Sage Journals: Discover world-class research

Abstract

The well-organized medical security system plays an important role in improving public health. This study investigates the effect of the diagnosis-related groups (DRG) policy, which is a way to control cost, on medical care services, including hospital revenue status. Based on the panel data on derived from public medical institutions in Sanming City from 2014 to 2022, the average length of stay for discharged patients, number of admissions per 100 visits, and bed utilization rate were used to measure medical services status. The difference-in-difference method was used to explore the impact of the DRG policy implementation on the overall hospital operation in medical care services in Sanming City, China. We found the DRG policy introduction has a negative effect on the average length of stay for discharged patients, emergency admissions per 100 visits, and the hospital bed occupancy rate. On the contrary, the DRG policy reduces the total medical revenue in terms of the reduced drug revenue. However, the DRG policy increases the examination and the test income, average charge per emergency visit, and drug cost per discharged patient. In sum, the introduction of the DRG policy reduces the medical care system operating cost by reducing drug revenue and bed occupancy, reducing the length of hospitalization, worsening hospital revenue, and increasing the charge for examination and test, emergency visits, and drug cost per discharge. The C-DRG policy implementation in Sanming city effected improvements in public health and overall medical care system services.

Keywords

DRG policy medical service difference-in-difference model

Introduction

The medical security system is an important component of the livelihood security system (SDGs goal 3), and its establishment and development can effectively guarantee the growing medical and health needs of people and continuously improve their health. By the end of 2022, the number of participants in the national basic medical insurance (including basic medical insurance for employees and urban and rural residents) had reached 1.345 billion, with a stable participation rate of over 95%. However, while gradually achieving a wide coverage, China has encountered the challenges of rapidly increasing medical expenses and unfavorable of medical services, posing a significant challenge to China’s medical security system (Yang & Lv, 2021). Reforming the medical insurance payment methods is an effective means of controlling the growth in medical expenses and improving the quality and efficiency of medical facilities. China has successively formulated many relevant policy documents, continuously strengthening its importance and elevating it to a level that impacts the overall effectiveness of medical reform(Zhang et al., 2017).

Diagnosis-related groups (DRG), as a prospective payment method for medical insurance, play an important role in stimulating hospitals to actively control costs and promote the balanced development of medical expenses and medical service efficiency (Li & Tang, 2023). At the National DRG Payment Reform Pilot Launch Conference in 2017, the National Health and Family Planning Commission designated Sanming City as one of the three pilot cities for DRG payment reform. Sanming City officially implemented the CDRG payment reform in 2018. On March 23, 2021, during his inspection at the General Hospital of Shaxian District in Sanming City. On June 17, the first article of the “Notice on Printing and Distributing the Key Tasks of Deepening the Reform of the Medical and Health System in 2021,” issued by the General Office of the State Council was to further promote the experience of medical reform in Sanming. On October 8th of the same year, the Leading Group of the State Council for Medical Reform issued the “Implementation Opinions on Deepening the Reform of the Medical and Health System in Fujian Province, Sanming City,” which clearly stated the need to further promote the experience of medical reform in Sanming, deepen the reform of medical insurance payment methods, and promote multiple compound medical insurance payment methods. By 2025, the proportion of DRG-based payments to the total eligible hospitalization medical insurance fund expenditure is expected to reach 70%.

Concurrent with the aging of the population, citizens’ expectations of the medical care system are increasing. Nevertheless, in China, compared with developed nations such as Japan and the United States, the health care system still needs substantial improvement in terms of efficiency (Piao, 2024). During the implementation of the DRG policy, issues with effectively controlling the growth of medical expenses while ensuring the efficiency of medical services existed (Wang & Cao, 2023). However, the DRG policy has significantly benefitted cost control(Zhu, 2023). Therefore, the Chinese government has been motivated to experiment with the C-DRG policy in a representative city. Drawing from international experience, Sanming City’s C-DRG policy was one of the first implemented DRG payment systems in China. The system has been in place for nearly 7 years (introduced in 2017); hence, the successes and shortcomings not only provide valuable guidance for future healthcare reforms in China but also offer useful insights into the implementation of DRG systems in other countries.

The challenge in terms of the DRG policy is how to improve the medical care system operation while controlling medical expenses. To this end, in this study, we aimed to explore the impact of DRG policy implementation on medical services based on the difference-in-difference model. We also discuss the underlying mechanisms in terms of cost-efficiency and resource reallocation to provide policy recommendations to facilitate the selection and improvement of the new round of medical insurance payment reforms.

The reminder of this paper is as follows. The section “Literature Review” presents the literature review, while the section “Materials and Methods” summarizes the materials and methods applied in this study. The section “Results” shows the results and conclusion is presented in section “Conclusion and Policy Recommendations.”

Literature Review

A well-organized medical security system enhances citizens’ health, and medical care service improvement and development guarantee that the growing medical and health needs of people are met, continuously improving their health (SDGs goal 3). In this regard, assessing the effects of DRG policy implementation on medical care cost-control effectiveness and service improvement is crucial.

Multiple previous studies have examined the implementation of DRG policies from the viewpoint of cost-control effectiveness, and consistent results have confirmed that these policies significantly reduce medical expenses (Bertoli & Grembi, 2017; Chanturidze et al., 2016; Chen et al., 2023; Chok et al., 2018; Hu et al., 2017; Iltchev et al., 2021; Jackson et al., 2015; Lee et al., 2020; Ljunggren & Sjödén, 2001; Messerle & Schreyögg, 2023; Milstein & Schreyögg, 2024; Reid & Sutch, 2008; Schmid & Götze 2009; Wu et al., 2021, 2022; Zou et al., 2020). The effect of a DRG policy on reducing unreasonable hospitalization costs has been confirmed in major Western countries, including the United States, United Kingdom, and Germany (Iltchev et al., 2021; Mistichelli, 2023; Wu et al., 2021). In China, the implementation of such policies has resulted in significant regional differences in terms of their effectiveness, with Beijing showing noticeable cost-control benefits (Hu et al., 2017). Similarly, Li et al. (2021) studied the effectiveness of DRG policy implementation nationwide in China. The pilot results showed that in 2019, among the 30 pilot cities, the outcomes in 29 were excellent, with a general downward trend in total hospitalization expenses (Wu et al., 2022).

