The Impact of Medical Insurance Payment Policy Reform on Medical Cost and Medical Burden in China

Abstract

Medical insurance that pays for the medical expenses of the insured is regarded as being maximally fair, but often leads to the excessive use of medical services, medical expenses rising too rapidly, and the waste of health resources. The deductible and reimbursement ratio defines the part of the cost to be borne by the patient to reduce moral hazard, preventing adverse selection and easing the burden of medical expenses. But will improving the deductible or reducing the reimbursement ratio reduce medical expenses and reduce the medical burden of the insured? By using panel data from 2011 to 2019, this paper examines the effect of the deductible and reimbursement ratio on reducing medical expenses. The following conclusions were obtained: First, the impact of the deductible on medical costs and the medical burden is significant and negative. Specifically, lowering the deductible can reduce medical costs and reduce the medical burden of patients. Second, the impact of the reimbursement ratio on medical expenses and the medical burden is significant and negative, that is, with all other factors remaining equal, an increase in reimbursement ratio will result in a reduction in medical expenses and medical burden. Third, the combined effect of the deductible and reimbursement ratio has a significant impact on both medical expenses and medical burdens. Specifically, changes in the deductible and reimbursement ratio in the same direction can reduce both the medical expenses and the burden on patients. Fourth, based on an analysis in which 31 provinces were divided into economically developed and less developed areas, it was found that the deductible and reimbursement ratio have a certain impact on medical expenses and medical burden; in relatively less developed areas, with increasing per capita GDP, the flow of patients in large hospitals will inevitably increase, and the overall medical expenses will rise.

1. Foreword

Medical care and national health are related to the national economy, as well as people’s livelihoods. China’s medical and health undertakings have always been widely valued by the government, and healthcare for residents and public health issues have also been of increasing significance to the people. COVID-19 swept the world in 2020, highlighting the importance and necessity of developing medical and health infrastructure. It is also a global problem. According to the Outline of the “Healthy China 2030” Plan, it is necessary to improve the medical insurance system and the universal medical insurance system. By 2030, the universal medical insurance system will be mature and finalized, and the negotiation and risk-sharing mechanism between medical insurance agencies and medical institutions will have been improved. Improving the medical insurance system is not only an important basis for the effective use of health resources, ensuring people’s physical and mental health, and promoting social stability; it is also an important cornerstone in the construction of the medical and health system [1]. In 2020, the total health cost increased by 76.47% compared with 2015. The proportion of the total health cost as a percentage of GDP has increased year by year. In 2020, it increased by 1.12 percentage points, and the per capita health cost has increased rapidly—by 72.65% compared with 2015. (As shown in

Table 1. List of national health expenses.

Metric	2015	2016	2017	2018	2019	2020
Total health cost (RMB 100 million)	40,974.6	46,344.9	52,598.3	59,121.9	65,841.4	72,306.4
Total health costs accounted for the GDP (%)	6.0	6.2	6.36	6.43	6.64	7.12
Per capita health cost (RMB 100 million)	2980.8	3351.7	3783.8	4237.0	4702.8	5146.4

).

In 2009, the Opinions of the CPC Central Committee and The State Council on Deepening the Reform of the Medical and Health System (hereafter referred to as the “New Medical Reform Plan”) were released, and China launched a new round of reform of the medical and health system to govern medical and health undertakings in a new way. Improving the medical insurance system is not only an important basis for the effective use of health resources, ensuring people’s physical and mental health, and promoting social stability, but it is also an important cornerstone in the construction of the medical and health system. The full utility of medical insurance is influenced by the design of the medical insurance system itself. The payment of medical insurance expenses is the key link in the system of medical insurance operation, and it is also one of the most important and basic aspects of medical insurance. Depending on the composition and operation process of the medical insurance system, medical insurance payment methods include the medical insurance buyer payment method and the supplier payment method, and the insured contributes a part of the medical expenses after receiving the services provided by the medical institution as defined by the deductible payment method, the proportional payment method, the ceiling line payment method, etc. [2]. The different medical insurance systems implemented by countries around the world demonstrate the range of medical insurance, fully reflecting the spectrum from cases where the insured are to pay all medical expenses to free medical treatment. But these extreme cases often lead to excessive utilization of medical services, medical expenses rising too fast, and the waste of health resources; therefore, to prevent this phenomenon, different countries have adopted all kinds of cost-sharing methods to replace full payment to effectively control medical expenses. The problem of medical expenses is also a concern of scholars all over the world, as evidenced in the work by Frees, E.W., J. Gao, and M. A. Rosenberg (2011) [3], a study predicting the frequency and amount of health care expenditure. Scholars G.P. Clemente, N. Savelli, G. A. Spedicato, and D. Zappa (2022) [4] believe that monitoring general practitioner prescribing costs is an important topic to efficiently allocate national health insurance resources. Liu, L., R. L. Strawderman, M. E. Cowen, and Y.-C. T. Shih. (2010) proposed a flexible “two-part” random effects model that was used to analyze pharmacy cost data on 56,245 adult patients clustered within 239 physicians in a mid-western US-managed care organization. [5].

In medical insurance policies, the deductible and the reimbursement ratio are the two most commonly used methods in China. The deductible policies and the reimbursement ratio policies describe the own costs incurred by the patients with the aim of reducing moral hazard, preventing adverse selections, and easing the burden of medical expenses. Reducing the deductible is equivalent to improving the average reimbursement ratio, but how should one determine whether improving the deductible and reducing the reimbursement ratio will be able to reduce the medical expenses and medical burden of patients?

This paper uses practice data from 2011 to 2019 to directly investigate the effect of the deductible and the reimbursement ratio on reducing medical expenses and medical burdens. The main contributions of this paper include the following three points: First, according to the statistical caliber of health expenditure, data on the deductible and the reimbursement ratio were collected to obtain an overview of medical expenses and the medical burden. Data availability is a common problem in health spending micro-research due to the fact that provincial medical insurance payment and medical expense data disclosure content are not the same thing, so this paper employed the China health and family planning statistical yearbook and the medical insurance bureau website provided by the public data for sorting. Second, the deductible and reimbursement ratio was quantified with respect to medical expenses, and the impact of medical insurance payment policies on medical expenses was directly evaluated. The ultimate goal of reforming the medical insurance payment mode is to improve the health level of residents and reduce the medical burden of individuals. The impact on medical costs directly reflects the effect of the reform of the medical insurance payment mode since the new medical reform. Based on provincial panel data, this paper explores the economic performance of the reform of medical insurance payment methods, enriches the research content in the field of medical insurance reform, and also answers questions as to whether medical insurance reform has made an impact. Third, the research results on the performance of medical insurance payment methods obtained in this paper can be used to evaluate the effectiveness of the new medical reform to a certain extent, while also providing empirical support for the reduction of medical expenses, as well as guidance and a decision basis for the policy trend of continuing medical insurance payment reform and the reduction of medical insurance expenses in the future. Fourth, there is a lot of research literature on the deductible and reimbursement ratio, but little of the literature combines the deductible and reimbursement ratio to study the influence of their joint action on medical expenses and medical burden. This paper innovatively adopts the cross term of deductible and reimbursement ratio to analyze the influence of their joint action on medical expenses and medical burden, further clarifying the impact of the simultaneous implementation of different medical insurance policies on medical expenses and medical burden. To a certain extent, this paper enriches the literature on medical insurance policy research.

