Impact of a new case-based payment scheme on volume distribution across public hospitals in Zhejiang, China: does ‘Same disease, same price’ matter

Abstract

Background

With the implementation of the hierarchical medical system (HMS) in China, Zhejiang Province introduced an innovative payment scheme called “payment method by disease types with point counting”. This scheme was initially adopted in Jinhua in July 2017, and was later integrated with the “same disease, same price” policy in Hangzhou in January 2020. This study aimed to investigate the impact of these reforms on the distribution of health service volume.

Methods

Data were obtained from 104 hospitals, including 12 tertiary and 14 secondary hospitals from each of four regions: Jinhua (intervention) vs. Taizhou (control), and Hangzhou (intervention) vs. Ningbo (control). A total of 3848 observation points were examined using two sets of controlled interrupted time series analyses to assess the effects of this new case-based payment, without and with “same disease, same price”, on the proportion of discharges, total medical revenue and hospitalization revenue. The Herfindahl–Hirschman Index (HHI) were analyzed to evaluate changes in market competition.

Results

Following the introduction of the new case-based payment without “same disease, same price”, secondary hospitals in Jinhua experienced a significant decline in the proportion of discharges (β₆ = -0.1074, p = 0.047), total medical revenue (β₆ = -0.0729, p = 0.026), and hospitalization revenue (β₆ = -0.1062, p = 0.037) compared to those in Taizhou, while tertiary hospitals showed a non-significant increase. After incorporating “same disease, same price”, the proportion of discharges (β₆ = 0.2015, p = 0.031), total medical revenue (β₆ = 0.1101, p = 0.041) and hospitalization revenue (β₆ = 0.1248, p = 0.032) in Hangzhou’s secondary hospitals increased compared with Ningbo’s, yet the differences in both the level and trend changes between tertiary hospitals in the two cities were not statistically significant. The HHI in Jinhua (β₇ = 0.0011, p = 0.043) presented an upward trend during the pilot period of the case-based payment, while the HHI in Hangzhou (β₆ = -0.0234, p = 0.021) decreased immediately after the introduction of “same disease, same price”.

Conclusion

This new case-based payment scheme may worsen the disproportionate distribution of service volume across hospitals of different levels. While “same disease, same price” shows potential benefits, further evidence is needed to assess its effectiveness in promoting HMS. Policymakers should consider hospital interests in payment design and address unintended strategic behaviors.

Introduction

Since the initiation of hospital accreditation efforts in 1989, Chinese hospitals have been divided into three levels: tertiary, secondary, and primary hospitals [1]. Unfortunately, the concentration of well-trained medical personnel and advanced equipment in tertiary hospitals has exacerbated the disparities in the distribution of health service volume. Secondary and primary hospitals experience relatively low service volume, while tertiary hospitals often face overcrowding [2]. This imbalance results in the inefficient use of medical resources and impedes access to high-quality care for patients with severe conditions, thereby contributing to and aggravating health inequities [3]. In response, China implemented the hierarchical medical system (HMS) in September 2015 as a strategy to address inefficiencies in health services and promote a more equitable allocation of medical resources [4]. Nevertheless, by 2017, the total number of hospitalizations in China had reached 189.15 million, with tertiary, secondary, and primary hospitals accounting for 44.39%, 42.32% and 6.18%, respectively. This widening disparity highlights the ongoing challenge in establishing a rational and well-ordered healthcare delivery system [5, 6]. Several studies have identified the dominant fee-for-service (FFS) payment model as a significant factor contributing to this disparity, since it has historically incentivized tertiary hospitals to over-admit patients in pursuit of increased revenue [7, 8].

