The universal zero markup drug policy and gastric cancer hospitalization expenses: an analysis of trends and influencing factors in Shanghai from 2014 to 2021

Abstract

Background

The increasing hospitalization expenses for Gastric Cancer (GC) impose a notable economic burden on society. Although the Chinese government has implemented the Universal Zero Markup Drug Policy (UZMDP) to control the growth of hospitalization expenditures, costs have continued to rise. Identifying the factors influencing the hospitalization expenses of GC patients is crucial. This study aimed to analyze the trends and factors influencing hospitalization expenses of GC patients in Shanghai from 2014 to 2021.

Methods

Data were sourced from the Health Network of Shanghai Economic Information Center. We employed interrupted time series analysis (ITSA) to analyze the trends in various medical expenditures before and after the implementation of the UZMDP. The degree of association between various medical expenditures and hospitalization expenditures of GC patients was calculated by using the new grey relational analysis (GRA). Furthermore, multiple linear regression was employed to identify the influencing factors.

Results

A total of 23,335 participants were included in this study. The ITSA results showed an increasing trend in hospitalization expenses following the implementation of UZMDP. Drug expenses decreased immediately post-UZMDP, but subsequently began to rise over time. Post-UZMDP, the expenses of medical consumables, examination, and healthcare services all showed an upward trend. The new GRA indicated that the influencing factors of hospitalization expenses, in order of importance, were expenses for drugs, consumables, healthcare services, and examination. Multivariable linear regression analysis revealed that GC patients aged 60 or below incurred lower hospitalization expenses (Coefficient = -780.06, P = 0.0398). However, factors associated with increased hospitalization expenses included longer length of stay (Coefficient = 1753.01, P < 0.001), surgeries (Coefficient = 29,047.26, P < 0.001), and hospitalization in the tertiary hospitals (Coefficient = 25,485.19, P < 0.001) or secondary hospitals (Coefficient = 17,755.12, P < 0.001).

Conclusions

Hospitalization expenses of GC patients in Shanghai have been rising annually from 2014 to 2021. Despite the implementation of the UZMDP, drug expenses remain a major factor in escalating hospitalization expenses. The hospitalization expenses of GC patients are significantly influenced by several factors, including the demographic characteristics of patients, the severity of diseases, and the levels of hospitals. These findings provide a basis for more effective management of the hospitalization expenses for GC patients.

Background

Gastric cancer (GC), a common cancer of the gastrointestinal tract, features high morbidity, significant mortality, and poor prognosis [1, 2]. China reports the highest cases of GC globally [3]. In 2018, annual treatment costs for GC in China was approximately 23.5 billion Chinese yuan (CNY), ranking third among all cancer-related expenditures [4]. The average per capita medical cost for GC patients was nearly 22,000 CNY, ranking second only to lung cancer [4]. Previous studies have highlighted that high treatment costs are largely driven by hospitalization expenses [5, 6], with drug costs representing a significant portion. Furthermore, hospitalization expenses for GC patients have increased in recent years [5]. Therefore, it is crucial to monitor trends in hospitalization expenses and identify the influencing factors for GC patients to prevent unjustified cost escalations.

The rapid rise in healthcare expenditures is a widespread challenge faced by most countries and is particularly severe in the low- and middle-income countries [7]. In response, China introduced a series of healthcare system reforms [8]. A key component of these reforms was the implementation of the Universal Zero Markup Drug Policy (UZMDP) in 2016, which mandated that healthcare institutions sell drugs at their procurement price without any markup, with the goal of reducing drug price and curbing the escalating healthcare expenditure [9, 10]. Existing studies have shown that the implementation of the UZMDP reduced patients’ drug expenses in the short term [11]. However, increases in other medical expenses have undermined the effectiveness of UZMDP in controlling overall medical costs [12, 13]. Given the impact of the UZMDP on healthcare spending, it is necessary to examine the trends and changes in hospitalization expenses for GC patients both before and after the policy’s implementation.

In Shanghai, the UZMDP was gradually phased in between 2015 and 2017. Drug price markups were initially reduced from 15 to 10% in 2015, further lowered to 5% in 2016, and completely eliminated in 2017 [14]. GC, one of the most prevalent cancers in Shanghai [15]. Despite its high prevalence, limited investigation has focused on the trends in patient hospitalization expenses of GC patients and the impact of UZMDP in this region. Therefore, this study aims to explore the trends in hospitalization expenses for GC patients before and after the implementation of the UZMDP and to identify the factors influencing these costs in Shanghai from 2014 to 2021.