From the perspective of medical care service improvement, studies examining the effects of the DRG policy have obtained inconclusive results(Gartner et al., 2015; Hu et al., 2019; Koné et al., 2019; Li et al., 2021). Some studies have provided positive evaluations, suggesting that the DRG policy helps improve medical quality (Gartner et al., 2015; Koné et al., 2019). However, some argue that the DRG policy implementation has either no impact on or leads to a deterioration in medical quality. Studies by Koné et al. (2019) in Western countries found no clear evidence that the DRG policy affects medical quality, and there is no link between the length of hospital stay and medical quality. In addition, Li et al. (2021) also believe that the DRG policy has little impact on medical service efficiency. Using interrupted time-series analysis, they compared the changes in medical service efficiency indicators before and after the DRG reform and found no significant changes in the average length of stay or other indicators (Hu et al., 2019). Some scholars have found that the implementation of the DRG policy leads to a deterioration in medical quality. Through a meta-analysis, Meng et al. (2020) found that the overall effects of the DRG payment reform on the average length of stay and re-admission rates were relatively small, leading to a deterioration in medical quality. In a study in Japan, Hamada et al. (2012) found that while the DRG system reduced the total medical expenses and average length of stay, it increased the re-admission rates. In a study in Switzerland, Fässler found that 2 years after the introduction of the DRG payment system, excessive economic pressure on medical personnel led to a deterioration in medical quality (Fässler et al., 2015). Jiaying used an interrupted time-series analysis method and found that after the implementation of the DRG policy, the average length of stay and time consumption index showed a downward trend (Zhu et al., 2021)., Summarizing the research progress of DRG domestically and internationally, found that DRG did not positively incentivize medical quality, but may lead to a decrease in quality as hospitals increase patient admissions to meet DRG's bundling requirements, ultimately reducing medical quality (Zhou et al., 2023). This is because the DRG payment method primarily targets the price control of single hospitalizations, leading medical institutions to distort their medical conduct. They may increase patient admission or re-admission rates to control costs, thereby reducing medical services. Alternatively, hospitals may reduce treatment costs by selecting cheaper drugs and services or reducing the necessary services to achieve cost-control goals, resulting in a decrease in medical quality (Zhu, 2023).

This study aims to investigate the effect of the DRG policy introduction on the medical care service using the panel data on the operation of public medical institutions in Sanming City from 2014 to 2022. It applied refined indicators, such as the average length of stay for discharged patients, emergency admissions per hundred visits, and bed utilization rate, to measure medical service. The difference-in-difference method is applied in this study to investigate the effect of the DRG policy on the medical care service. This study examines the impact of DRG policy implementation on medical services and the underlying mechanisms of the implementation effects, focusing on the medical revenue, healthcare services revenue, examination and test income, drug revenue, average daily charge per bed, average charge per emergency visit, and average drug cost per discharge. Furthermore, suggestions have been proposed to promote the implementation of the in-depth DRG policy.

Compared with existing studies, the present study primarily contributes in the following ways: (1) In terms of research focus, it centers on the efficiency of medical services in public hospitals, analyzing the impact of DRG policy implementation on medical service efficiency. While previous research has not specifically refined the concept of medical service efficiency, in this study, it is measured using indicators such as the average length of stay for discharged patients, emergency admissions per hundred visits, and bed occupancy rate. (2) In terms of research methodology, while earlier studies have mostly employed case analysis or interrupted time-series methods, the present study applies the difference-in-difference model to estimate the effects of the DRG policy, providing empirical evidence on its impact on improving the efficiency of public hospital services. (3) In terms of research findings, this study enriches the existing literature by not only confirming the positive effect of the DRG policy reform on the efficiency of public hospital services but also discovering that the DRG policy improves medical service efficiency by reducing drug costs and lowering medical expenses. Moreover, we analyze the mechanisms behind efficiency improvement from the perspective of DRG-related supporting policies.

Materials and Methods

Data Source

To investigate the effects of the DRG policy on medical care system sustainability in China, data from January 2014 to December 2022 were extracted from the “Sanming Health Management Platform” and “Sanming Statistical Yearbook.” The hospital-related information were extracted from the Sanming Health Management Platform (2023), published by the Sanming Municipal Government, specifically from the “Report on the Operation of Public Medical Institutions in Sanming City.” The platform includes data on public hospitals, public community health service centers, and township health centers in Sanming City, and comprises medical income data and relevant operational indicators of public medical institutions. Additionally, relevant city-level indicators were obtained from the “Sanming Statistical Yearbook” (2023), including data on resident population, per capita disposable income, and per capita GDP.

The implementation of the policy in Sanming City began on January 1, 2018. Therefore, this study was conducted from January 2014 to December 2022. The data organization derived from the “Sanming Health Management Platform” involved a policy group and control group. The policy group comprised 12 representative public hospitals from Sanming City that implemented the DRG policy—the data from these hospitals were relatively complete before and after the policy implementation. The Sanming Municipal People’s Government has only released data from public hospitals on its official platform. The control group comprised the remaining 140 hospitals or medical institutions that did not implement the policy. Their data were tracked and observed on a monthly basis, allowing for dynamic monitoring of the effects of DRG policy implementation. Furthermore, the provided data covered different dimensions of indicators to measure the operational status of the medical institutions and met the data requirements for this study. The data mainly included information on the average length of stay, drug revenue proportion, emergency admissions per hundred visits, average charge per bed day, bed occupancy rate, average medicine cost per discharge, total medicine revenue, medicine fee, medicine proportion, total healthcare revenue, healthcare fee, examination revenue, examination fee, drug revenue, test fee, material fee, and average charge per emergency visit at the county level and above for public hospitals and grassroots medical institutions in Sanming City.

In total, the study obtained monthly operational data for 152 continuously operating medical institutions, resulting in 1,368 samples. Among these, the policy group (12 hospitals) comprised 108 samples, while the control group (140 public hospitals and medical institutions) comprised 1,260 samples. This provided a solid data source for studying the medical service efficiency of public hospitals in Sanming City. Given some deficiencies in the statistics of this database, the data were processed as follows: samples with zero or missing values for the average length of stay of discharged patients, emergency admissions per hundred visits, and bed occupancy rate as well as samples with incomplete data or interruptions in time were excluded.