To manage the medical insurance budget and regulate health service delivery, the Chinese government has proposed transitioning from FFS to prospective payment systems, with a particular focus on diagnosis-related group (DRG)-based payment systems [9]. DRG-based payment systems have proven effective in reducing the out-of-pocket (OOP) costs and length of stay (LOS) for patients with targeted diseases, though they may not necessarily improve the quality of medical care [10,11,12]. In recent years, growing concern regarding the relationship between emerging DRG-based payment programs and the implementation of HMS [13,14,15]. As an economically developed province in China, Zhejiang redesigned an innovative case-based payment method, named “payment method by disease types with point counting”, which was first piloted in Jinhua and has been progressively adopted in inpatient services at secondary and tertiary hospitals across the province since January 2020 [16]. Further details were provided in Sect. “Institutional Background”.

Medical insurance payment methods have long been regarded as leverage to guide healthcare providers, and DRG-based payment is no exception. Theoretically, the payment standard for each DRG serves as the basis for prepaying hospitals, enabling them to estimate the maximum resource consumption threshold before patient admission [17]. To enhance their case mix index (CMI), which reflects treatment capabilities, tertiary hospitals are incentivized to proactively manage complex cases and transfer milder ones to lower-level facilities. Secondary hospitals, in turn, can admit more patients with common diseases and transfer cases that exceed their capacity to higher-level hospitals [18]. Under the DRG mechanism, the functional roles of hospitals at different levels are clarified, thereby promoting a more rational distribution of service volume. However, with the continuous exploration of DRG-based payment mechanism, researchers have found that under the current economic incentives, the design of this payment system fails to curb the expansion of certain more lucrative services and cannot achieve the desired effect [13, 19]. Furthermore, potential behaviors (e.g., patient selection, cost-shifting, readmissions) may adversely affect competition among hospitals [20, 21] (Fig. 1).

Fig. 1

The potential mechanism for the effects of DRG-based payment method on service volume distribution

Prior studies have primarily focused on the impact of DRG-based payment systems on medical costs, efficiency and quality [10, 11, 22]. While some findings provide insights into how DRG-based payment affects the implementation of HMS, most research is limited to specific hospital levels or focuses solely on the individual behavior of healthcare providers, thus empirical evidence on public hospitals at different levels remains scarce [23,24,25,26]. Additionally, most existing studies examine the average effects of DRG-based payment at specific time points following implementation, with limited consideration of the evolving trends over time.

Based on hospital-level data in Zhejiang province, China, this study was intended to investigate the impact of the new case-based payment reforms on the distribution of service volume between secondary and tertiary hospitals through interrupted time series analysis (ITSA). Moreover, this article further identified potential defects of existing DRG-based payments, with the aim of providing valuable insights for regions facing similar challenges in optimizing the distribution of health services.

Institutional background

In June 2017, the General Office of the State Council of China issued guidelines on further deepening the reform of basic medical insurance payment methods, encouraging qualified regions to actively explore the integration of point-based methods with case-based payment [27]. This is exactly the highlight of Zhejiang’s payment reform. In July 2017, following a year of pilot testing, Jinhua comprehensively implemented the “payment method by disease types with point counting” across 49 urban hospitals. To be specific, after determining the regional global budget and DRG classification, the actual reimbursement for each hospital is obtained by the following two steps [16, 28] (Fig. 2).

Fig. 2

Core components of “same disease, same price” under the new case-based payment

The first step involves calculating the points for each DRG in a specific hospital. The point volume for a specific group is determined by the average expenditure per case in this DRG relative to the average expenditure for all cases. The adjustment factor reflects the proportional relationship between the average expenditure of each DRG in a specific hospital and the average expenditure of this DRG across the region. Then, the points for a specific DRG in this hospital are equal to the point volume for a specific group multiplied by the adjustment factor. Finally, the total points for this hospital are calculated by adding or subtracting the assessment points (rewards or penalties) from the total points of all hospitalized cases.

The second step is the payment settlement for each hospital. The monetary value for one point is determined by the total DRG expenditure relative to the total number of points in the whole region. As a result, the actual reimbursement for a specific hospital are equal to the sum of the monetary value of the total points.

