurance system covering the entire population. Data from China National Health Insurance Administration (CNHIA) in 2023 showed that the number of people covered by China’s social medical insurances has reached 1.36 billion, with the participation rate remaining stableat over 95% for four consecutive years.8 Numerous studies have shown that China’s health reform has significantly increased health service usage and improved the population’s overall health.9–11

Nonetheless, despite significant improvements in the overall health of Chinese residents, the new health reform has also created new issues in the health services market, one of the main challenges is the siphon effect from the overconcentration of healthcare resources in general hospitals.12,13 Siphon effect in health service market is a general term used to denote the phenomenon where larger hospitals, by leveraging superior resources and job opportunities, attract higher-level medical teams, subsequently drawing talent that could have worked at the grassroots level and patients who might have sought treatment at primary healthcare institutions.16,17

Contrary to integrated healthcare systems based on primary care (such as UK), China has not set up a comprehensive ‘gatekeeper’ systems and patients seem inclined toward larger and more reputable hospitals.2,17–19 As a result, patients previously treated in smaller facilities are easily attracted once they seek treatment at larger medical institutions.20,21 In China, the hospital classification system comprises three tiers: primary, secondary, and tertiary hospitals (secondary and tertiary hospitals are commonly categorized as general hospitals). Primary hospitals serve as local medical hubs, focusing on basic medical care including diagnosis, treatment for common ailments, and routine medical procedures. Secondary hospitals, located at the city/county level, feature specialized departments and advanced technologies, which mainly provide a higher standard of medical services. Tertiary hospitals, as regional or national medical centres, feature the most advanced medical equipment and technologies. These institutions provide comprehensive and high-quality medical services. The main aim of classification regarding hospitals is to streamline patients’ access to appropriate medical care according to the severity of their conditions and treatment needs, and thus meeting their specific healthcare requirements.21–23

However, as individual’s income and the benefits level of medical insurances continue to rise, the demand for high-quality health services increased drastically among Chinese residents. Consequently, the siphon effect is increasingly becoming serious in high-grade hospitals. Individuals tend to choose general and special hospitals for medical treatment, and primary hospitals failed to play the role of ‘gatekeeper’. The siphon effect has broken the division of labour among hospitals at different levels, aggravated the tension in quality medical resources, and affected the efficiency of the healthcare service system. In addition, the price and costs of tertiary hospitals are higher than that of primary hospitals, and medical sources are limited, which will further aggravate the problem of ‘difficult and expensive access to health care’.16,17,25,26 Furthermore, studies have shown that long waiting time and overworked doctors have become the norm in general hospitals, with over 90% of doctors working overtime.27,28

Regarding the determinants of siphon effect in health service market, scholars have argued that the lack of an effective hierarchical medical system (HMS) contributed to the siphon effect in general hospitals at the macro level.19,29,30 Although the differential reimbursement ratios at different hospital levels have a particular effect in guiding residents to seek medical treatment in primary hospitals, the current disparity in the reimbursement rate of health insurance is insufficient to guide residents to seek medical treatment rationally.31 On the micro level, increasing household income, lacking of trust in primary care, increasing individual health awareness, disease severity, and higher education attainment were the key factors contributing to people’s preference for general hospitals.32,33 Shen & Zhang (2016) found that the inadequate capacity and equipment of medical services were the main reason for the low willingness of residents to seek treatment in primary hospitals.30 Meanwhile, the accessibility of healthcare facilities can also influence patients’ choices of medical institutions, the distance and the time taken to the hospital were the primary consideration factors. Previous studies showed that long waiting time for patients to access healthcare would affect their decision to medical institutions, and some studies even suggested that sensitivity for waiting time is higher than quality.28,34

Some scholars also explored the impact of medical insurance on residents’ healthcare choices, and found that medical insurance not only directly affected the accessibility to healthcare, but also significantly influenced their healthcare decisions.35–38 For instance, in the early years of the New Rural Cooperative Medical System (NRCMS), the insured residents preferred township health centres because of higher health insurance subsidies.39–41 A wealth of studies assessed the utilization of medical services among residents covered by NRCMS in rural China, and found that NRCMS could significantly increase the utilization of inpatient services.42 Hadley et al (1997) compared the effects of different health insurance systems on the choice of medical institutions for young women with breast cancer in the United States and found that female enrolled in health maintenance organizations were more likely to bypass the nearest cancer hospital.43 Jiang and Zheng (2022) verified that the proportion of residents visiting primary care facilities decreased by 1.5% after the NRCMS was integrated into the Basic Medical Insurance System for Urban and Rural Residents (BMISURR).44