The datasets derived from the “Sanming Health Management Platform” and “Sanming Statistical Yearbook” are publicly available. The study protocol was approved by the legal and ethics committee.

Model Specification

The difference-in-difference method, also known as difference-in-differences (DID), is currently widely used in academia; it not only effectively avoids endogeneity issues caused by omitted variables or reverse selection bias but also controls for common time trends experienced by the policy and control groups over time. Finally, by taking the difference-in-difference, we obtain the net effect of policy implementation. This method has been used extensively to evaluate the effects of policy implementation(Zheng et al., 2022). Therefore, this study adopted a difference-in-difference model with the baseline regression model formulated as follows:

\begin{matrix} Effec t_{it} = β_{0} + β_{1} Polic y_{it} \times Tim e_{it} + β_{2} contro l_{it} \\ + μ_{i} + ϑ_{t} + ε_{it} \end{matrix}

(1)

Where $Effec t_{it}$ denotes the $i$ th observation Among them, $Effec t_{it}$ represents the policy effect indicator of the $i$ th observation at time t, where $Polic y_{it}$ is the policy dummy variable, with a value of 1 for the policy group and 0 for the control group; $Tim e_{it}$ is the time dummy variable, with a value of 0 for the years before policy implementation (2014–2017) and 1 for the years after policy implementation (2018–2022); $Polic y_{it} \times Tim e_{it}$ represents the interaction term between the policy dummy variable and the time dummy variable, that is, the DID policy effect, with the coefficient $β_{1}$ representing the net effect of the DRG policy on healthcare service; $contro l_{it}$ represents control variables, $μ_{i}$ denotes individual fixed effects, $ϑ_{t}$ denotes time fixed effects, and $ε_{it}$ represents the random disturbance term.

An important assumption for estimating using the DID model was that the parallel trend hypothesis was satisfied in advance; that is, before the implementation of the DRG policy, there was a similar trend of change between the policy and control groups, and there was no systematic difference. Based on this, we constructed a parallel trend model as follows:

\begin{matrix} Effec t_{it} = β_{0} + \sum_{k = - 4}^{4} δ_{k} Polic y_{i} \times Yea r_{k} + β_{2} contro l_{it} \\ + μ_{i} + ϑ_{t} + ε_{it} \end{matrix}

(2)

The coefficient $δ_{k}$ of $Polic y_{i} \times Yea r_{k}$ measures the difference between the policy and control groups in the $k$ th period, where $k$ represents the relative time of policy implementation (the year of the sample minus the year of policy implementation), with actual values ranging from −4 to 4, based on the implementation of the DRG policy in various healthcare institutions.

Variable Definitions

The dependent variable was medical service related variables, which was measured using representative indicators of hospital services, including the average length of stay for discharged patients (Elliott et al., 2014; Jiang, Bai, et al., 2020), emergency admissions per hundred visits(Cheng & Ruan, 2023), and bed occupancy rate (Wang & Han, 2021). The average length of stay is an essential indicator for evaluating the service of medical institutions. This directly reflects the length of stay of hospitalized patients, with shorter lengths of stay indicating reduce the medical cost. The number of emergency admissions per hundred visits indirectly reflects the service capacity of the hospital. A decrease in hospital admissions improves the efficiency of medical services. Bed occupancy rate is an important indicator that truly and effectively reflects the level of medical service. This directly reflects the workload of hospital beds, and a decreasing bed occupancy rate can reflect an improvement in hospital work efficiency.

The core explanatory variables in this study were represented by the interaction terms of the policy and time dummy variables. According to the research design, two sets of dummy variables were set: one set was the policy dummy variable, with medical institutions that implemented the DRG policy in 2018, defined as 1, and those that did not implement the policy, defined as 0; the other set was the time dummy variable, with the implementation year of the policy being 2018. Therefore, the years 2014 to 2017 were defined as 0, and the years 2018 to 2022 were defined as 1. The coefficient of the interaction term represented the net effect of DRG policy on medical service. To minimize the estimation bias caused by other omitted variables and based on data availability and previous research (Jiang, Gao, et al., 2020; Li et al., 2021; Lin et al., 2021; Luan & Shu 2022; Zhang et al., 2021), a series of control variables were selected. The definitions of the main variables are listed in Table 1.

Table 1.

Variable Definitions.

Variable name	Variable definition
Average length of stay	Total number of bed days occupied by discharged patients/number of discharged patients
Drug revenue proportion	Drug revenue/total medical revenue × 100%
Emergency admissions per hundred visits	Admissions/emergency department visits × 100
Average charge per bed day	Hospitalization revenue/total number of bed days actually occupied by inpatients
Bed occupancy rate	Total bed days actually occupied/total bed days available × 100%
Average charge per emergency visit	Outpatient revenue/number of outpatient visits
Resident population	Population actually residing in a certain area for more than 6 months
Average medicine cost per discharge	Average charge per bed day × average length of stay
Per capita GDP	Per capita GDP = total GDP/average annual population (logarithm)
Per capita disposable income	Disposable income of residents/resident population (logarithm)
Total medicine revenue	Total revenue from medical services, examinations, drugs, materials, and physical examinations
Medicine fee	Average drug cost for inpatients
Medicine proportion	Drug cost/average medicine cost per discharge × 100%
Total healthcare revenue	Total revenue from hospital services, registration, surgeries, etc.
Healthcare fee	Average healthcare service fee for inpatients
Examination revenue	Total revenue from examinations and tests
Examination fee	Average examination fee for inpatients
Drug revenue	Total revenue from various drugs
Test fee	Average test fee for inpatients
Material fee	Average material fee for inpatients
Classification of Chinese hospitals	Hospitals in China are organized according to a 3-tier system that recognizes a hospital's ability to provide medical care, medical education, and conduct medical research
Average material fee per discharged patients	Average material fee for inpatients/average medicine cost per discharged patients × 100%
Healthcare fee per total healthcare revenue	Healthcare fee/total healthcare revenue × 100%
Medical service fee per discharged patients	Healthcare fee/average medicine cost per discharged patients × 100%
Examination and test revenue per total healthcare revenue	Examination and test revenue/total healthcare revenue × 100%

Results

Descriptive Analysis

Table 2 presents the means and standard deviations of the average length of stay, emergency admissions per hundred visits, bed occupancy rates, and related control variables for both the policy and control groups. There were significant differences between the policy and control groups.