Due to the relatively higher medical costs, it is generally observed that the adjustment factor for tertiary hospitals tends to be higher than that for secondary hospitals when treating the same disease. Previous research has proposed that this payment mechanism may induce tertiary hospitals to siphon more patients from secondary hospitals in order to earn more points [29]. To address this issue and achieve the goal of HMS, Hangzhou, the capital city of Zhejiang, integrated the “same disease, same price” policy into the case-based payment in January 2020, namely, to cancel the adjustment factor for specific basic DRGs and adopt a unified payment standard based on the average expenditure for the same disease across hospitals at different levels [30]. Up to September 2021, Hangzhou has implemented “same disease, same price” for a total of 50 DRGs [31]. Currently, “same disease, same price” has been considered as a critical strategy to reduce disparities in service volume across hospitals and to accelerate the development of HMS. Figure 2 summarizes the calculation rules of this new case-based payment and outlines the possible consequences of implementing or not implementing the “same disease, same price” policy.

Methods

Study designing and setting

Our study employed an ecologically controlled ITSA design, incorporating pre-intervention and post-intervention periods. To minimize time-varying confounding, Jinhua and Hangzhou were respectively set as intervention groups and compared with a control group not exposed to the intervention. Considering comparability with Jinhua and Hangzhou, we selected neighboring Taizhou and Ningbo as control groups respectively, both of which are located on the eastern coast of Zhejiang Province, as shown in Fig. 3. Table 1 shows that the two sets of intervention and control groups have similar characteristics in terms of economic development and health resource allocation [32]. In addition, Hangzhou launched the payment reform in January 2020, while Ningbo did not start officially until December 2020. Accordingly, the temporal gap between the intervention and control groups in terms of implementation time points enhances the accuracy and credibility of the policy effect assessment.

Fig. 3

Geolocation of sample cities in Zhejiang Province, China

Table 1 Socioeconomic status and health resources of sample cities in 2022

Data source and sampling

We employed a code random sampling strategy to select public hospitals from both the intervention and control cities. In total, we extracted discharge and revenue data from 104 local hospitals, including 12 tertiary hospitals and 14 secondary hospitals in each sample city. To avoid potential confounding due to significant differences in technical capabilities, top-level provincial hospitals were excluded from the analysis.

For each hospital, we collected monthly data points over a span of 37 consecutive months: from January 2016 to January 2019 for Jinhua (intervention group) and Taizhou (control group), and from July 2018 to July 2021 for Hangzhou (intervention group) and Ningbo (control group). After verifying the accuracy and completeness of the data, we analyzed a total of 3,848 observation points.

Outcome variables

Considering the main purpose of the study and the availability of data, we selected three proportion indicators to measure the distribution of health service volume: 1) proportion of discharges (%); 2) proportion of total medical revenue (%); 3) proportion of hospitalization revenue (%).

Herfindahl–Hirschman Index (HHI) is commonly used to assess market concentration and potential monopolistic behavior within an industry. By summing the squares of the discharge shares from secondary and tertiary hospitals within the entire inpatient services market, we can determine the degree of competition. A higher HHI value indicates a greater degree of monopoly in the inpatient services market, which also indirectly reflects the uneven distribution of inpatient service between secondary and tertiary hospitals [2, 33,34,35].

All outcome variables were calculated based on raw data extracted from the hospital management system. Table 2 provides detailed measurement methods and descriptive interpretations for each outcome variable.

Table 2 Measurements and interpretations of outcome variables

Statistical analysis

For the outcome variables, we computed and compared means and standard deviations (SDs) before and after the intervention using Student’s t-tests. Using multi-group ITSA, we evaluated both the instantaneous and trend changes of outcome variables before and after the payment reforms [36]. The model adopts the following equation:

$${Y}_{t}={\beta }_{0}+{\beta }_{1}{T}_{t}+{\beta }_{2}{X}_{t}+{\beta }_{3}{X}_{t}{T}_{t}+{\beta }_{4}Z+{\beta }_{5}Z{T}_{t}+{\beta }_{6}Z{X}_{t}+{\beta }_{7}Z{X}_{t}{T}_{t}+{\varepsilon }_{t}$$