Presently, the Chinese government has established two basic social health insurance systems covering all residents: the Basic Medical Insurance System for Urban Employees (BMISUE) and BMISURR (See Figure 1). However, these two programs are significantly different in financing and benefits levels. Specifically, BMISUE is financed mainly by individuals and enterprises, but BMISURR is mainly funded by government subsidies and individuals. Overall, the benefits level of BMISUE are higher than BMISURR when residents use same medical services, the data from China’s National Health Insurance Bureau showed that the average reimbursement rate for hospitalization in BMISUE is about 80%, which is higher than that of BMISURR (approximately 70%). Nevertheless, whether these differences will cause the differences in medical institution choice between BMISUE and BMISURR need more empirical evidence.

Figure 1 The framework of China’s social medical insurance systems.

The siphon effect in healthcare market is notably pronounced in many countries, and undeveloped ‘gatekeeper’ systems exacerbate disparities in accessing healthcare services and resources. How to measure this effect within healthcare markets presents a substantial challenge. Currently, to our knowledge, there is nearly no literature exploring how social medical insurances influence siphon effect of general hospitals. Therefore, bridging this research gap would provide crucial insights to Chinese healthcare systems reform. The marginal contributions of this study are as follows: first, by utilizing representative regional survey data (2021 Health Life Satisfaction Survey of Yangtze River Delta), we introduced a new metric denoted the “distance-healthcare choice” variable. This metric comprehensively quantified the prevalent siphon effect within healthcare service market. Second, this study further explored the influence of different medical insurances on the siphon effect of general medical institutions by using the maximum likelihood method (logit model), which had substantial implications for Chinese healthcare system reforms.

Materials and Methods Data Source

The data used in this study were from the “Health Life Satisfaction Survey of Yangtze River Delta (HLSSYRD)”, implemented by Shanghai Jiao Tong University in 2021. The survey used all urban residents in the Yangtze River Delta in China (YRDC) as the basic sampling frame and used a stratified, multistage, and population size-proportional method for sampling. The survey comprised 10 sections, including basic demographic information, medical insurance, choices of medical institutions, health environment, and health society of residents in the YRDC.

A total of 19,870 questionnaires were distributed in HLSSYRD, and 18,031 people participated in this survey. The response rate was 90.74%. According to research needs of this study, we cleaned the data as follows (see Figure 2): First, the main aim of this study was to analyze the effect of social medical insurance programs on the siphon effect in the health service market, so we only retained these samples covered by BMISUE and BMIISURR. Second, we removed duplicate enrolment samples to get a sample size of 14,591. Finally, after removing those samples with responses such as “I don’t know”, “I refuse to answer” and those with missing values of key variables (such as income and the choices of medical institutions), the final sample used for this study was 14,425.

Figure 2 The flowchart of sample selection.

Variable Measurement Dependent Variable

Based on existing research and questionnaire design, this study used the first choice of medical institutions to measure residents’ willingness to choose medical institution when seeking for medical treatment.45,46 Specifically, the question “What type of medical institutions is your first choice for daily consultations?” was utilized to determine the respondents’ willingness to seek for medical treatment. Respondents were asked to choose from four options: tertiary hospitals, secondary hospitals, community health centres/township health centres and private clinics. A value of 1 was assigned when the respondent selected tertiary and secondary hospitals (general hospitals) and a value of 0 was assigned when the respondent selected primary hospitals.

We used the following two questions to identify the siphon effect in the health service market, namely, “What type of medical institutions is your first choice for daily consultations (tertiary hospitals, secondary hospitals, community health centres/township health centres and private clinics)?” and “What’s the nearest medical institutions in your residence (tertiary hospitals, secondary hospitals, community health centres/township health centres and private clinics)?”. Specifically, if the nearest medical institution to a respondent is a primary institution, but their first choice of medical institution is a general hospital (tertiary hospitals and secondary hospitals), we suggest that siphon effect of preferring general hospitals may exist. Therefore, we construct a “distance-first visit choice” variable, assigning a value of 1 when the respondent chose a general hospital when the nearest medical institution is a primary hospital. Otherwise, we assigned a value of 0. Table 1 presented the values of new synthetic variables.