As mentioned above, the medical problem is a worldwide problem. As one of the emerging countries, China’s reform experience in medical insurance payment policy and other aspects can provide inspiration to many developing countries. By studying the impact of the deductible and reimbursement ratio on medical expenses and medical burden, this paper hopes to provide some guidance to other countries’ medical insurance policies and research related to medical expenses.

2. Literature Review

In medical insurance policy, the deductible policy and the reimbursement ratio policy are two commonly used practices in China. Both the deductible policy and the reimbursement ratio policy reduce moral hazard and ease the burden of medical expenses by the patients themselves. It is of practical significance to study the impact of the deductible and the reimbursement ratio on medical expenses. In the existing literature, scholars have arrived at different conclusions regarding the impact of the deductible and the reimbursement ratio on medical expenses or medical burden, but the overall conclusions mainly belong to two categories.

First, if the deductible decreases or the reimbursement ratio increases, medical expenses will increase, and the same goes for individual medical burden; that is, the higher the proportion of individual payment, the lower the amount of medical services that will be consumed.

In foreign studies, on the effect of the medical insurance coverage rate (out-of-pocket ratio) on medical expenses, Pauly (1968) [6] indicated that when there is a proportion of medical insurance cost sharing, individuals consume medical services less, and that the best healthcare policy would be one in which a mechanism is established whereby the state and the patient pay jointly and an insurance deductible standard is established. Other studies have opined that medical insurance can share medical costs for residents because of its risk-sharing function. For example, the famous Rand experiment showed that increased reimbursement ratios increased patients’ use of medical services (Manning et al., 1987) [7]. Feldman and Dowd (1991) [8] proposed a cost-sharing system in which the insured and the medical institutions shared the medical expenses, thus preventing the insured from overconsuming medical services and reducing the phenomenon of doctors inducing demand. Subsequent scholars began to study how much the specific co-insurance rate should be to achieve the specific effect of reducing medical costs, e.g., Feldstein (1994) [9], pointed out that in co-insurance, when the co-insurance rate (out-of-pocket ratio) is typically in the range of 20% to 25%, the demand for medical services is significantly reduced, and the international co-insurance ratio (out-of-pay ratio) of medical insurance is generally controlled at about 20%. Dulamsuren (2013) noted that Japan had adopted several reforms to control increases in healthcare costs, the co-pay rate (out-of-pocket ratio) of the insured increased from 10% to 20% in 1997 and from 20% to 30% in 2003. The data from 1996 to 2007 were divided into 1996–2002 and 2002–2007 [10].

Subsequently, scholars investigated different policies in medical insurance, such as the deductible policy. In the United States, after enterprises replaced their original insurance policies with a high deductible, the medical expenses of employees decreased significantly (Brot-Goldberg et al., 2017) [11]. Studies of drug purchase programs in US geriatric programs also showed that nonlinear reimbursement rules such as the deductible are able to significantly reduce medical costs (Dalton et al., 2020) [12]. Remmerswaal et al. (2019) compared two Dutch medical insurance plans and found that adopting the deductible policy was able to save more medical costs [13].

From foreign studies on the impact of the deductible and reimbursement ratio on medical expenses, targeted studies have been conducted in the United States, Japan, the Netherlands, and other countries, which once again shows that medical insurance policy and medical expenses are a worldwide problem, and scholars from all countries pay close attention to it; moreover, from the early impact study to the later specific impact degree study, step by step. Of course, it is also the orientation of medical issues. However, although the impact of medical insurance policies on medical expenses is relatively mature in various countries, there are few examples from the literature on the impact of the combination of the deductible and reimbursement ratio of medical insurance policies on medical expenses or medical burden.

Domestic scholars have performed a greater amount of study on reimbursement ratio policies, but few scholars have studied the deductible policy. Wang Zhen et al. (2019) used the medical insurance data of urban workers in Shanghai and found that for elderly people around 60, every 1 percentage point reduction in the patient’s out-of-pocket payment ratio increased the number of inpatients by about 0.355% [14]; Da Yuqi et al. (2020), using the homepage data of a hospitalized city, found that reducing the deductible and increasing the reimbursement ratio significantly improved the total inpatient medical expenses and service utilization intensity [15]; Xiang Hui et al. (2020) studied the impact of an increase in the hospitalization reimbursement ratio caused by an increase in the pooling level, and found that this caused a significant increase in hospitalization expenses [16].

It can be seen that, compared with foreign research, Chinese scholars’ research is relatively simple and has only started in recent years. It can also be seen that Chinese scholars’ research mainly focuses on the impact of a certain medical insurance policy on medical expenses.

Second, the deductible and the reimbursement ratio policy have an impact on medical expenses. Most studies at home and abroad show that the co-insurance rate of medical insurance, including the deductible and the reimbursement ratio policy, can reduce medical expenses while increasing the out-of-pocket ratio, while other scholars only report the impact on medical expenses. Siviero (2012) studied the practice of medical payments in some regions of Italy in 2001, analyzing the impact of medical payments on medical consumption expenditure based on time series data, and conducted case studies in two regions of Italy: Piedmont and Puglia. The results showed that the introduction of co-payments did not achieve the expected effect of reducing health consumption expenditure [17]. Huang and Gan’s (2017) study also found that changes in healthcare reimbursement ratios significantly affected medical expenses [18]. Feng Jin et al. (2021) believed that the deductible policy had a significant impact on patient groups with intermediate medical expenses of a relatively large number [19].