where ${Y}_{t}$ represents the outcome variable measured at each monthly time point t; ${T}_{t}$ is a continuous variable representing the months since the start of observation period. ${X}_{t}$ is a dummy variable with a value of either 0 or 1, representing the period before or after the intervention, respectively. $Z$ is also a dummy variable with a value of either 0 or 1, indicating whether a hospital was in the intervention or control group. ${X}_{t}{T}_{t}$, $Z{T}_{t}$, and $Z{X}_{t}{T}_{t}$ represent all interaction terms among the variables described above. ${\beta }_{1}$ and ${\beta }_{3}$ represent the slope of the control group in pre- and post-intervention periods, respectively. ${\beta }_{2}$ is the change in the trend of the control group after the intervention. ${\beta }_{4}$ and ${\beta }_{5}$ represent the differences in intercepts and slopes between the intervention and control groups prior to the intervention, while ${\beta }_{6}$ and ${\beta }_{7}$ indicate the differences in levels and slopes changes after the intervention between the two groups. Ultimately, the effects of the two different stages of the new case-based payment reforms (without or with “same disease, same price”) were estimated by observing whether the intervention groups deviated from baseline levels (${\beta }_{6}$) and trends (${\beta }_{7}$) to a greater extent than the control groups.

ITSA was conducted separately for the tertiary hospitals and secondary hospitals to assess the effects on hospitals at different levels. July 2017 was set as the intervention point for the new case-based payment without “same disease, same price” (Jinhua vs. Taizhou), while January 2020 for the payment with “same disease, same price” (Hangzhou vs. Ningbo). To address autocorrelation and potential heteroskedasticity, we employed ordinary least squares regression with Newey-West standard errors [36]. All statistical analyses were performed using Stata 14.0.

Results

Characteristics of outcome variables

Before the introduction of the new case-based payment scheme, the four outcome variables were similar between the intervention and control groups. Overall, notable changes were observed after the payment reforms (Tables 3 and 4). Figure S1 in the Additional file 1 displays the time series graphs for outcome variables in the intervention and control groups.

Table 3 Monthly mean (SD) values of proportion indicators in intervention and control groups

Table 4 Monthly mean (SD) values of the HHI in intervention and control groups

Table 3 reports the monthly mean (SD) values for the proportion indicators of tertiary and secondary hospitals in the intervention and control groups. After the reform of the new payment without “same disease, same price”, the proportion of discharges (t = -5.2920, p < 0.001), total medical revenue (t = -4.4960, p < 0.001) and hospitalization revenue (t = -3.5437, p = 0.003) in secondary hospitals in Jinhua significantly decreased, in contrast to the control group (Taizhou). Meanwhile, the three indicators for tertiary hospitals increased throughout the study period. Following the payment reform combined with the “same disease, same price” policy, the proportion of total medical revenue (t = 2.2641, p = 0.037) and hospitalization revenue (t = 1.6400, p = 0.119) in secondary hospitals in Hangzhou increased to some extent, whereas the corresponding indicators for tertiary hospitals exhibited a downward trend.

Table 4 presents the monthly mean (SD) values of the HHI for the intervention and control groups. After the reform without the “same disease, same price” policy, the HHI in Jinhua (t = 4.1014, p = 0.001) increased significantly compared to the control group (Taizhou). However, following the implementation of the payment reform with the “same disease, same price” policy, the HHI in Hangzhou (t = -2.6955, p = 0.015) decreased significantly compared to the control group (Ningbo).

Effects of the new case-based payment without “same disease, same price” on volume distribution

Table 5 and Fig. 4 show the controlled ITSA results for the proportion indicators of secondary and tertiary hospitals in Jinhua and Taizhou. Prior to the payment reform, the intervention and control group exhibited similar and nearly parallel proportion of discharges, total medical revenue and hospitalization revenue, indicating that the two groups of regions are highly comparable. After the reform of the new case-based payment without the “same disease, same price” policy, unlike control group (Taizhou), secondary hospitals in Jinhua experienced a significant immediate decrease in the average proportion of discharges (β₆ = -0.1074, p = 0.047), total medical revenue (β₆ = -0.0729, p = 0.026), and hospitalization revenue (β₆ = -0.1062, p = 0.037). Despite the increase in proportion indicators for Jinhua’s tertiary hospitals, the overall differences in both level and trend changes between two groups were not statistically significant.