Table 2.

Descriptive Statistics of the Variables for the Policy and Control Groups.

Variable	Policy group mean	Control group mean	Difference	t-Test	p-Value
Average length of stay (day)	8.411	5.019	−3.392	−20.002***	<.001
Emergency admissions per 100 visits (time)	4.774	3.161	−1.613	−5.226***	<.001
Bed occupancy rate (%)	80.68	38.76	−41.93	−16.917***	<.001
Classification of Chinese hospitals	2.25	1	−1.25	−102.395***	<.001
Resident population (10,000)	78.07	24.51	−53.56	−17.725***	<.001
Per capita GDP (logarithm)	11.32	11.34	0.016	0.433	.665
Per capita disposable income (logarithm)	10.07	10.09	0.022	0.813	.416
Total medical revenue (10,000 yuan)	24,439	238.2	−24000	−40.136***	<.001
Medical service revenue (10,000 yuan)	10,881	78.9	−11000	−44.832***	<.001
Healthcare fee per total healthcare revenue (%)	46.44	35.85	−10.6	−10.966***	<.001
Examination and test revenue (10,000 yuan)	6,045	26.79	−6000	−38.759***	<.001
Examination and test revenue per total healthcare revenue (%)	24.51	10.38	−14.13	−21.779***	<.001
Drug revenue (10,000 yuan)	5,483	131.2	−5400	−38.612***	<.001
Drug revenue ratio (%)	22.29	52.9	30.61	22.980***	<.001
Average charge per bed-day (yuan)	609.6	133.2	−476.4	−65.674***	<.001
Average charge per emergency admission (yuan)	166.1	60.72	−105.4	−27.106***	<.001
Average medical expenses per discharge (yuan)	5,163	703.8	−4500	−78.546***	<.001
Drug expenses (yuan)	878.2	212.4	−665.8	−45.109***	<.001
Hygiene material fee (yuan)	612.5	5.18	−607.3	−58.526***	<.001
Average material fee per discharged patients (%)	11.24	0.641	−10.6	−75.115***	<.001
Medical service fee (yuan)	740.2	162.4	−577.8	−64.158***	<.001
Medical service fee per discharged patients (%)	38.23	0.427	−37.8	−72.642***	<.001
Examination fee (yuan)	654.8	121.2	−533.6	−57.334***	<.001
Inspection fee	656.6	42.55	−614	−66.783***	<.001
Sample size	108	1,260

Note. Yuan denotes the basic unit of the Chinese Currency.

***p < .01.

Effects of the DRG Policy on Healthcare Service

Based on equation (1), the difference-in-difference method was used to estimate the effect of the DRG policy on the baseline regression. The results are summarized in Table 3. Columns (1) to (3) show the effects of the DRG policy on medical service, controlling for different variables and fixed effects. The coefficients of the Diagnosis-related group's policy are negative values statistically significant at 5% and 1% for the average length of the stay, the emergency admissions per 100 visits, and the bed occupancy rate. The results suggest that the introduction of the DRG policy has a favorable effect on the improvement of health services.

Table 3.

Effects of the DRG Policy on Healthcare Service.

Variable	Average length of stay	Emergency admissions per 100 visits	Bed occupancy rate
Diagnosis-related groups policy	−1.776***	−0.943**	−18.801***
Diagnosis-related groups policy	(0.314)	(0.415)	(4.412)
Control variables	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes
Time fixed effects	Yes	Yes	Yes
N	1,367	1,367	1,367
Model	DID	DID	DID
R ²	0.604	0.476	0.229

Note. Standard errors are in parentheses.

***

p < .01. **p < .05.

Parallel trends test and policy effectiveness

Parallel Trends Test

The first was the baseline year selected for this study. The dynamic effects are presented graphically to illustrate the efficiency of the DRG policy on healthcare services in different years, with the average length of stay for discharged patients serving as a representative analysis indicator.

Figure 1 shows that the regression coefficients fluctuate around 0 before the policy shock. This indicates that the policy and control groups exhibited similar trends before the DRG policy shock, meeting the premise of parallel trends. The short vertical lines perpendicular to the horizontal axis in Figure 1 represent the 95% confidence intervals of the δ_k regression coefficients. The significant estimated coefficients exhibit a downward trend over time. In the 3 years prior to the baseline year (2018), the coefficients were not significant (95% confidence intervals cross the horizontal dashed line at coefficient = 0). However, starting from the year of the DRG policy implementation and continuing for the subsequent 4 years, the coefficients were significant. This trend suggests that the DRG policy yields immediate effects upon implementation and maintains a certain level of sustained impact. In addition, within the sample period, the third year after DRG policy implementation exhibited the maximum effect, with a slight decrease in the fourth year.

Figure 1.

Parallel trends in average length of stay for discharged.

Effectiveness of DRG Policy

As depicted in Figure 1, since the implementation of the DRG policy, the significant estimated coefficients have exhibited a decreasing trend annually; however, after the fourth year, they gradually approach insignificance. While the DRG policy initially demonstrated an “immediate effect” upon implementation and its effects gradually increased, with the third year showing the most pronounced effect, subsequent years witnessed a decline, resulting in diminished effectiveness of the policy. The analysis, in conjunction with practical situations, suggests that the failure of the DRG policy to achieve its expected long-term effectiveness mainly stems from new issues arising during policy implementation.

First, as DRG policy is gradually implemented and deepened, government efforts and public resources are initially effective. However, after the DRG policy implementation stabilizes, there may be a lack of timely follow-up and supervision from relevant medical insurance authorities, leading to a lag or termination of government input in later stages, ultimately resulting in diminishing the effectiveness of the DRG policy.