Theoretical analyses of the necessity of adopting cost-sharing performed by many foreign scholars provide a theoretical basis for the positive role of cost-sharing systems in terms of controlling the excessive growth of medical insurance costs and enhancing the cost awareness of patients. Research into the optimal personal out-of-pocket ratio provides specific theoretical guidance for the formulation of medical insurance cost-sharing policies in various countries. Most scholars believe that increasing the deductible and reducing the reimbursement ratio can effectively reduce medical expenses and reduce the medical burden of patients. Therefore, the authors believe that China should learn from the relevant experience of foreign countries by its specific national conditions. But health insurance policies, such as starting lines and reimbursement rates, are often implemented simultaneously, and the effects of both should be tested simultaneously to be more realistic.

Based on this research, under the background of the new health reform, a quantitative analysis of 31 provinces in China, representing an innovation in the process of research analysis using payment line and reimbursement ratio, was performed to analyze the influence of payment line and reimbursement ratio on medical expenses and medical burden, providing a certain guiding significance in the formulation of the medical insurance policy.

The reason why medical service consumers face moral hazard with respect to overconsuming medical services after participating in medical insurance is mainly that the demand and price elasticity of medical services are greater than zero. If patients buy medical insurance, patients’ demand for medical services increases as the medical insurance co-insurance rate gradually decreases. It is easy to overconsume medical services. Excessive consumption of medical services can result in the inefficiency of medical services and the unnecessary loss of social welfare.

Deductible and reimbursement ratios are two ways of structuring payment in order to prevent medical insurance providers from excessively providing medical services to consumers. Setting a deductible aims to control moral hazard in the context of minor illness, and emphasizes those with poor health sharing medical expenses because the expected medical expenses would not exceed the deductible of medical expenses with stronger constraints, because their reimbursement ratio is 0, while above the deductible, the insured have a higher-than-average annual reimbursement ratio. Both patients reduce the moral hazard by paying part of the cost. Reducing the deductible is equivalent to increasing the average reimbursement ratio and reducing the burden on patients. Therefore, generally lowering the deductible is similar to raising the reimbursement ratio, which will lead to an increase in medical expenses [19]. This is supported by research both at home and abroad. In the United States, after companies replaced their original insurance items with higher starting payment lines, their medical expenses decreased significantly (Brot-Goldberg et al., 2017) [11]. Studies of drug purchase programs in the US geriatric care program also show that nonlinear reimbursement rules such as deductible lines can significantly reduce medical costs (Dalton et al., 2020) [12].

In the case of increased reimbursement levels, medical insurance is not only an insurance system but also a welfare benefit. The absolute value of the price elasticity of the medical service demand of insured people is greater than the absolute value of income elasticity. High compensation levels stimulate the medical service demand of insured people, and will also induce an increase in the medical expenses of the insured group (Zhou Jian, Shen Shuguang, 2010) [20], or the treatment of minor diseases using more expensive medical equipment, medicines, or services (Zhao Man, Lu Guo, 2007a) [21].

Based on this, this paper puts forward the following assumptions about the relationship between medical insurance payment policies that employ a starting payment line and those that employ a reimbursement ratio, and medical expenses and medical burden:

H1. 

When other factors are controlled, lowering the deductible will increase medical expenses and increase the medical burden of patients.

H2. 

When other factors are controlled, increasing the reimbursement ratio will increase medical expenses and increase the medical burden of patients.

As posed by assumptions H1 and H2, it is known that lowering the deductible and increasing the reimbursement has the same effect on medical expenses and medical burden, so hypothesis H3 is proposed:

H3. 

Under the condition of controlling other factors, both the deductible and the reimbursement ratio of medical expenses will increase the medical expenses and increase the medical burden of patients.

3. Empirical Design

3.1. Data Sources

The data in this paper were acquired from the “China Statistical Yearbook” and the “China Health and Family Planning Statistical Yearbook” from 2011 to 2019, the official websites for the human resources and social security bureaus of provinces and cities, or the official websites of medical insurance bureaus of provinces and cities, and included data from 31 provinces.

3.2. Variables

3.2.1. Dependent Variable

In combination with the results of previous studies, this paper used the ratio of the total income of health institutions and population as the variable for measuring the cost of medical care per capita, and the proportion of per capita healthcare expenditure to per capita disposable income was used as a measure for individual medical burden. Through the normality test, the p value of per capita medical cost was 0.0778 and that of individual medical burden was 0.8396, both greater than 0.05. Both variables passed the normality test.

3.2.2. Independent Variables

(1) The deductible of the medical insurance is the standard threshold following which treatment is paid for by basic medical insurance. Depending on the medical insurance fund, the fund protects personnel from common individual burdens with respect to hospital medical treatment costs, which is a principle of the basic medical treatment insurance system reform. The fund covers treatments for personnel that can be found in designated medical institution “catalogs” that define the scope of eligible medical treatment in hospitals. If the cost is within the scope, after the patient first bears the initial portion themselves, the medical insurance fund will pay in accordance with the regulation proportions. Hospitalization medical expenses below the starting payment standard is borne by the individual patients.

(2) Reimbursement ratio

For on-the-job workers, medical insurance submits an expense account proportion, the hospitalization cost that the worker that attends medical treatment insurance enjoys is the proportion that can reimburse according to the regulation.

Controlled variable. In order to ensure the reliability of the empirical results, it was necessary to control for the influence of economic development, supply of medical resources, demographic characteristics, environment, etc. This paper selected GDP per capita, number of beds per thousand, medical and health institution beds, the proportion of private hospitals, insurance participation rate, literacy rate, gender ratio, dependency ratio of the elderly population, medical health and family planning expenditure, and local general public budget expenditure as control variables, the specific definitions and descriptive statistics of which are shown in

Table 2. Descriptive statistical results of the main variables.

Class	Variable	Symbol	Average Value	Standard Deviation	Least Value	Crest Value	Skewness
Explanatory variable	Deductible	Deductible	963.2518	797.5328	146.8500	6567.1100	0.5766
	Reimbursement ratio	Reimbursement	0.7431	0.1646	0.3500	0.9700	−0.7675
	Payment items of the deductible and the reimbursement ratio	DERT	735.7659	703.8037	113.8087	5713.3860	0.3001
Explained variable	Per capita medical expenses	Perhexp	3345.6870	1689.0630	1220.9100	13,766.7700	0.4766
	Personal medical burden	Hexpincome	0.0570	0.0158	0.0176	0.1013	0.0615
Controlled variable	GDP per capita	Pergdp	54,017.9300	26,223.3600	16,413.0000	164,220.0000
	Number of beds per thousand	Bed	5.1132	1.0096	2.7150	7.5434
	Medical and health institution beds	Insbeds	22.3411	14.8824	0.8352	64.0147
	The proportion of private hospitals	Private	0.4937	0.1529	0.0485	0.7888
	rate of joining insurance	Coverage	0.5206	0.2838	0.1286	1.1686
	illiteracy rate	Illiteracy	6.0825	6.2019	1.2300	41.1800
	sex ratio	Sexratio	104.9575	4.1063	95.7700	123.1700
	The dependency ratio of the elderly population	Dependency	13.8321	3.4092	6.7100	23.8200
	Medical and health care and family planning expenditure	HCexpenditure	368.9198	241.2193	35.3000	1579.6010
	Local general public budget expenditure	Budget	4748.1460	2683.6000	705.9100	17,297.8500

. Among them, the average “deductible” was 963.2518, and the average “reimbursement ratio” was 0.7431. The skewness of the main variables shows that the absolute values are all less than 1, so it can be considered that these main variables are approximately subject to Normal distribution.