Table 5 Coefficients (standard errors) for effects of the new case-based payment without “same disease, same price”

Fig. 4

Interrupted time series graphs for the proportion of discharges, total medical revenue and hospitalization revenue in Jinhua and Taizhou. a, b, c are proportion indicators of secondary hospitals; d, e, f are proportion indicators of tertiary hospitals

Effects of the new case-based payment with “same disease, same price” on volume distribution

Table 6 and Fig. 5 present the controlled ITSA results for the proportion indicators of secondary and tertiary hospitals in Hangzhou and Ningbo. Before the policy intervention, the differences in both the levels and trends between the intervention and control group were not statistically significant. Following the implementation of the new payment combined with “same disease, same price”, compared to the control group (Ningbo), the overall changes in tertiary hospitals were not remarkable, whereas the average proportion of discharges (β₆ = 0.2015, p = 0.031), total medical revenue (β₆ = 0.1101, p = 0.041) and hospitalization revenue (β₆ = 0.1248, p = 0.032) for secondary hospitals in Hangzhou displayed a large increase within a short period. Nevertheless, the differences in trend changes of these three indicators between the intervention and control group were not significant.

Table 6 Coefficients (standard errors) for effects of the new case-based payment with “same disease, same price”

Fig. 5

Interrupted time series graphs for the proportion of discharges, total medical revenue and hospitalization revenue in Hangzhou and Ningbo. a, b, c are proportion indicators of secondary hospitals; d, e, f are proportion indicators of tertiary hospitals

Effects of the new case-based payment on inpatient services market competition

Table 7 and Fig. 6 present the results of controlled ITSA regarding the degree of competition in the inpatient services market across the two sets of regions. Prior to the policy intervention, although there were substantial differences between the initial intercepts in Jinhua and Taizhou, as well as the trends observed in Hangzhou and Ningbo, the payment reform notably affected the HHI. After the reform, the HHI in Jinhua (β₇ = 0.0011, p = 0.030) showed an upward trend, in contrast to Taizhou. In comparison to Ningbo, the HHI in Hangzhou (β₆ = -0.0234, p = 0.021) significantly decreased in the short term following the implementation of the case-based payment combined with the “same disease, same price” policy.

Table 7 Coefficients (standard errors) for effects of the new case-based payment on HHI

Fig. 6

Interrupted time series graphs for inpatient services market competition in intervention and control groups. a is the HHI in Jinhua and Taizhou; b is the HHI in Hangzhou and Ningbo

Discussion

This study employed a quasi-experimental design to evaluate the impact of the new case-based payment on service volume distribution across public hospitals in Zhejiang, China. Our findings suggest that the new case-based payment system may exacerbate the disparities in service volume between secondary and tertiary hospitals. While the “same disease, same price” policy has the potential to alleviate the over-concentration of patients in tertiary hospitals, its effectiveness in addressing the underlying imbalance in health service distribution remains uncertain.

Following the reform of the new case-based payment scheme without “same disease, same price”, the proportion indicators for secondary hospitals decreased significantly, and the HHI showed a noticeable upward trend. According to previous studies, hospitals are strongly incentivized to increase service volume under DRG-based payment system due to the potential for enhanced profitability within DRGs [37, 38]. Similarly, under this new case-based payment scheme, we speculate that the differential compensation caused by the adjustment factor in the insurance reimbursement formula may encourage tertiary hospitals to leverage their competitive advantages to accumulate more points, thereby placing secondary hospitals at a significant competitive disadvantage in the inpatient services market [29]. As a result, this scheme appears to have further entrenched the concentration of inpatients in tertiary hospitals, which hinders the effective implementation of the HMS.