Second, healthcare institutions’ profit-oriented behaviors may lead to moral hazards. Institutions may encourage the development of DRG-advantageous disease projects, blindly reducing costs for projects affecting profits or causing losses, and even engaging in unethical behaviors, such as deferring hospitalization of patients or reducing rehabilitation services for patients.

Finally, traditional performance evaluation indicators in healthcare institutions do not entirely align with the background of the DRG policy, failing to completely eliminate negative medical behaviors. Actions such as reducing hospitalization days, preoperative examination procedures, using low-cost drugs and medical consumables, and discharging patients prematurely decrease the medical services they receive, contradicting the original intention of the DRG policy.

To some extent, these factors affect the later-stage implementation effectiveness of the DRG policy, ultimately leading to diminished effectiveness or even termination of the policy. This conclusion provides guidance for adjusting and improving the DRG policy. In the later stages, government departments should continuously strengthen performance evaluation intensity and regulatory efforts to ensure that the DRG policy continually promotes healthcare service efficiency.

Robustness Test

Typically, there is a lag between the announcement and actual implementation of any policy, commonly observed as a lag in policy implementation. The reform of DRG medical insurance payment methods, a critical focus of recent healthcare reforms, has been implemented only in certain regions, involving collaboration among various entities, such as government health insurance bureaus, pharmaceutical bureaus, hospitals, and patients. Considering that each entity’s response to the policy has a certain implementation cycle, that is, the effect of the DRG policy on medical service may have a lag, this study conducted a lagged regression analysis on the core explanatory variables, as shown in Table 4. The results indicate that after regressing with the lagged one-period core explanatory variable data, only the significance of emergency admissions per hundred visits as the dependent variable was affected. The results with different control variables for robustness check are displayed in Appendix.

Table 4.

Lagged One-Period Regression Results (Robustness Check).

Variable	Average length of stay	Emergency admissions per 100 visits	Bed occupancy rate
Diagnosis-related groups policy	−1.519***	−0.296	−17.433***
Diagnosis-related groups policy	(0.321)	(0.374)	(4.132)
Control variables	Yes	Yes	Yes
Individual fixed effects	Yes	Yes	Yes
Time fixed effects	Yes	Yes	Yes
Model	Lagged one-period	Lagged one-period	Lagged one-period
N	1,215	1,215	1,215
R ²	0.637	0.434	0.227

Note. Standard errors are in parentheses.

***

p < .01.

However, the coefficient remained negative, indicating that the core explanatory variables still had a reducing effect. Furthermore, the coefficients of the lagged one-period regression for the average length of stay for discharged patients and bed occupancy rate were significantly negative, indicating a significant impact on medical service and further validating the robustness of the baseline regression results.

Identification of Mechanisms of DRG Policy Impact on Healthcare Service

Based on the results in Table 5, it is evident that the DRG policy significantly affected seven variables: proportion of medical service revenue, proportion of inspection and testing revenue, proportion of drug revenue, drug expenses, medical materials expense, proportion of medical service expenses, laboratory fees. None of the other variables showed significant effects.

Table 5.

Identification of Mechanisms of DRG Policy Impact on Healthcare Service.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
Variables	Proportion of medical service revenue	Proportion of inspection and testing revenue	Proportion of drug revenue	Drug expenses	Medical materials expense	Proportion of medical materials expense	Medical service expenses	Proportion of medical service expenses	Laboratory fees	Examination fees
Diagnosis-related groups policy	0.096***	0.394***	−0.128***	0.438***	0.476***	0.083	0.043	1.876***	−0.291***	−0.026
Diagnosis-related groups policy	(0.028)	(0.047)	(0.024)	(0.064)	(0.154)	(0.133)	(0.044)	(0.153)	(0.061)	(0.099)
Individual fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Time fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Model	DID	DID	DID	DID	DID	DID	DID	DID	DID	DID
Observations	1,364	1,352	1,365	1,288	582	582	1,289	1,288	1,279	1,255
R-squared	0.116	0.174	0.319	0.162	0.025	0.024	0.563	0.400	0.227	0.089

Note. Standard errors are in parentheses.

***

p < .01.

The DRG policy tends to significantly increase the average charge per emergency visit, mainly due to the fact that the DRG payment reform primarily targets inpatient costs during the acute phase of treatment. Generally, DRG payments are implemented in acute phase treatment services or within certain medical insurance scopes. Many medical institutions, in order to control costs, may transfer some inpatients to outpatient departments, leading to cost shifting between departments. The implementation of the DRG payment system incentivizes hospitals to transfer the costs of inpatient care and services from inpatient departments to outpatient and long-term care facilities. The result is a significant increase in emergency visits and charges (Zhu et al., 2021). Therefore, to prevent moral hazards, the government needs to continually strengthen institutional mechanisms by establishing clear rules and regulations to curb unethical medical practices (Li & Yu, 2022). Additionally, hospitals should enhance assessment and evaluation systems, combining both penalties and rewards, to encourage improvements in service quality and ensure the orderly progression of DRG medical insurance payment reform in China.

The DRG policy also leads to a significant increase in the average drug cost per discharge. The reform of the DRG payment system may result in medical institutions shifting some costs onto patients covered by medical insurance. Through adjustments in the internal structure of different patient groups, medical institutions can ensure that DRG treatment groups are not financially at a loss. However, this adjustment also increases the drug costs for inpatients (Zhu, 2023). The increase in medical expenses may also be due to the inherent strength of hospitals included in the DRG policy, where the number of inpatients and medical expenses are relatively advantageous. Therefore, the impact of the DRG reform on discharge costs is minimal (Li et al., 2021). Additionally, it may be due to the fact that cost reduction was a key focus of China’s healthcare reform before the implementation of the DRG policy. Many policies had already been implemented to control cost growth, but the expected reduction in costs was not achieved. The continued increase in emergency visits and inpatient costs leads to a significant decrease in healthcare service. Therefore, the increase in the average charge per emergency visit and the average drug cost per discharge cannot be regarded as effective mechanisms for improving healthcare service under the DRG policy.