Table 1 The Values of Distance-First Visit Choice Variable

Independent Variables

The core explanatory variable of this study was the types of social medical insurance. The social medical insurance in China includes BMISUE and BMISURR. Owing to China’s universal medical insurance reform, only 0.5% samples were not covered by medical insurance, so we dropped these samples. We generated a “social medical insurance” dichotomous variable and assigned a value of 1 when a respondent was covered by BMISUE. Otherwise, we gave a value of 0 when people insured by BMISURR.

Control Variables

In order to prevent estimation bias in the model due to missing variables, we applied the Anderson Health Model,47 which has been widely used to analyse the factors influencing health service utilization behaviour, to select control variables. Three types of control variables were selected in this study: predisposing characteristics, enabling factors, and need factors. Specifically, predisposing characteristics included age, gender, education, occupation and marital status; enabling factors included household income; needs factors included chronic diseases and self-rated health. The definition and descriptive results of variables are shown in Table 2.

Table 2 Definition and Descriptive Results of Variables

Analytical Plan

Stata 14.0 was used to analyze the relationship between social medical insurance and choices of medical institution among urban residents in YRDC. Firstly, we used descriptive statistics to describe the overall sociodemographic characteristics of the samples, and Chi-square test was used to investigate the differences between choice of medical institution and social medical insurances. Finally, we conducted logistics regression model to explore the association between social medical insurance and choices of medical institution, and propensity score matching (PSM) method was used to examine the robustness of the regression results.

Results Descriptive Results

Table 2 presented the status of choice of medical institution, participation of social medical insurance and control variables of the study samples. More than 66.97% of total residents prefer to chosen general hospital when seeking for medical treatment. Furthermore, nearly 45% of people chosen general hospitals although they were close to primary hospitals. In terms of medical insurance participation, BMISUE had the highest participation rate of 73.78%, and 26.22% of people insured by BMISURR.

In terms of control variables, male and female accounted for 50.37% and 49.63% of the total sample, respectively. More than 83% of total adults were married with a spouse, and only 16.5% of people remained unmarried. The education attainment of all residents was extremely optimistic, more than 57% of all urban residents had attended college and above, and followed by high school and vocational school, accounting for 32.95%, the lowest was middle school and below, with only 9.98% of total samples.

From socio-economic indicators, the ratio of people having a job was 66.50%. More than 55% of residents’ monthly household income ranging from 10,001 to 20,000 CNY, the ratio of monthly household income ranging from 20,001 to 40,000 CNY was 22.42%, and followed by 0–10,000 CNY. Only 2.96% of total samples’ monthly household income was 40,001 CNY and above. In terms of health status, the ratio of people with chronic disease was 24.34%, and more than 75% of total samples did not suffer from chronic disease. The overall health status was optimistic, over 74% of all residents reported good health level, and only 25.78% of residents had poor health conditions.

Table 3 showed the Chi-squared test between social medical insurances and medical institution choice. There were significant differences in the choice of medical institution and siphon effect between BMISUE and BMISURR (P < 0.001).

Table 3 The Results of the Chi-Square Test

Regression Results

We adopted the maximum likelihood method to evaluate the association between social medical insurance and individual’s choice of medical institutions. Column (1) in Table 4 showed that compared to residents insured by BMISURR, those covered by BMISUE were more likely to choose general hospital when seeking medical treatment (OR = 5.377, 95% CI:4.887, 5.915). Column (2) in Table 4 reported the regression results for the impact of social medical insurances on the siphon effect in medical service market. The results showed that compared to people insured by BMISURR, those residents covered by BMISUE were more likely to choose general hospitals despite the fact that they were close to primary hospitals (OR = 3.240, 95% CI:2.945, 3.565). In other words, participating in BMISUE was more likely to cause siphon effect in medical service market.