3.2.3. Empirical Model and Estimation Methods

(1) The empirical model

There is a certain time stickiness in medical expenses, and the current values may be affected by the previous period (Finkelstein et al., 2009) [22]. In order to avoid missing important variables, the explained variables and lag items need to be added to the regression. Therefore, this paper uses a dynamic panel model for GMM estimation to test the effect of the deductible and reimbursement ratio on medical expenses. Only when the deductible is reached will reimbursement start at the given ratio, and the reimbursement ratio of the deductible is zero, so the self-payment line and reimbursement ratio appear at the same time in practice.

At the same time, although the deductible and the reimbursement ratio are two medical insurance payment methods, they have a certain correlation. Therefore, another regression equation is set to add a cross-multiplication, which is able to reflect the impact of the two on the medical cost and the medical burden.

$$y_{it}={\beta }_0+{\beta }_{1}L.y_{it}+{\beta }_{2}Deductible+{\beta }_{3}Reimbursement+{\beta }_{4}DERT+{\beta }_k{X}_{it}+{\epsilon }_{it}$$

(1)

where $y_{it}$ represents the explained variable, including medical expenses and medical burden, L. represents the first-order lag of the corresponding variables; Deductible and Reimbursement represent the “deductible” and “reimbursement ratio”; DERT represents the deductible and reimbursement cross-multiplying term; $X_{it}$ represents other control variables, including GDP per capita, number of beds, beds in private hospitals, participation rate, illiteracy rate, sex ratio, elderly ratio, health and family planning expenditure and local general public budget expenditure; and ${\epsilon }_{it}$ represents the residual.

(2) Estimation method

For the dynamic panel models, the most commonly used estimation method is differential GMM with the system GMM, but only for short panels with larger cross-section dimensions N and smaller time dimensions T. Because the sample data in this paper comprised a short panel with larger N (N = 31) and smaller T (T = 9), the GMM estimation method can be used directly. We found that the first-order autoregression coefficients of the explained variables were all significantly close to 1, indicating that the explained variables have a strong sequence correlation. When using the difference GMM estimator, it is susceptible to weak instrumental variables. Therefore, this paper mainly uses the systematic GMM estimator.

Usually, we use least squares estimation (OLS) for the panel model, but the influence of the deductible and reimbursement ratio may have endogenous problems with respect to medical expenses; secondly, the effects of the deductible and reimbursement ratio on medical expenses may be interrelated, and the explanatory and explained variables are causal, greatly leading to endogeneity.

Based on the above considerations, in order to reduce the impact of the endogenous problem, this paper considers the introduction of the tool variable method for estimation. Before introducing the tool variable, we first check the endogeneity of the explanatory variable and conduct the Hausman test on the two groups of models using the deductible, reimbursement ratio, and their cross-multiplying terms as explanatory variables (see

Table 3. Results.

Project	Model (1)
	Perhexp	Hexpincome
statistics	37.02	53.04
p-value	0.0001	0.0000

). The result is that the models formed by the two explained variables in Model (1) reject the original hypothesis. With reference to the systematic generalized moment estimation method proposed by Arellano and Bond (System GMM, SGMM) [23,24], in Model (1), we used lag phase 2 of the explained variables (Perhexp, Hexpincome) as the tool variable, and adopted the dynamic panel GMM to estimate the two-step system GMM.

4. Analysis of Empirical Results

The rapid increase in medical expenses and the excessive burden of personal medical treatment are prominent problems in the development of China’s health infrastructure, and they have become one of the social problems that are highly significant to the public. Medical insurance payment policies adjust the medical burden assumed by patients. By lowering the deductible or increasing the reimbursement ratio, is possible to effectively reduce personal medical expenses, reduce personal medical burden, and truly alleviate the problem of “expensive medical treatment”? This paper tries to answer the above questions by examining the impact of the deductible and reimbursement ratio on medical expenses and medical burdens. The regression results are shown in

Table 4. Regression results of medical costs.

Variable	Model (1)
	OLS
L. Perhexp	1.1089 *** (0.0299)
Deductible	−0.2484 ** (0.1172)
Reimbursement	−258.023 * (146.6804)
DERT	0.3209 ** (0.1402)
Pergdp	0.0010 (0.0010)
Bed	−31.6983 (23.6033)
Insbeds	0.0934 (1.6593)
Private	−53.2527 (116.4395)
Coverage	−31.6381 (48.6713)
Illiteracy	0.4493 (2.4855)
Sexratio	−7.0451 * (4.1880)
Dependency	−10.7287 * (5.5077)
HCexpenditure	0.27855 (0.1825)
Budget	−0.0116 (0.0194)
Constant	1213.874 ** (556.0202)
N	248

Note: ***, **, and * are significant at 1%, 5%, and 10%, respectively; AR (2) is the p-value of the residual second-order sequence correlation test, and the original hypothesis is “no second-order sequence correlation”.

and

Table 5. Regression outcomes of healthcare burden.

Variable	Model (1)
	Two-Step System, the GMM
L. Hexpincome	0.8648 *** (0. 0741)
Deductible	−0.00001 * (0.00001)
Reimbursement	−0.0332 * (0.0178)
DERT	0.00002 * (0.00001)
Pergdp	−1.02 × 10−8 (6.02 × 10−8)
Bed	0.0009(0.0028)
Insbeds	0.00006 (0.0001)
Private	−0.0052 (0.0200)
Coverage	0.0036 * (0.0021)
Illiteracy	−0.0001 (0.0003)
Sexratio	−0.00007 (0.0001)
Dependency	0.0003 (0.0007)
HCexpenditure	0.00003 (0.00002)
Budget	−2.83 × 10−6 (1.91 × 10−6)
Constant	0.0392 (0.0282)
N	217
AR(1)	0.0000
AR(2)	0.441
Sargan	0.961

Note: *** and * are significant at 1%, and 10%, respectively; AR (2) is the p-value of the residual second-order sequence correlation test, and the null hypothesis is “no model residual term sequence correlation of the second order”; Sargan is the p-value of the overidentification test, and the null hypothesis is “the tool variable is valid”.