Methods

Data source

The study data were obtained from the Health Network of the Shanghai Economic Information Center, which includes the Health Data Reporting System, Hospital Information System (HIS), and Laboratory Information Management System (LIS). Patient case data were extracted from the Health Data Reporting System, while information regarding examination and testing expenses was sourced from the LIS. Additional relevant data were retrieved from the HIS.

Study participants

The study enrolled individuals diagnosed with GC in Shanghai from 2014 to 2021. The collected data primarily include: 1) Patient case information: gender, age, length of stay (LOS), department visited, hospital level, insurance type, hospital nature, etc.; 2) Hospitalization cost information: total hospitalization cost, laboratory fees, bed fees, nursing fees, surgical fees, etc. Outliers were excluded based on the following criteria: 1) Obvious logical discrepancies between detailed and the total hospitalization cost; 2) Hospitalization days less than 1 day or greater than 100 days; 3) Cases with missing basic information; 4) Outlier cases with abnormal total hospitalization costs. Individuals with missing data and outliers (as specified in the "Measures of independent variables" section) were excluded from the analysis. Following a rigorous screening process to eliminate missing values, outliers, and duplicate entries, a total of 23,335 study subjects were included in the final analysis.

Measures of independent variables

The main independent variables included in this study were sex (male; female), age (0–45 years; 45–60 years; > 60 years), LOS (0–7 days; 8–15 days; 16–30 days; > 30 days), surgery (yes/no), medical insurance (yes/no), and the level of the hospital (unclassified hospital; community healthcare center; primary hospital; secondary hospital; tertiary hospital).

Measures of dependent variables

Hospitalization expenses were the outcome variable of interest for the study, which include the costs associated with drugs, medical consumables, examinations (testing, radiology, and laboratory testing), medical services (beds, consultation, surgery, blood transfusion, oxygen, nursing) and other medical costs.

Statistical analysis

The study initially utilized frequencies and percentages to outline the basic characteristics of the study subjects.

Subsequently, interrupted time series analysis (ITSA) was employed to analyze hospitalization expense trends before and after the UZMDP implementation. ITSA recognized for its ability to discern instantaneous and trend changes in interventions, facilitated assessing intervention effects by examining data at multiple time points. In this study, the full enactment of the UZMDP in Shanghai in January 2017 served as the intervention point (Inter-UZMDP). The period before January 2017 was designated as the Pre-UZMDP phase, and the period after January 2017 was referred to as the Post-UZMDP phase. The specific model is as follows:

$${Y}_{t}= {\beta }_{0}+ {\beta }_{1} Time+{\beta }_{2} Intervention+{\beta }_{3 }Posttime+ {\varepsilon }_{t}$$

${Y}_{t}$ represents the outcome variable. Time is the time-counting variable, starting from the first observation point up to the last observation point; Intervention is the intervention variable, with values of 0 before the intervention and 1 after the intervention; Posttime is the post-intervention time-counting variable. The observation points before the intervention have Posttime = 0, and for each subsequent observation point after the intervention, the values of Posttime increase sequentially (e.g., 0 for the first post-intervention point, 1 for the second post-intervention point, and so on);${\varepsilon }_{t}$ represents the residual at time t, indicating the variation unexplained by the regression model.

By fitting the regression model, the parameters ${\beta }_{0}$, ${\beta }_{1}$, ${\beta }_{2}$ and ${\beta }_{3}$ are estimated. ${\beta }_{0}$ is the intercept, representing the initial value of the outcome variable;${\beta }_{1}$ is the slope before the intervention; ${\beta }_{2}$ represents the immediate change in the level of the outcome, calculated as the difference between the indicator level at the end of the pre-intervention period and the beginning of the post-intervention period; ${\beta }_{3}$ represents the change in slope, calculated as the difference between the slopes before and after the intervention.

A new Grey Relational Analysis model (GRA) was used to assess the strength of the relationship between healthcare expenditures and hospitalization expenses for GC patients. This new GRA method calculated correlation coefficients and the degree of association by directly assessing the absolute difference between individual data series and reference series. Total hospitalization expenses served as the reference series, while various healthcare costs were treated as the comparison series. The calculation formula is as follows:

• (1) Determine the reference series and comparison series.