Table 4 The Regression Results for the Association Between Health Insurance and Choices of Medical Institution

Among controlled variables, increasing age (OR = 0.881, 95% CI: 0.778, 0.999; OR = 0.744, 95% CI: 0.649, 0.865; OR = 0.656, 95% CI: 0.538, 0.799) was negatively associated with the siphon effect of general hospitals. Conversely, higher education level was positively related to the siphon effect of general hospitals (OR = 1.336, 95% CI: 1.132, 1.578; OR = 1.454, 95% CI: 1.213, 1.743). In terms of socio-economic conditions, higher monthly household income was positively associated with choosing general hospitals (OR = 2.625, 95% CI: 2.363, 2.916; OR = 1.440, 95% CI: 1.271, 1.632). Conversely, having a job was negatively related to choosing general hospitals relative to those without a job (OR = 0.679 CI: 0.606, 0.763). From health status, chronic disease would exacerbate siphon effect (OR = 1.370, 95% CI: 1.219, 1.540), people with chronic disease were more likely to choose general hospitals despite they were closed to primary hospitals. Conversely, self-rated health was negatively associated with siphon effect, the better the health level of a resident, the lower the likelihood of siphon effect.

Robust Test

In order to reduce the potential selective bias caused by social medical insurance participation behavior and examine the robustness of basic regression results, the PSM method was used to analyzed the relationship between social medical insurance and choices of medical institutions. We used the K-nearest neighbor matching (one to four), Caliper matching and Kernel matching to calculate the average treatment effect (ATT) of BMISUE and BMISURR on medical institution choice. Column (1) in Table 5 showed that the ATT coefficients for K-nearest neighbor matching, Caliper matching, and Kernel matching were 0.329, 0.332, and 0.335, respectively (P < 0.001), which indicated that people insured by BMISUE were more likely to choose general hospitals when seeking medical treatment. Column (2) in Table 4 showed that the ATT coefficients for K-nearest neighbor matching, Caliper matching, and Kernel matching were 0.247, 0.243, and 0.244, respectively (P < 0.001), implying that people insured by BMISUE were more likely to cause siphon effect relative to those covered by BMISURR.

Table 5 PSM Results for the Association Between Social Medical Insurance and the Choice of Medical Institutions

Heterogeneity Analysis

Table 6 delineated the relationship between social medical insurance programs and siphon effect of general hospitals, segmented by age and monthly household income. Column (1)-(2) showed that BMISUE was positively associated with siphon effect among different age groups, and this kind of positive association among individuals aged 15–59 years was higher than older people. Column (3)-(5) showed that BMISUE was positively related to siphon effect among different income groups, and the siphon effect increased gradually as income increased.

Table 6 The Regression Results of Heterogeneity Analysis

Discussion

Primary care has been the first line of defence against chronic diseases in developed countries,48 and primary medical institutions are the best choice for chronic disease management.49 However, most high-quality medical resources in China are concentrated in general hospitals, and patients lack trust in primary healthcare institutions. Currently, the Chinese government has not established a comprehensive HMS, so patients can easily give up primary care institutions and choose general hospitals for medical treatment. This situation will lead to a severe underutilization of primary health services and increase the pressure on general hospitals, which poses a severe threat to the efficiency and effectiveness of China’s healthcare system. Although several studies have analysed the association between medical insurances and choice of medical institutions,35–38,50,51 to the best of our knowledge, this study represented the first attempt to measure the siphon effect of medical service market and further explored the association between medical insurance programs and siphon effect of general hospitals.

By using data from 2021 HLSSYD, this study found that residents covered by BMISUE were more likely to choose a general hospital when seeking medical treatment than those insured by BMISURR. Further analysis found that BMISUE may lead to the siphon effect of general hospitals, residents covered by BMISUE were more likely to choose general hospitals despite being close to primary hospitals compared to those insured by BMISURR.

Social medical insurance can guide patients’ access to care by increasing their financial accessibility and changing the price they face through differentiated reimbursement rate designs.25,52,53 Our research supported this view. Moreover, BMISUE and BMISURR differ significantly in terms of medical financing and benefit levels, and the medical reimbursement ratio for BMISUE is significantly higher than that for BMISURR. Consequently, BMISUE can provide greater financial accessibility to patients than BMISURR, and participants covered by BMISUE were more likely to choose general medical institutions to seek high-quality health services.