.

4.1. The Impact of Medical Insurance Policy on Per Capita Medical Expenses

In Model (1), the p-values of AR (1) and AR (2) in the residual autocorrelation test were 0.004 and 0.106, and the two-step system GMM regression equation showed a significant first-order sequence correlation and no second-order correlation, indicating the plausibility of the chosen instrumental variables. However, the p-value of the Sagan statistic (Sargan) of the two-step system GMM was 0.007, thus rejecting the null hypothesis that the tool variable is valid.

In Model (1), the estimated coefficient was 1.1089 and significant, which fully indicates that the lag effect of the explained variable is significant. It is therefore of great importance to study medical costs using dynamic panel models. Accordingly, dynamic panels were used to analyze per capita medical cost as the explained variable.

On the basis of the regression results, the coefficients of the core explanatory variable deductible (Deductible), reimbursement ratio (Reimbursement), and multiplication (DERT) in the model (1) were −0.2484, −258.023, and 0.3209, respectively. The deductible and reimbursement ratio was significant at 5%, and the DERT was significant at 1%. The fitting degree R^2 of the equation is 0.9911, and the fitting degree is very good. The results of the regression were as follows:

First, the influence of the deductible on medical expenses was significantly negative, in that a decrease in the deductible led to an increase in per capita medical expenses, which is consistent with hypothesis H1. The reason for this is that reducing the deductible significantly increases the average reimbursement ratio. An increase in the average reimbursement ratio leads to an increase in the number of patient visits [23]. For large hospitals, in particular, an increased reimbursement ratio inevitably leads to an increase in medical expenses of large hospitals due to moral hazard.

Second, the impact of the reimbursement ratio on medical expenses was significantly negative, whereby an increased reimbursement ratio led to a decrease in per capita medical expenses, which is contrary to hypothesis H2. There are two main reasons for this. First, because the sample data were obtained from provincial capital cities, they were dependent on the economic conditions of various places. On the basis of the regression results, it can be seen that the per capita GDP was positively related to per capita medical expenses. Nevertheless, moral hazard is a factor, and when the reimbursement ratio increases, medical expenses increase. However, due to differences in regional economic conditions, even if the reimbursement ratio is high, the per capita medical expenses will not increase in economically less developed regions. Second, the increase in reimbursement ratio and the reduction in personal medical burden are due to the effect of price elasticity of demand. The insured often chooses services with relatively low prices, thus reducing the price of health services and driving the insured to seek medical services with lower prices, in order to be able to control their medical expenses. Whether this incentive effect is effective depends on the level of reimbursement ratio and the price elasticity of medical service demand [24].

Third, changes in the deductible and reimbursement ratio in the same direction led to a significant increase in per capita medical expenses. This is consistent with the separate impacts of the deductible and the reimbursement ratio on medical expenses. The impacts of the deductible and the reimbursement ratio on medical expenses were significantly negative, and the traffic items of the two were only positive and significant, which indicates that in cases where the deductible and the reimbursement ratio of a policy work together, or where the deductible acts indirectly through the reimbursement ratio or the reimbursement ratio through the deductible, the deductible and reimbursement ratio will have played a significant role in the change in medical expenses.

$$\frac{{\partial }_{Perh\exp }}{{\partial }_{Deductible}}={\beta }_2+{\beta }_{4}Reimbursment$$

(2)

$$\frac{{\partial }_{Perh\exp }}{{\partial }_{Reimbursment}}={\beta }_3+{\beta }_{4}Deductible$$

(3)

The regression result coefficient is embedded in Model (1). Then, the first derivative of the deductible on the per capita medical expenses and the reimbursement ratio on the per capita medical expenses can be calculated, that is, the marginal effect of the deductible on the per capita medical expenses, and the marginal effect of the reimbursement ratio on the per capita medical expenses, as shown in (2) and (3).

According to the regression results, β₂ was −0.2484, β₃ was −258.023, β₄ was 0.3209, the marginal effect of the deductible on per capita medical expenses was −0.2484 + 0.3209 $Reimbursment$, the marginal effect of reimbursement ratio on per capita medical expenses was −258.023 + 0.3209 $Deductible$, the value range of the deductible was 0 to infinity, and the value of the reimbursement ratio was 0 to 1. ① When the minimum reimbursement ratio is 0, that is, the medical expenses do not reach the deductible, the marginal effect of the deductible on per capita medical expenses is negative; when the maximum reimbursement ratio is 1, that is, the medical expenses exceed the deductible, the marginal effect of the deductible on per capita medical expenses is positive. ② When the deductible is 0, that is, all medical expenses are reimbursed, the marginal effect of reimbursement ratio on per capita medical expenses is negative; when the deductible is infinite, no medical expenses are reimbursed, and the marginal effect of the reimbursement ratio on per capita medical expenses is positive.

Therefore, in order to reduce per capita medical expenses through the deductible and reimbursement ratio, it is necessary to inversely adjust the deductible and reimbursement ratio, or retain the deductible unchanged while increasing the reimbursement ratio, or retain the reimbursement ratio unchanged while reducing the deductible.

4.2. The Impact of Medical Insurance Policy on Medical Burden

In Model (1), the p-values of AR (1) and AR (2) in the residual autocorrelation test were 0.0000 and 0.441. The first-order sequence correlation of the two-step GMM regression equation was significant, and there was no second-order correlation, indicating that the selected tool variables were reasonable. The p-value of the Sargan statistic of the two-step system GMM was 0.961, accepting the null hypothesis that the tool variables are valid.

The regression results showed that the estimated coefficient of the medical cost lag item was 0.8648 and significant, thus showing that the lag effect of the explained variable was significant. It is therefore extremely important to study medical expenses using the dynamic panel model. Therefore, this paper adopts dynamic panel estimation for the model, the explanatory variable of which was personal medical burden.