Encouragingly, we observed that after the introduction of “same disease, same price” in the new case-based payment system, the proportion indicators for secondary hospitals all increased, while the parameter β₆ for these indicators for tertiary hospitals became negative, leading to a significant drop in the HHI. This suggests that the service volume is more evenly distributed compared to the past, with less concentration in large tertiary hospitals. One possible explanation is that with the “same disease, same price” policy, tertiary hospitals are assigned the same number of points as secondary hospitals when treating patients from the same DRG [39, 40]. Due to higher medical costs and reduced benefits, tertiary hospitals may consciously limit the intake of patients from these groups. Simultaneously, the more equitable insurance payment system also incentivizes secondary hospitals to improve operational efficiency and, to some extent, treat a greater number of patients with common diseases.

Notably, after the introduction of the “same disease, same price” policy, the reduction in the corresponding proportion indicators for tertiary hospitals was not as pronounced as expected, which means the increase in service volume at secondary hospitals may not be entirely attributable to patient referrals from tertiary hospitals. Previous studies have indicated that DRG-based payment can incentivize violations such as code creep (i.e., the substitution of less severe diagnoses with more severe ones) and unnecessary admissions [41,42,43]. Therefore, some tertiary hospitals may continue to treat common cases for profit, and the increased service volume in secondary hospitals might stem from their admission of patients who require only outpatient care. This finding highlights the need for policymakers to incorporate referral rates into the performance evaluation system for public hospitals and to closely monitor the accuracy of medical record documentation and coding.

Furthermore, compared to immediate changes, the long-term trends in outcomes resulting from the new case-based payment scheme with the “same disease, same price” policy seem insignificant. Firstly, the short-term effect may be due to the fact that the policy has just been implemented, and hospitals have yet to adopt strategic actions, thereby showing the expected effects. However, since DRG-based payment is still in its initial stage in China, and effective supervision is not in place, large tertiary hospitals may make strategic adjustments later, such as reducing LOS, decomposing hospitalizations, increasing readmissions or exploiting other inappropriate avenues to offset revenue losses [44,45,46,47,48,49]. Hence, it is imperative for policymakers to continuously regulate hospitals’ emerging strategic behavior and evaluate the system barriers. Additionally, our study was conducted in relatively developed cities where high-quality healthcare resources are primarily concentrated in large tertiary hospitals. Patients in these cities generally have greater financial capacity, enabling them to afford the high hospitalization costs associated with these hospitals. As a result, the impact of the “same disease, same price” policy on their choice of inpatient care may be less significant [50,51,52].

Limitations

This study has several limitations. First, the interpretation of the results should be interpreted with caution, as we selected a developed province for our sample, meaning that the policy effects may not be generalizable to other regions. Second, we analyzed only 19 months of post-policy implementation data, limiting our ability to assess the longer-term effects of the payment reforms. Third, due to the unavailability of referral data, we could not determine whether the increase in service share at secondary hospitals under the “same disease, same price” policy was due to patient transfers from tertiary hospitals, despite observing non-significant parameters for tertiary hospitals. Finally, during the study period, the COVID-19 pandemic and related policy changes (e.g., encouraging residents to delay non-emergency medical visits and concentrating medical resources on severe cases) occurred. Although a quasi-experimental design was employed to control for these effects, intervention and control groups may have responded differently, potentially leading to the deviations in the estimated effects of the payment reforms [53,54,55].

Conclusion

The new case-based payment scheme may exacerbate the disparities in service volume distribution between secondary and tertiary hospitals. While the “same disease, same price” policy holds potential for increasing the service volume in secondary hospitals, the decline in tertiary hospitals remains minimal. Thus, further research using long-term observational data is essential to determine whether this policy can effectively address the imbalanced service distribution. Additionally, policymakers should consider a more equitable allocation of benefits across hospitals when designing payment schemes, and continuously monitor hospitals’ strategic responses to payment reforms.