Second, this study suggested that BMISUE may exacerbate the siphon effect of preferring general hospitals. Although people insured by BMISUE were close to primary medical institutions, they were still more likely to choose general medical institutions when they first seek medical treatment. Previous studies suggested that differentiated reimbursement rates of health insurance among different medical institutions can reduce the siphon effect of higher medical institutions and effectively guide price-sensitive patients to primary hospitals.16,19,31 Although the Chinese government has established different reimbursement rates among different levels of medical institutions, the reimbursement rate for BMISUE is significantly higher than that for BMISURR when seeking healthcare service at the same medical institution (see Table 7). This differentiated reimbursement treatment can cause BMISUE participants to prefer higher-rated medical institutions for health services, thus exacerbating the siphon effect of general hospitals.

Table 7 Reimbursement Rates in Different Hospital Levels in YRDC

More importantly, moral hazard is common in the health insurance market. The Chinese government stipulates that, except for urban employees, residents can also participate in BMISUE as flexible employees if they afford all the premiums (8% of the average local social wage). Residents with poor health status are more likely to participate in BMISUE to enjoy higher medical insurance benefits and better medical services. Consequently, this exacerbates the siphon effect of high-rated medical institutions. Feng et al (2018) found that the medical expenditure of flexible workers who chose to join BMISUE was 78% higher than that of urban employees.54 Furthermore, the medical expenditure of those who chose to join BMISUE was approximately 45% higher than that of those who chose to join BMISURR.

Finally, aligning with previous research, household income exhibits positive correlations with the inclination towards general hospitals.25,55 This might stem from income’s multifaceted influence on medical preferences, encompassing socioeconomic status and people with higher incomes tend to have more opportunities to participate in BMISUE.53,55,56 Furthermore, prior studies suggested that the elderly were more likely to choose general hospitals,56 which was contrary to our findings. This might be attributed to poorer health conditions resulting in increased healthcare needs and frequent hospital visits, where general hospitals involve longer waiting time. Additionally, limitations in daily living abilities and financial constraints restrict the elderly’s access to larger medical facilities.56

Conclusion and Limitation

This study has several limitations. First, due to data limitations, this study only verified the impact of social medical insurances on the siphon effect of higher medical institutions in the YRDC. However, considering the significant differences in economic and social development between the eastern, central and western regions of China, further research is needed to determine whether there are regional differences. Moreover, this study only analyzed the effect of medical insurances on urban residents’ first choice of medical treatment, and more evidence is needed to regarding rural residents. Third, this study used cross-sectional data, so we failed to verify the causal relationship between social medical insurance, the first choice of medical treatment and the siphon effect of general hospitals. Finally, this article included three types of variables in the Anderson Health Model, such as predisposition characteristics, enabling resources and needs. However, due to data limitation, this study failed to add all variables related to the Anderson Health Model variables. We will continue to conduct relevant research once data available.

Despite the acknowledged limitations, this study has significant implications for the Chinese government to further improve its health reform system. First, effectively solving the “expensive and difficult access to health care” issue is a prerequisite for achieving the “Health China Strategy”. The government should continue to enhance the health reform system, improve the quality of medical services of primary hospitals and maintain their role as gatekeepers. Furthermore, health reform should also highlight the fairness of the medical security system, and measures should be taken to narrow the gap between BMISUE and BMISURR, strengthen the subsidies for poor residents caused by disease and provide more medical resources transfer payments to the vulnerable groups.

Abbreviations

BMISUE, Basic Medical Insurance System for Urban Employees; BMISURR, Basic Medical Insurance System for Urban and Rural Residents; YRDC, Yangtze River Delta in China; HLSSYRD, Health Life Satisfaction Survey of Yangtze River Delta; HMS, Hierarchical Medical System; NRCMS, New Rural Cooperative Medical System.

Ethics Approval

The data was collected using a semi-structured questionnaire contained participants’ socio-demographics and healthy lifestyle information in Yangtze River Delta. Obtained informed consent from all study participants before the start of the study. The study protocol was approved by the ethics review committee of Shanghai University of Medicine & Health Sciences, China based on the assessment of design and process of the study and was conducted in accordance with the principles of the Declaration of Helsinki. The approval number is 2021-gskyb-02-372424198012222511. Before starting the research study, the respondents were informed about the anonymity of this study with relevant details. The questionnaire-related information is as follows: “In order to ensure the authenticity and confidentiality of the research, this research is anonymous, and the content of the survey is only used for relevant research, and the content of your answers will be kept confidential in accordance with the requirements of the Statistical Law of the People’s Republic of China”.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Funding

This research was supported by the National Social Science Fund of China (22AZD082) and China Scholarship Council program (202306010171).