In Model (1), the coefficients of the deductible, the reimbursement ratio, and their DERT are −0.00001, −0.0332, and 0.00002, respectively. The deductible and the reimbursement ratio are significant at the 10% level, and the DERT is significant at the 10% level. The following can be observed from the regression results:

First, the impact of the deductible on the medical burden was significantly negative, that is, a decrease in the deductible resulted in an increase in personal medical burden, which is consistent with assumption H1. The reason for this is that reducing the deductible significantly increases the average reimbursement ratio. The increase in the average reimbursement ratio increases the number of patient visits [23]. For large hospitals, in particular, the increase in reimbursement ratio inevitably increases personal medical burden due to the moral hazard of patients.

Second, the impact of the reimbursement ratio on medical burden was significantly negative; that is, when the reimbursement ratio increased, there was a decrease in personal medical burden, which is contrary to hypothesis H2. This is because the sample data obtained were medical insurance data of provincial capital cities and were dependent on the economic conditions of various places. The regression results show that the per capita GDP is negatively related to the personal medical burden. Although moral hazard is present, the reimbursement ratio will increase, thus resulting in an increase in the expenditure for medical care. However, due to the differences in regional economic conditions, even if the reimbursement ratio is high, the medical care expenditure in economically less developed regions will not go up. Furthermore, an increase in the reimbursement ratio and the reduction in personal medical burden is due to the role of price elasticity of demand. The insured often chooses services that have relatively low prices in order to reduce the price of health services, prompting the insured to seek medical services with lower prices, so as to control their medical expenses. Whether this incentive role is effective depends on the level of reimbursement ratio and the price elasticity of medical service demand.

Third, changes in the deductible and reimbursement ratio in the same direction will lead to a significant increase in personal medical burden. This is consistent with the separate impacts of the deductible and reimbursement ratio on medical expenses. The impact of the deductible and reimbursement ratio on medical burden is significantly negative, and the traffic items of the two are only positive and significant, thus demonstrating that the deductible and reimbursement ratio plays a significant role in changing the medical burden when the policy of the deductible and reimbursement ratio work together or when the deductible indirectly acts through the reimbursement ratio.

By calculating the first derivative of the medical burden model, it is possible to calculate the first derivative of the deductible on the personal medical burden to the deductible and the personal medical burden on the reimbursement ratio, that is, the marginal effect of the deductible on the personal medical burden and the marginal effect of the reimbursement ratio on the personal medical burden, as shown in (4) and (5).

$$\frac{{\partial }_{H\exp income}}{{\partial }_{Deductible}}={\beta }_2+{\beta }_{4}Reimbursment$$

(4)

$$\frac{{\partial }_{H\exp income}}{{\partial }_{Reimbursment}}={\beta }_3+{\beta }_{4}Deductible$$

(5)

In terms of regression results, β₂ was −0.00001, β₃ was −0.0332, β₄ was 0.00002, the marginal effect of the deductible on the personal medical burden was −0.00001 + 0.00002 $Reimbursment$, the marginal effect of the reimbursement ratio on the personal medical burden was −0.0332 + 0.00002 $Deductible$, the value range of the deductible was 0 to infinity, and the reimbursement ratio was 0 to 1. (1) When the lowest reimbursement ratio is 0, that is, the medical expenses do not reach the deductible, the marginal effect of the deductible on the personal medical burden is negative. When the reimbursement ratio is a maximum of 1, that is, the medical expenses exceed the deductible, the marginal effect of the deductible on the personal medical burden is positive. (2) When the deductible value is 0, that is, all medical expenses will be reimbursed, the marginal effect of the reimbursement ratio on the personal medical burden is negative. The marginal effect of the personal medical burden decreased by −0.0332 when the reimbursement ratio of one company was increased. When the value of the deductible is infinite, that is, no medical expenses will be reimbursed, regardless of the reimbursement ratio, the marginal effect on the personal medical burden is positive.

Therefore, in order to reduce the personal medical burden by means of altering the deductible and reimbursement ratio, it is necessary to inversely adjust the deductible and reimbursement ratio, or retain the deductible unchanged while increasing the reimbursement ratio, or retain the reimbursement ratio unchanged while reducing the deductible, which is consistent with the practice of reducing per capita medical expenses.

(3) Regional analysis

It can be seen from the above analysis that the regression results are affected by per capita GDP. This paper divides the sample data of 31 provinces into two small samples on the basis of per capita GDP. Firstly, the samples were sorted according to the size of their GDP. During the 9 years from 2011 to 2019, those provinces that remained in the top 15 for at least 8 years were taken as a small sample. These more developed provinces are referred to as sample 1, mainly including Liaoning, Shandong, Hubei, Beijing, Tianjin, Jiangsu, Shanghai, Zhejiang, Inner Mongolia, Fujian, Guangdong, Jilin, Chongqing, Shaanxi, and Ningxia. The rest constitute another sample, made up of less developed regions, and they are referred to as sample 2. Next, a regression analysis was performed on the two samples.

The following results (see

Table 6. Small-sample Haumann test.

Project	Sample 1		Sample 2
	Perhexp	Hexpincome	Perhexp	Hexpincome
statistics	18.89	48.92	5.71	35.36
p-value	0.0261	0.0000	0.8389	0.0001

) were obtained from the Haumann test of two samples:

With respect to the results of the model, sample 1 totally refuted the null hypothesis, the sample 2 medical cost model supported the null hypothesis, while the medical burden refuted the null hypothesis, and sample 2′s medical cost was estimated using dynamic panel OLS.

The p-values of AR (1) and AR (2) in the residual autocorrelation test of sample 1 were 0.915 and 0.997, and there was no first-order or second-order sequence correlation in the GMM regression equation of the two-step system. The p-value of the Sargan statistic of the two-step system GMM was 0.961, thus rejecting the original assumption that the tool variable was valid; therefore, OLS regression was adopted. Moreover, the square of regression model R of the two samples was 0.9927 and 0.9849, respectively, which passed the equation fitting test.

The estimated coefficients of the medical cost lag in the regression results in

Table 7. Regression results for a small sample of medical costs.