• (2) Calculate the difference series, $\Delta y(i)$, between the reference and comparison series.

  $$\triangle y\left(i\right)=\left|{\mathrm X}_0\left(\mathrm i\right)-{\mathrm X}_{\mathrm y}\left(\mathrm i\right)\right|,\mathrm{where}\;\mathrm y\;=\;1,\;2,\;3,\;4;i=1,\;2,\;3,\cdots,\;8.$$

  The difference series is the absolute difference between the reference series and the comparison series.

• (3) Calculate the correlation coefficient:

$$\varepsilon_y\left(i\right)=\frac{min\triangle y\left(i\right)+\rho max\triangle y\left(i\right)}{\triangle y\left(i\right)+\rho max\triangle y\left(i\right)},\;\mathrm{where}\;\mathrm y\;=\;1,2,3,4;\;\mathrm i\;=\;1,\;2,3,\cdots,\;8.$$

• (4) Determine the degree of association. The magnitude of the degree of association directly reflects the relative influence of each comparison series on the reference series. The greater the degree of association, the stronger the influence of the comparison series on the reference series. The specific calculation is as follows:

  $$\gamma_y=\frac1N{\textstyle\sum_{i=1}^n}\varepsilon_y\left(i\right),\;\mathrm{where}\;i\;=\;1,2,3,\cdots,\;8;N\;=\;8$$

The factors influencing on hospitalization expenses were identified through multivariable linear regression analysis. Variables that were not statistically significant were sequentially removed using backward regression to construct the final model.

Microsoft Office Excel 2016 was used for database establishment and data cleaning, while R version 4.3.1 software performed the analyses. A two-sided P-value less than 0.05 was considered statistically significant.

Results

A total of 16,109 male and 7,226 female patients with GC in Shanghai from 2014 to 2021 were included in this study. The findings indicated that over 68% of GC patients were aged 60 or older, with nearly 50% having a hospital stay for 8 to 15 days. Additionally, more than half of the patients had undergone surgery and were covered by health insurance, and the majority received treatment at tertiary care hospitals. Detailed participant characteristics are provided in Table 1, and year-specific data can be found in Supplement Table 1.

Table 1 Baseline characteristics of the study participants

As shown in Fig. 1, overall hospitalization expenses for GC patients decreased initially from 2014 to 2021, but began to rise steadily after 2017. The implementation of the UZMDP led to a marked 9.03% reduction in total drug expenses (Table 2). Specific details are available in Supplement Table 2.

Fig. 1

Trend of the hospitalization and different medical expenses

Table 2 Amount and proportion of average hospitalization expenses for GC patients at different stages

Table 3 displays the segmented regression outcomes from ITSA. The immediate shift in total hospitalization expenses after the implementation of the UZMDP was not statistically significant. However, average total hospitalization expenses exhibited a significant upward trend over time (P < 0.01), manifesting a monthly increase of approximately 271.73 CNY, as shown in Fig. 2.

Table 3 The impact of the universal zero mark-up drug policy on changes in medical costs for GC patients

Fig. 2

Trend of the hospitalization expenses from 2014 to 2021

In contrast, drug expenses underwent an immediate and marked reduction following the UZMDP, though they gradually increased thereafter. Expenses for consumables did not show a significant immediate change, but increased by about 106.41 CNY per month post-UZMDP. Meanwhile, both examination and healthcare services expenses showed a consistent rise post-UZMDP, as depicted in Fig. 3.

Fig. 3

Trend of the different medical expenses from 2014 to 2021

Table 4 presents the outcomes of the new GRA, illustrating the degree of association and relative ranking of average expense levels for each category. The study underscores that drug and consumable costs are primary factors significantly influencing hospitalization expenses for GC patients in Shanghai from 2014 to 2021. Their degree of association is notably higher than that of other variables, with healthcare services expenses and examination following, but to a lesser degree.