Disclosure

The authors report no conflicts of interest in this work.

References

1. Kangovi S, Barg FK, Carter T, Long JA, Shannon R, Grande D. Understanding why patients of low socioeconomic status prefer hospitals over ambulatory care. Health Affairs. 2013;32(7):1196–1203. doi:10.1377/hlthaff.2012.0825

2. Faden RR, Chalkidou K, Appleby J, Waters HR, Leider JP. Expensive cancer drugs: a comparison between the United States and the United Kingdom. Milbank Q. 2009;87(4):789–819. doi:10.1111/j.1468-0009.2009.00579.x

3. Yu W, Li M, Ge Y, et al. Transformation of potential medical demand in China: a system dynamics simulation model. J biomed informat. 2015;57:399–414. doi:10.1016/j.jbi.2015.08.015

4. Yip W, Mahal A. The health care systems of China and India: performance and future challenges. Rev Health Aff. 2008;27(4):921–932. doi:10.1377/hlthaff.27.4.921

5. Wagstaff A, Yip W, Lindelow M, Hsiao WC. China’s health system and its reform: a review of recent studies. Rev Health Econ. 2009;18(SUPPL. 2):S7–S23. doi:10.1002/hec.1518

6. Meng Q. Strengthening public health systems in China. Note. Lancet Public Health. 2022;7(12):e987–e988. doi:10.1016/S2468-2667(22)00206-7

7. Yip WCM, Hsiao WC. What drove the cycles of Chinese health system reforms? Health Care Policy in East Asia. 2019;2019:5–23.

8. China National Health Insurance Administration. Statistical report on the development of national healthcare security in 2022; 2023. Available from http://www.nhsa.gov.cn/art/2023/7/10/art_7_10995.html. Accessed December12, 2023.

9. Chen Q, Chu X, Wang S, Zhang B. A triple-difference approach to re-evaluating the impact of China’s new cooperative medical scheme on incidences of chronic diseases among older adults in rural communities. Risk Manag Health Policy. 2020;643–659. doi:10.2147/RMHP.S244021

10. Wang C, Cha Z. Has the reform of public hospitals improved the capacity of medical and health services? Analysis of the promotion effect based on financial input. J Guang Univer Finan Econom. 2022;37(04):100–111.

11. Wang C, Cha Z. Has the comprehensive medical reform pilot alleviated the problems of the accessibility and affordability of health care? Public Finan Res. 2021;(12):79–92. doi:10.19477/j.cnki.11-1077/f.2021.12.005

12. Yip WCM, Hsiao W, Meng Q, Chen W, Sun X. Realignment of incentives for health-care providers in China. Review. Lancet. 2010;375(9720):1120–1130. doi:10.1016/S0140-6736(10)60063-3

13. Hung LM, Shi L, Wang H, Nie X, Meng Q. Chinese primary care providers and motivating factors on performance. Family Practice. 2013;30(5):576–586. doi:10.1093/fampra/cmt026

14. Shen M, Li Y. “Siphon Effect” of tertiary hospitals and the effects of changes in cost-sharing on patient behavior—an empirical study in a large city. J Public Administ. 2021;04:61–84+229–230.

15. Chen Y, Wang L, Cui X, et al. COVID-19 as an opportunity to reveal the impact of large hospital expansion on the healthcare delivery system: evidence from Shanghai, China. Ann Translat Med. 2021;9:16.

16. Zhan J, Fu H. Hospital reputation, spatial distance and patient choice—evidence from hospital discharge data. Chin Economic Quart. 2022;22(01):343+364+344–363. doi:10.13821/j.cnki.ceq.2022.01.17

17. Feng J, Lv S, Wang Z. Medical resource sharing and patients’ choice: evaluation of the medical consortium construction policy in China. J Manag World. 2022;38(10):144–157+173+158. doi:10.19744/j.cnki.11-1235/f.2022.0142

18. Pang R, Li S. Structural imbalance of medical resource allocation and ‘Kanbing Gui’ in China: from the perspective of referral system. Mod Econom Sci. 2022;44(03):97–110.