Variable	Sample 1	Sample 2
	OLS	OLS
L. Perhexp	1.1411 *** (0.0402)	0.9634 *** (0.0433)
Deductible	−0.3344 (0.2527)	0.0679 (0.1696)
Reimbursement	−287.9567 (368.6366)	−48.6911 (174.7755)
DERT	0.4248 (0.3247)	−0.0873 (0.1993)
Pergdp	−0.0005 (0.0020)	0.0110 *** (0.0029)
Bed	−88.5949 ** (35.9134)	57.5606 ** (26.3999)
Insbeds	0.1259 (2.3544)	0.9457 (2.1230)
Private	−354.3504 (235.7753)	90.0188 (126.7292)
Coverage	57.1432 (77.8652)	−52.3570 (59.3172)
Illiteracy	5.6946 (12.8605)	9.7559 *** (2.2533)
Sexratio	−18.6533 *** (6.2081)	10.7349 ** (4.8934)
Dependency	−13.2598 * (7.7552)	5.4474 (6.6281)
HCexpenditure	0.1249 (0.3007)	0.1317 (0.3186)
Budget	0.0079 (0.0282)	−0.0367 (0.04345)
Constant	2831.349 *** (830.4246)	−1488.26 ** (599.8418)
N	120	128
AR(1)
AR(2)
Sargan

Note: *, **, *** denote significance at 10%, 5%, and 1% statistical levels, respectively; figures in parentheses are standard errors.

for sample 1 and sample 2 were 1.1411 and 0.9634, respectively, which are significant, and show that the lag effect of the explained variable was significant. It is highly important to study medical expenses through a dynamic panel model.

First, in sample 1 and sample 2, the regression coefficients of the deductible, reimbursement ratio and their multiplying terms (DERT) were −0.3344, −287.9567, 0.4248; and 0.0679, −48.6911, and −0.0873; respectively, but they were not significant.

It can be seen from the regression results that the impact of the deductible line on medical expenses in the economically developed areas of sample 1 was negative, that is, when the deductible line decreased, there was an increase in per capita medical expenses, which is consistent with the assumption of H1. The reason for this is that reducing the deductible significantly increases the average reimbursement ratio. The increase in the average reimbursement ratio leads to an increase in the number of patient visits [23]. For large hospitals, in particular, the increase in the reimbursement ratio inevitably leads to an increase in medical expenses due to moral hazard.

In sample 2, where the economy is relatively less developed, the impact of the deductible on medical expenses was positive, that is, when the deductible increased, the medical expenses rose. In less developed areas, the per capita medical expenses were significantly positively correlated with the per capita GDP. Further analysis is shown in

Table 8. Return results of sample 2 payment line and GDP on medical costs.

Variable	OLS
Deductible	0.3056 *** (0.1183)
Pergdp	0.0644 *** (0.0110)
Degdp	−0.00001 *** (3.20 × 10−6)
Bed	511.4561 *** (75.9194)
Insbeds	−25.7773 * (15.5846)
Private	170.6748 (545.8219)
Coverage	254.7593 (163.5759)
Illiteracy	27.4465 *** (4.5324)
Sexratio	4.2851 (12.7528)
Dependency	8.7172 (22.1219)
HCexpenditure	−0.0154 (0.9455)
Budget	0.0209 (0.1682)
Constant	−2647.788 * (1569.587)

Note: *, *** denote significance at 10% and 1% statistical levels, respectively; figures in parentheses are standard errors.

, where Degdp is the multiplier between the deductible and the per capita GDP. The R square of the equation is 0.9107, and the fitting is good. The regression results show that the multiplier between the deductible and the per capita GDP was significantly negative, that is, with increasing per capita GDP, lowering the deductible increases medical expenses, which is consistent with both theory and practice. With increasing per capita GDP and a decrease in the deductible, the flow of patients into large hospitals inevitably increases, and the overall medical expenses rise.

$$Perh\exp ={\beta }_0+{\beta }_{1}Deductible+{\beta }_{2}Pergdp+{\beta }_{3}Degdp+{\beta }_k{X}_{it}+{\epsilon }_{it}$$

(6)

Secondly, the impact of the reimbursement ratio on the medical burden is negative in both economically developed and economically less developed regions, that is, the reimbursement ratio increases, and the per capita medical cost decreases, which is contrary to the assumption of H2. This is mainly because the sample data were obtained from provincial capitals and are dependent on the economic conditions of various places. The regression results show that the per capita GDP is significantly positively related to per capita medical expenses. Although moral hazard is present, at a higher reimbursement ratio, medical costs increase. However, due to differences in the regional economic conditions, even if the reimbursement ratio is high, the per capita medical expenses will not increase in economically less developed regions.

In sample 1, AR (1) was 0.842 and AR (2) was 0.925, and in sample 2, AR (1) was 0.246 and AR (2) was 0.725, both of which fail to pass the test of the rationality of the tool variables of the two-step system GMM. The regression results show that the estimation of the medical cost lag was significant, thus indicating that the lag effect of the explained variables is significant. It is highly important to study medical expenses using the dynamic panel model. Therefore, the paper adopted dynamic panel estimation for the model, the explanatory variable of which was personal medical burden. Additionally, the squared R of the two sample models are 0.9380 and 0.9392, respectively, passing the fitting test of the equation.

In sample 1 and sample 2, the coefficients of the deductible, reimbursement ratio, and their DERT in

Table 9. Regression results of a small sample of healthcare burden.

Variable	Sample 1	Sample 2
	OLS	OLS
L. Hexpincome	0.9300 *** (0.0484)	0.8858 *** (0.0555)
Deductible	−6.97 × 10−6 (8.58 × 10−6)	−1.05 × 10−6 (6.31 × 10−6)
Reimbursement	−0.0095 (0.0119)	−0.0014 (0.0065)
DERT	9.54 × 10−6 (0.00001)	8.10 × 10−7 (7.42 × 10−6)
Pergdp	−3.66 × 10−10 (1.93 × 10−8)	−1.17 × 10−7 (9.12 × 10−8)
Bed	0.0015 * (0.0008)	0.0026 ** (0.0011)
Insbeds	0.00007 (0.00005)	−7.79 × 10−6 (0.00008)
Private	0.0007 (0.0057)	−0.0025 (0.0052)
Coverage	0.0011 (0.0020)	0.0026 (0.0020)
Illiteracy	−0.00009 (0.0003)	−0.000000009
Sexratio	0.0001 (0.00008)	−0.0003 (0.0002)
Dependency	−0.0002 (0.0002)	0.0002 (0.0003)
HCexpenditure	2.47 × 10−6 (6.30 × 10−6)	0.00002 (9.55 × 10−6)
Budget	−6.50 × 10−7 (5.19 × 10−7)	−1.98 × 10−6 (1.30 × 10−6)
Constant	−0.0035 (0.0172)	0.0395 (0.0253)
N	120	128
AR(1)
AR(2)
Sargan

Note: *, **, *** denote significance at 10%, 5%, and 1% statistical levels, respectively; figures in parentheses are standard errors.

were −6.97 × 10⁻⁶, −0.0095, 9.54 × 10⁻⁶, −1.05 × 10⁻⁶, −0.0014, and 8.10 × 10⁻⁷, respectively.