The DRG policy confirms a significant decrease in medical revenue, particularly in drug. However, it has a significant positive effect on examination and test income. This indicates that the DRG policy primarily reduces drug expenses to lower overall medical costs and ultimately promotes the improvement of healthcare service. This is mainly because the implementation of the DRG policy simultaneously requires local governments to limit the use of high-value drugs, reduce the duration of drug use, substitute generic drugs for original drugs, and promote policies such as centralized procurement of drugs and medical consumables. These related policies have a positive cumulative effect with the DRG policy. Additionally, after the implementation of the DRG policy, the medical insurance payment standard for the same treatment group is fixed. The drugs used by patients within the same treatment group are part of the hospital’s costs. If excessive drug use leads to medical expenses exceeding the payment standard, the hospital will incur losses. Therefore, hospitals proactively control costs by reducing the consumption of resources during medical treatment (Wang & Cao, 2023). Hence, the government’s implementation of the DRG policy has a certain inhibitory effect on drug and medical revenue, while medical expenditure has always been an important factor affecting healthcare service (Luan & Shu, 2022). The above analysis demonstrates that the DRG policy primarily reduces drug expenses to lower medical costs, thereby promoting the improvement of healthcare service.

Conclusion and Policy Recommendations

The medical security system is an important component of the livelihood security system, and its establishment and development can effectively guarantee the growing medical and health needs of people and continuously improve their health. When the universal health coverage is discussed along with the population ageing, the cost control of the medical care service become a crucial issue. The DRG policy is a known method to reduce the medical care service referring to the previous studies. The focal point of healthcare payment reform and is currently in the process of nationwide implementation. Therefore, research on the effectiveness of DRG policy implementation in improving healthcare remains to be explored. This study primarily investigates the impact of DRG policy implementation on healthcare using data from healthcare institutions in Sanming City, Fujian Province from 2014 to 2022. Average length of hospital stay, emergency room visits per 100 admissions, and bed occupancy rate were selected as representative indicators of healthcare. The empirical analysis was conducted using a difference-in-differences model to examine the impact of the DRG policy on healthcare. The research findings are as follows:

The DRG policy can to some extent improve healthcare, mainly manifesting in significantly negative effects on average length of hospital stay, emergency room visits per hundred admissions, and bed occupancy rate. The results are consistent with the previous studies (Chen et al., 2023). Although the degree of significance varies, the results confirm the positive impact of DRG policy implementation on healthcare. The primary reason for this outcome is that DRG, as a prospective payment method, encourages medical institutions to engage in refined management, thereby strengthening quality control and effectively reducing average length of hospital stay and bed occupancy rate. Furthermore, the more detailed the DRG grouping, the more standardized the behavior of physicians becomes, driving them to reduce resource wastage in terms of medication and medical supplies, thereby enhancing healthcare.

The DRG policy can positively impact healthcare quality and service efficiency, benefiting from the implementation of other supporting policies. First, Sanming City has implemented a diverse and composite payment method, including payments by service item, bed day, and average daily surgery cost. This approach encourages medical institutions to actively shorten the average length of stay and reduce the number of admissions while ensuring healthcare service efficiency and quality. Second, Sanming City has established a close-knit urban medical group and county-level medical community system, thereby creating a unified management mechanism across eight key areas: human resources, medical operations, financial systems, performance evaluation, resource allocation, centralized procurement, information technology, and pre-paid medical insurance. This has fundamentally improved healthcare service quality at grassroots medical institutions. Third, a salary system has been introduced, with a basic annual salary supplemented by a performance-based annual salary. The wages of staff in township health centers and publicly-run village clinics have been included in the total salary budget of general hospitals, ensuring unified accounting and distribution.

Compared with DRG policies implemented in other countries, the highlight of Sanming City’s DRG policy is the performance-based annual salary system for employees. An annual salary assessment is conducted in three main steps: determining the hospital’s total wage budget, establishing the hospital’s internal distribution plan, and allocating salaries to individuals. The standard total wage budget for hospitals is uniformly set by the Sanming municipal government, and it comprises four parts: the total annual salary, performance assessment incentive funds based on DRG, incentive funds for chronic disease—integrated management, and income from family doctor contracting services. At the end of each year, hospitals allocate the total salary according to occupational team categories, and the internal distribution plan is formulated. The main occupational teams in the hospital include doctors (technicians), nursing and pharmacy staff, and administrative support staff. Individual annual salaries are composed of three parts: base points, workload points, and reward/penalty points. For different occupational teams, points are calculated based on workload and work quality. For the doctor team, workload points are calculated using resource-based relative value scales and DRG for outpatient visits and discharged patients. For the medical technical and nursing teams, points are calculated using relative value ratios for 400 medical service items and 122 nursing tasks. For the administrative support team, key performance indicators are used to calculate points based on the hospital’s strategic goals and corresponding assessment criteria.

On the other hand, although the implementation of the DRG policy promotes the improvement of healthcare, this effect gradually diminishes over time. Through parallel trend tests, it was determined that the DRG policy initially had an immediate effect, reaching its peak effectiveness in the third year after implementation, with diminishing returns thereafter. In other words, from the evolution trend of time, the impact of the DRG policy on healthcare exhibits dynamic marginal effects, with policy efficacy showing a declining trend over time.

The cause of this issue primarily stems from the medical insurance reimbursement cap. As mentioned earlier, after the implementation of the DRG policy, hospitals have linked the surplus from medical insurance with the performance salary level of the doctors. When a patient’s medical expenses exceed the amount covered by medical insurance, the hospital initially bears the excess cost, which is then deducted from the doctor’s individual performance pay. This measure increases doctors’ motivation in the short term; however, over time, they may avoid treating patients with complex conditions to mitigate risks. Alternatively, they may split hospital stays or divide hospital costs to meet the “average length of stay” and “per-hospitalization cost” assessment criteria. Additionally, owing to localized management of medical insurance in China, the quality control standards for diagnosis and treatment, as well as medical insurance fund supervision standards, have not yet been completely unified across provinces. This lack of standardization makes it difficult for medical institutions across provinces to share data comprehensively and in real time, creating regulatory barriers for local medical insurance providers.