19. Liu Y, Zhong L, Yuan S, Van de Klundert J. Why patients prefer high-level healthcare facilities: a qualitative study using focus groups in rural and urban China. BMJ Global Health. 2018;3(5):e000854. doi:10.1136/bmjgh-2018-000854

20. Qian D, Lucas H, Chen J, Xu L, Zhang Y. 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

21. Barber SL, Borowitz M, Bekedam H, Ma J. The hospital of the future in China: china’s reform of public hospitals and trends from industrialized countries. Health Policy Plann. 2014;29(3):367–378. doi:10.1093/heapol/czt023

22. Pan J, Qin X, Hsieh C-R. Is the pro-competition policy an effective solution for China’s public hospital reform? Health Econ Policy Law. 2016;11(4):337–357. doi:10.1017/S1744133116000220

23. Zhao X, Xiao J, Chen H, et al. Patient preferences and attitudes towards first choice medical services in Shenzhen, China: a cross-sectional study. BMJ Open. 2022;12(5):e057280. doi:10.1136/bmjopen-2021-057280

24. Shen M, He W, Li L. Incentives to use primary care and their impact on healthcare utilization: evidence using a public health insurance dataset in China. Soc sci med. 2020;255:112981. doi:10.1016/j.socscimed.2020.112981

25. Hu Y, Zhang Z. Skilled doctors in tertiary hospitals are already overworked in China. Lancet Glob Health. 2015;3(12):e737. doi:10.1016/S2214-109X(15)00192-8

26. Sivey P. The effect of waiting time and distance on hospital choice for English cataract patients. Health Econom. 2012;21(4):444–456. doi:10.1002/hec.1720

27. Zhang L. A study on the effective strategies of urban hierarchical medical system in the post-epidemic era. Nanj J Soc Sci. 2020;(04):7–13. doi:10.15937/j.cnki.issn1001-8263.2020.04.002

28. Shen S, Zhang B. Hierarchical medical system, primary care and construction of primary health care institutions. Academ Bimest. 2016;(02):48–57. doi:10.16091/j.cnki.cn32-1308/c.2016.02.008

29. Li X, Gao G, Hu X, Zhang D, Tian J, Deng Q. The influencing factors of residents’medical treatment flow under differentiated reimbursement policy of medical insurance. Chin J Health Policy. 2020;13(12):23–29.

30. Tay A. Assessing competition in hospital care markets: the importance of accounting for quality differentiation. Rand J Econ. 2003;34:786–814. doi:10.2307/1593788

31. Smith H, Currie C, Chaiwuttisak P, Kyprianou A. Patient choice modelling: how do patients choose their hospitals? Health Care Manag Sci. 2018;21:259–268. doi:10.1007/s10729-017-9399-1

32. Gaynor M, Propper C, Seiler S. Free to Choose: Reform and Demand Response in the British National Health Service. Unpublished manuscript, Stanford University; 2011.

33. Wagstaff A, Lindelow M, Jun G, Ling X, Juncheng Q. Extending health insurance to the rural population: an impact evaluation of China’s new cooperative medical scheme. J Health Econ. 2009;28(1):1–19. doi:10.1016/j.jhealeco.2008.10.007

34. Chandra A, Gruber J, McKnight R. Patient cost-sharing and hospitalization offsets in the elderly. Am Econ Rev. 2010;100(1):193–213. doi:10.1257/aer.100.1.193

35. Manning WG, Newhouse JP, Duan N, Keeler EB, Leibowitz A. Health insurance and the demand for medical care: evidence from a randomized experiment. Am Econom Rev. 1987;1987:251–277.

36. Normand C. Free for all: lessons from the RAND health insurance experiment. BMJ. 1994;308(6945):1724–1725. doi:10.1136/bmj.308.6945.1724a

37. Brown PH, Theoharides C. Health‐seeking behavior and hospital choice in China’s new cooperative medical system. Health Economics. 2009;18(S2):S47–S64. doi:10.1002/hec.1508

38. Ren X, Wang Y. The Effect of China’ s new cooperative medical scheme on the farmers’ medical behavior under the background of urbanization. Finan Econom. 2015;03:121–130.