The following observations can be made from the regression results:

First, the impact of the deductible on the medical burden was negative, that is, when the deductible decreases, the personal medical burden increases, which is consistent with assumption H1. The reason for this is that reducing the deductible significantly increases the average reimbursement ratio. Increasing the average reimbursement ratio increases the number of patient visits [19]. For large hospitals, in particular, the increase in reimbursement ratio inevitably increases the personal medical burden due to the moral hazard to patients.

Second, the impact of the reimbursement ratio on the medical burden is negative, that is, when the reimbursement ratio increases, the personal medical burden decreases, which is contrary to the assumption of H2. This is mainly because the sample data were obtained from provincial capitals and are dependent on the varying economic conditions of different places. The regression results show that the per capita GDP is negatively related to the personal medical burden. Although moral hazard is present, when the reimbursement ratio increases, there is an increase in medical care expenditure. However, due to the differences in regional economic conditions, even if the reimbursement ratio is high, medical care expenditure does not increase in economically less developed regions.

5. Conclusions and Policy Recommendations

Since the new medical reform, China’s medical insurance system has been further improved. The deductible and the reimbursement ratio, as the two payment methods employed by medical insurance providers, are tools conducive to adjusting the proportion of medical expenses shared between the government and individuals. This paper made the best use of public data and used provincial panel data from 2011 to 2019 to study the performance of government medical and health expenditure since the new medical reform, and to examine the impact of two provider payment methods, i.e., the deductible and the reimbursement ratio, on reducing medical expenses and personal medical burden. The following conclusions were drawn: first, the deductible has a significantly negative impact on medical expenses and medical burdens. Specifically, lowering the deductible reduces medical expenses and reduces the medical burden of patients. Second, the impact of the reimbursement ratio on medical expenses and the medical burden was significantly negative. That is, a higher reimbursement ratio results in lower medical expenses and reduced medical burden if all other factors remain unchanged. Third, the combined effect of the deductible and reimbursement ratio had a significant impact on both medical expenses and medical burden. Specifically, changes in deductible and reimbursement ratio in the same direction can reduce medical expenses and reduce the burden on patients. Fourth, according to the regional analysis, it was found that the deductible and the reimbursement ratio had a certain impact on medical expenses and medical burdens. In relatively less developed economic areas, when the per capita GDP increases and the deductible decreases, the flow of patients in large hospitals inevitably increases, and the overall medical expenses increase.

The research reported in this paper provides illumination regarding the evaluation of the effectiveness of the new medical reform with respect to the aim of reducing medical expenses, reducing the medical burden, and the future trends of the payment methods in policies provided by health insurance providers in the future:

First, adjust the deductible in a timely fashion in accordance with the local economy and medical policy. By adjusting the deductible, the impact on medical expenses and medical burden can achieve the desired effect. The level of the deductible directly affects the utilization efficiency of medical services and the medical behavior of the insurer [2]. This study also found that adjustment of the deductible line, by influencing the insurer’s behavior, ultimately affects the medical expenses and the burden of medical expenses; reductions in the deductible line urge the insurer to increase medical treatment behavior as a result of moral hazard, indirectly increasing the average reimbursement ratio, promoting the growth of medical expenses, and increasing the medical burden of the insurer. Raising the deductible line can effectively curb the moral hazard arising from insurance, thereby reducing unnecessary medical consumption, saving medical resources, and diverting patients to different levels of hospitals, although a deductible line that is too high can deter patients from paying for treatment when they are ill, which is contrary to the original intention of medical insurance. Further research found that in places with different levels of economic development, the impact of the deductible on medical expenses is different, especially in less developed areas. With increasing GDP, lowering the deductible increases medical expenses. Therefore, when formulating the deductible policy, the level of the deductible should be adjusted according to the local economic situation to achieve the due effect of the deductible policy.

Second, the current reimbursement ratio policy can reduce medical expenses and medical burdens, but this requires timely adjustment in line with economic fluctuations. As one payment method employed by providers, the reimbursement ratio directly affects the medical treatment behavior of the insured. The current reimbursement ratio adopts the same standard for different populations and income conditions, which may lead to inequity in health services. It can be seen from the regression analysis that the reimbursement ratio has a significant impact on medical expenses and medical burden and is related to the local economic situation. If the economic situation is relatively less developed, increasing the reimbursement ratio does not stimulate the insured’s behavior. This is also similar for families. If family income is relatively low, the cost of medical care is less, especially for families with serious illnesses under poor economic conditions. If the reimbursement ratio is not sufficiently high, the medical treatment will be abandoned or the medical treatment will be terminated in advance, thus reducing the effectiveness of medical treatment. Therefore, the reimbursement ratio can be increased level by level with the increase in medical expenses. At the same time, it is necessary to adjust the reimbursement ratio in a timely fashion in line with the economic development of each region. Due to the different levels of social and economic development in the region, the gap between rich and poor is large, resulting in huge differences in the economic conditions and physiological conditions of people in different regions, as well as the demand for medical insurance [25]. Therefore, it is necessary to adjust the reimbursement ratio in different regions depending on their consumption level and the medical needs of people in different regions. It should be adjusted in accordance with local per capita medical expenses and annual income per capita. For regions with relatively low levels of economic development, the scheme of appropriately reducing the deductible and increasing the reimbursement ratio should be adopted to reduce the burden of medical expenditure on low-income people, increase the investment of medical insurance funds in regions with relatively low levels of economic development, and help to achieve equitable distribution of social resources [1].

Third, the correlation between the deductible and the reimbursement ratio has a common impact on medical expenses and medical burdens. When formulating medical insurance compensation policies, it is necessary to consider the impact of the relationship between the two. The regression results show that performing an inverse adjustment of the deductible and reimbursement ratio, or increasing the reimbursement ratio while the deductible remains unchanged, or decreasing the deductible when the reimbursement ratio remains unchanged, can significantly reduce medical expenses and reduce the medical burden for patients.

Limitations

At present, the problem of medical insurance policies and medical expenses is more prominent in China, which is also a common problem faced by developing countries. Therefore, based on the research on China’s situation, this paper uses China’s panel data from 2011 to 2019 to investigate the effect of the starting line and reimbursement ratio on reducing medical expenses. Some experiences and practices of the article may be limited by the Chinese cultural context. However, as mentioned above, the issue of medical insurance policy and medical expenses is very realistic and worthy of in-depth study both in practice and in theoretical research. The article will further study the situation of other countries in the later stage, hoping to make breakthroughs in the field of cross-cultural research in the future. It is hoped that this topic will be extended to the whole world to solve health problems faced by people around the world.