Based on the results of this study, it is recommended to enhance and implement the DRG policy reform with a focus on its long-term and consistent effects. Strengthening the implementation and supervision of the DRG policy is crucial. Medical institutions, as the main implementers of the policy, hold the initiative in medical service provision and may exhibit moral hazards during implementation, such as by shifting costs through decommissioning hospital beds, discharging patients prematurely, or transferring treatment from inpatient to outpatient settings, thereby reducing the medical services. To eliminate the risk factors that decrease medical, it is necessary to establish and implement a medical behavior supervision mechanism, incentivize and constrain medical behaviors, encourage medical institutions to take responsibility proactively, and employ third-party audits by the government to establish a diverse medical insurance fund supervision mechanism. Real-time monitoring and reward-penalty mechanisms should be implemented for medical institutions, with rewards for those performing well under the DRG policy and penalties for those performing poorly. Continuous strengthening of regulatory oversight on the evaluation of DRG policy effectiveness will ultimately achieve Pareto optimality in the long-term effects of the DRG policy and realize the goal of sustained improvement in healthcare. Healthcare institutions can further improve their compensation distribution system by centering it around “health performance.” This would involve incorporating the recurrence rate of diseases among discharged patients or outpatients into doctors’ performance evaluations. As such, fewer illnesses or recurrences in signed patients after treatment would result in higher performance rewards for doctors, thereby incentivizing them to improve the quality of diagnosis and treatment.

Additionally, it is crucial to continuously improve the relevant supporting measures. For instance, drug and medical consumable markups affect the long-term effectiveness of DRG policy implementation. As such, the government can adopt measures such as centralizing the procurement of drugs and medical consumables, limiting the use of high-value drugs, and substituting generic drugs for original drugs to reduce drug prices and continually promote the enhancement of healthcare services. Strengthening the pre-hospital emergency system to prevent sudden cardiac death is also essential. By following the principles of comprehensive planning, resource integration, rational allocation, and efficiency improvement, efforts should be made to establish a “10-minute emergency response circle” in urban areas. This initiative would form a pre-hospital medical emergency network led by emergency centers and coordinated with emergency network hospitals. In densely populated public areas, emergency stations should be established and equipped with automated external defibrillators, and regular professional emergency demonstrations should be conducted for the public.

Finally, the DRG evaluation criteria should be dynamically adjusted. It is essential to optimize the grouping of DRG complications and comorbidities, especially as the aging population increases together with the incidence of chronic diseases in the presence of other comorbidities. The focus should be on diseases such as hypertension, diabetes, severe mental disorders, and tuberculosis, and studies on the spectrum of diseases and causes of death in hospitalized patients should be regularly conducted to improve the accuracy of case grouping for complications/comorbidities. A diversified DRG evaluation system should be established, which should include not only assessment indicators such as cost control, service capacity, and medical behavior but also appropriate indicators of patient satisfaction. Establishing a patient feedback mechanism, encouraging patients to participate in the medical decision-making process, and reducing the issue of “information asymmetry between doctors and patients” will help improve patient satisfaction and effectively enhance the long-term impact of the DRG policy on improving healthcare services.

This study had certain limitations. The Sanming Municipal People’s Government has only released data from public hospitals on its official platform, the “Sanming Public Health Information Service Management Platform.” In China, compared to private clinics, public hospitals play a crucial role and cover major health services. However, the inclusion of data from the private sector may be valuable in future studies.

Footnotes

Appendix

ORCID iDs

Qi Hu

Yue Lei

Xiangdan Piao

Statements and Declarations

References

Bertoli

Grembi

(2017). The political economy of diagnosis-related groups. Social Science & Medicine, 190, 38–47.

Chanturidze

Esau

Hölzer

Richardson

(2016). Introducing diagnosis-related groups in Kazakhstan: Evolution, achievements, and challenges. Health Policy, 120(9), 987–991.

Chen

Hua

Liu

(2022). Difficulties and countermeasures of implementing C-DRG in Sanming City. Chinese Health Economics, 41(1), 36–40.

Chen

Y. J.

Zhang

X. Y.

Yan

J. Q.

Xue-Tang Qian

M. C.

Ying

X. H.

(2023). Impact of diagnosis-related groups on inpatient quality of health care: A systematic review and meta-analysis. Inquiry, 60, 469580231167011.

Cheng

Z . X.

Ruan

X . J.

(2023). A study on the impact of C-DRG payment policy implementation on inpatient costs and efficiency in public hospitals of Sanming City. Chinese Hospital Management, 43(4), 29–33

Chok

Bachli

E. B.

Steiger

Bettex

Cottini

S. R.

Keller

Maggiorini

Schuepbach

R. A.

(2018). Effect of diagnosis related groups implementation on the intensive care unit of a Swiss tertiary hospital: A cohort study. BMC Health Services Research, 18, 1–10.

Elliott

D. J.

Young

R. S.

Brice

Aguiar

Kolm

(2014). Effect of hospitalist workload on the quality and efficiency of care. JAMA Internal Medicine, 174(5), 786.

Fässler

Wild

Clarinval

Tschopp

Faehnrich

J. A.

Biller-Andorno

(2015). Impact of the DRG-based reimbursement system on patient care and professional practise: perspectives of Swiss hospital physicians. Swiss Medical Weekly, 145, w14080.

Gartner

Kolisch

Neill

D. B.

Padman

(2015). Machine learning approaches for early DRG classification and resource allocation. INFORMS Journal on Computing, 27(4), 718–734.

10.

Hamada

Sekimoto

Imanaka

(2012). Effects of the per diem prospective payment system with DRG-like grouping system (DPC/PDPS) on resource usage and healthcare quality in Japan. Health Policy, 107(2–3), 194–201.

11.

Liu

(2019). Research on the effect of DRG-PPS reform based on interrupted time series analysis. Chinese Health Policy Research, 12(10), 23–28.

12.

Iltchev

Sierocka

Marczak

(2012). The use of DRGs in hospital management. Studies in Logic, Grammar and Rhetoric, 29, 129–142.

13.

Jackson

Dimitropoulos

Madden

Gillett

(2015). Australian diagnosis related groups: Drivers of complexity adjustment. Health Policy, 119(11), 1433–1441.

14.