39. Jiang J. NCMS and health care provider choices in Rural China. South Chin J Econom. 2013;(02):56–66. doi:10.19592/j.cnki.scje.2013.02.006

40. Yu B, Meng Q, Collins C, et al. How does the New Cooperative Medical Scheme influence health service utilization? A study in two provinces in rural China. BMC Health Serv Res. 2010;10:1–9. doi:10.1186/1472-6963-10-116

41. Hadley J, Mitchell JM. Effects of HMO Market Penetration on Physicians’ Work Effort and Satisfaction: with the rise in managed care, a majority of physicians report having reservations about the way they now practice medicine. Health Affairs. 1997;16(6):99–111. doi:10.1377/hlthaff.16.6.99

42. Jiang Y, Zheng K. Unified management of the medical insurance for Urban and Rural Residents, and medical treatment behavior and expenditure in rural areas. J Beijing Univer Aeronaut Astronaut. 2022;35(05):126–138. doi:10.13766/j.bhsk.1008-2204.2022.0450

43. Avdic D, Moscelli G, Pilny A, Sriubaite I. Subjective and objective quality and choice of hospital: evidence from maternal care services in Germany. J Health Econo. 2019;68:102229. doi:10.1016/j.jhealeco.2019.102229

44. Santos R, Gravelle H, Propper C. Does quality affect patients’ choice of doctor? Evidence from England. Econ J. 2017;127(600):445–494. doi:10.1111/ecoj.12282

45. Andersen RM. Revisiting the behavioral model and access to medical care: does it matter? J Health Social Behav. 1995;36:1–10. doi:10.2307/2137284

46. Thomson S, Osborn R, Squires D, Reed SJ. International profiles of health care systems 2011: Australia, Canada, Denmark, England, France, Germany, Iceland, Italy, Japan, the Netherlands, New Zealand, Norway, Sweden, Switzerland, and the United States; 2011.

47. Wang J, Wang P, Wang X, Zheng Y, Xiao Y. Use and prescription of antibiotics in primary health care settings in China. JAMA Intern Med. 2014;174(12):1914–1920. doi:10.1001/jamainternmed.2014.5214

48. Song M, Song H. Disagreement between a public insurer’s recommendation and beneficiary’s choice of long-term care services in Korea. Article. Health Policy. 2020;124(8):881–887. doi:10.1016/j.healthpol.2020.04.012

49. He X, Li L, Bian Y. Analysis on the health care provider choice of outpatients in underdeveloped areas and its influencing factors: An empirical study in county-level hospitals and township health centers of Western China. Chin J Health Policy. 2018;11(7):61–66. 1674–2982.

50. Finkelstein A, Taubman S, Wright B, et al. The Oregon health insurance experiment: evidence from the first year. Q J Econ. 2012;127(3):1057–1106. doi:10.1093/qje/qjs020

51. Wang Z, Li X, Chen M, Si L. Social health insurance, healthcare utilization, and costs in middle-aged and elderly community-dwelling adults in China. Internat J Equit Health. 2018;1717. doi:10.1186/s12939-018-0733-0

52. Feng J, Wang Z, Song H. Self-selection and medical expenditure in Chinese health insurance system: evidence from informal employees’ participation. J Finan Res. 2018;08:85–101.

53. Liu Y, Kong Q, Yuan S, van de Klundert J. Factors influencing the choice of health system access level in China: a systematic review. Meeting AbstLancet. 2018;392:39.

54. Zeng Y, Wan Y, Yuan Z, Fang Y. Healthcare-seeking behavior among Chinese older adults: patterns and predictive factors. Int J Environ Res Public Health. 2021;18(6):2969. doi:10.3390/ijerph18062969

55. Zhang S, Chen Q, Zhang B. Understanding healthcare utilization in China through the Andersen behavioral model: review of evidence from the China health and nutrition survey. Risk Manag Healthcare Policy. 2019;Volume 12:209–224. doi:10.2147/RMHP.S218661

56. Syed ST, Gerber BS, Sharp LK. Traveling towards disease: transportation barriers to health care access. J Comm Health. 2013;38(5):976–993. doi:10.1007/s10900-013-9681-1