Jiang

Bai

Wang

(2020). The influence of online and offline medical service quality on patients’ offline medical decisions. Management Science, 33(1), 46–53.

15.

Jiang

Gao

Guo

(2020). Evaluation of medical service efficiency and its influencing factors in China. Medicine and Society, 33(3), 32–36.

16.

Koné

Maria Zimmermann

Nordström

Simone Elger

Wangmo

(2019). A scoping review of empirical evidence on the impacts of the DRG introduction in Germany and Switzerland. The International Journal of Health Planning and Management, 34(1), 56–70.

17.

Lee

S. B.

Kim

S. I.

Choi

B. Y.

(2020). The analysis of medical care behaviors influencing new diagnosis-related groups (DRG) based payment-focused on hospitalized patients with medical illness. Korea Journal of Hospital Management, 25(2), 45–56.

18.

Tang

Xue

(2021). The impact of diagnosis-related group policy implementation on the efficiency and costs of hospitalization services in Sanming City public hospitals. Chinese Health Resources, 24(4), 393–396+404.

19.

Tang

(2023). Research on the evolution and diffusion mechanism of DRG policy in China from the perspective of policy diffusion theory: A quantitative text analysis based on 2009–2022 DRG policy. Social Security Studies, 2023(2), 62–74.

20.

(2022). Government regulation, benchmark competition, and medical insurance payment reform. Chinese Public Administration, 2022(10), 90–98.

21.

Lin

Xia

Zhu

(2021). Research on the impact of DRG payment reform on hospital operation efficiency. Health Economics Research, 38(12), 62–65.

22.

Ljunggren

Sjödén

P. O.

(2001). Patient reported quality of care before versus after the implementation of a diagnosis related groups (DRG) classification and payment system in one Swedish county. Scandinavian Journal of Caring Sciences, 15(4), 283–294.

23.

Luan

Shu

(2022). Statistical measurement of medical service efficiency in China and comparison of regional differences: Based on stochastic frontier models. Chinese Health Care Management, 39(5), 325–330.

24.

Messerle

Schreyögg

(2023). Country-level effects of diagnosis-related groups: Evidence from Germany’s comprehensive reform of hospital payments. The European Journal of Health Economics, 25(6), 1013–1030.

25.

Milstein

Schreyögg

(2024). The end of an era? Activity-based funding based on diagnosis-related groups: A review of payment reforms in the inpatient sector in 10 high-income countries. Health Policy, 141, 104990.

26.

Piao

(2024). Multinational medical care evaluation concerning aging populations. Journal of Sustainability Research, 6(2), e240011.

27.

Reid

Sutch

(2008). Comparing diagnosis-related group systems to identify design improvements. Health Policy, 87(1), 82–91.

28.

Sanming Health Management Platform. (2023). Report on the operation of public medical institutions in Sanming City (2014–2022) [EB/OL]. https://www.sm.gov.cn/zw/ztzl/shyywstzgg/ygyxbb/

29.

Sanming Statistical Yearbook. (2023). Statistical bulletin on national economic and social development of Sanming City (2015–2023) [EB/OL]. https://www.sm.gov.cn/zw/tjxx/tjgb/

30.

Schmid

Götze

(2009). Cross-national policy learning in health system reform: The case of Diagnosis Related Groups. International Social Security Review, 62(4), 21–40.

31.

Wang

Cao

(2023). Healthcare payment reform and the high-quality development of public hospitals: An analysis of perfecting the cost management mechanism of public hospitals. Price Theory and Practice, 2023(2), 80–83.

32.

Wang

Han

(2021). Population development, fiscal investment, and efficiency of local medical and health expenditure. Journal of Guizhou Normal University. Social Science Edition, 2021(3), 100–107.

33.

X.-Z.

Deng

S.-J.

(2022). Analysis of the effect of DRG payment reform on medical expenses of medical insurance patients in a tertiary hospital in Panzhihua City. Medicine and Society, 35(3), 76–79.

34.

Y. T.

Lin

Y. N.

Cheng

C. T.

C. Y.

Liao

C. H.

Hsieh

C. H.

(2021). Diagnosis-related group (DRG)-based prospective hospital payment system can be well adopted for acute care surgery: Taiwanese experience with acute cholecystitis. World Journal of Surgery, 45, 1080–1087.

35.

Yang

C.-C.

(2021). Ten years of “New Medical Reform”: The regional differences, dynamic evolution and influencing factors of China’s medical and health service efficiency. Chinese Journal of Management Science, 31(2), 162–172.

36.

Zhang

X.-X.

Sun

Y.-Y.

Zhou

B. S.

(2021). A comparative study of healthcare service efficiency in the Yangtze River Delta Region from 2010 to 2018. Chinese Health Resources, 24(1), 52–58.

37.

Zhang

Z.-Z.

Jiang

L. H.

(2017). The overall design of the National Diagnosis-Related Groups (DRG) Payment Standard. Chinese Health Economics, 36(06), 5–8.

38.

Zheng

Wang

X. J.

Sun

(2022). A study on the urban-rural medical insurance pooling policy, health risk impacts, and the performance of targeted poverty alleviation. Journal of Public Management,19(1),146–158+176.

39.

Zhu

F. M.

(2023). DRG/DIP payment reform, cost control, and impact mechanisms: A multiple-period difference-in-differences analysis based on sample data from 19 regions. Chinese Social Security Review, 7(5), 54–67.

40.

Zou

H. Y.

Zhou

Liao

Z. J.

(2020). The effects of diagnosis-related groups payment on hospital healthcare in China: A systematic review. BMC Health Services Research, 20, 1–11.

The Effects of Diagnosis-Related Groups Policy on Medical Care System Sustainability in China

Abstract

Keywords

Introduction

Literature Review

Materials and Methods

Data Source

Model Specification

Variable Definitions

Results

Descriptive Analysis

Effects of the DRG Policy on Healthcare Service

Parallel trends test and policy effectiveness

Parallel Trends Test

Effectiveness of DRG Policy

Robustness Test

Identification of Mechanisms of DRG Policy Impact on Healthcare Service

Conclusion and Policy Recommendations

Footnotes

Appendix

ORCID iDs

Statements and Declarations

References