Table 4 The results of the new grey relational analysis

After verification, total hospitalization expenses demonstrated a normal distribution. Univariate analysis revealed significant differences in expenses based on gender, age, length of stay, surgery, health insurance, and hospital level (Supplement Table 3). All variables were included in a multivariable linear regression model, which was refined using a backward elimination process, as detailed in Table 5. Findings indicated a positive correlation between incurred expenses and LOS. Additionally, GC patients who underwent surgery incurred higher costs than those who did not. Hospitalizations at tertiary and secondary care hospitals were associated with higher expenses compared to those at community health service centers and primary care hospitals. Furthermore, hospitalization costs were lower for those aged ≤ 60 compared to those aged > 60.

Table 5 Multivariable analyses of hospitalization expenses for gastric cancer patients in Shanghai from 2014 to 2021

Discussion

This study retrospectively analyzed the trends and influencing factors of hospitalization expenses for GC patients in Shanghai, using ITSA and the new GRA. The findings revealed a rising trend in hospitalization expenses for GC patients from 2014 to 2021, with the increase primarily driven by the growing costs of drugs and consumables both before and after the implementation of UZMDP. Additionally, demographic characteristics, disease severity, and hospital-level factors collectively influence hospitalization expenses for GC patients.

Trends in hospitalization expenses and the impact of the UZMDP

Our study indicated that a general upward trend in hospitalization expenses before and after the implementation of the UZMDP. The varying growth rates of hospitalization expense during the pre- and post- UZMDP periods were linked to changes in the components of these expenses. Specifically, drug expenses, which were the primary contributor to hospitalization costs, decreased before the UZMDP, potentially contributing to the slower overall growth in hospitalization expenses during that period.

Meanwhile, expenses for medical consumables, healthcare services, and diagnostic tests continued to rise, ultimately contributing to the overall increase in hospitalization expenses, consistent with previous studies [12, 13]. This phenomenon reflects the complex influence of provider-induced demand (PID), which occurs when healthcare providers influence patients’ demand for care in ways that may not align with their best interests [14]. The implementation of the UZMDP did not reduce hospitalization expenses, as it may have reduced drug expenses driven by PID but instead shifted its effects to other revenue sources, such as consumables and examinations, to compensate for the revenue loss [14]. This suggests that reducing hospitalization expenses requires not only adjusting drug prices but also monitoring the prices of other healthcare services, such as consumables and examinations, while establishing effective incentive mechanisms for healthcare providers.

Trends in components of hospitalization expenses and the impact of the UZMDP

Our study indicated that drug expenses were a key component contributing to the increasing trends in hospitalization expenses from 2014 to 2021. Although drug expenses initially decreased in Shanghai, the ITSA revealed a steady increase over time after the full implementation of the UZMDP. On the one hand, this may be attributed to a decrease in drug prices, which could lead patients to purchase larger quantities or opt for more expensive drugs within a fixed budget [16]. Additionally, the increase in the annual number of patients could also explain this trend [14]. On the other hand, healthcare providers remain incentivized to adjust their strategies to maximize profits, which could potentially result in increased drug expenses after the implementation of the UZMDP. For example, doctors may still be incentivized to overprescribe drugs or prescribe more expensive drugs to patients in exchange for rebates and higher profits [16]. In the future, there is still a need to further regulate prescribing practices so as to reduce the burden of medication on patients.

Since the full implementation of the UZMDP in 2017, other healthcare expenses in hospitals have continued to rise due to insufficient subsidies and cost-effectiveness motives [17]. Our study showed that the impact of consumables expenses on hospitalization expenses for GC patients is almost equivalent to that of drug expenses. The trend in consumable expenses increased both before and after the implementation of the UZMDP. This may be due to the fact that consumables account for a relatively large portion of hospitalization expenses for GC patients [18]. After the implementation of the UZMDP, healthcare providers compensated for the reduction in drug revenues by increasing the provision of other services or products, leading to increases in consumable expenses [19]. To reduce consumables expenses, China implemented the universal consumables markup policy in 2019 [20]. However, this study found that only a 2.7% decrease in consumables expenses for GC patients in 2020 compared to 2019, followed by a continued upward trend thereafter. The main reason may be that there were more elderly GC patients who used more high-value consumables in this study, which made the change in consumables expenses less noticeable [18]. Further research is needed to focus on establishing and improving the medical consumables management system based on patient’s characteristics, as well as regulating the consumables use behavior of medical personnel to reduce patient’s consumables expenditure [18].

It is also noteworthy that the study identified an upward trend in healthcare services expenses from 2014 to 2021. This trend may be linked to China’s public hospital reforms launched in 2015, which aimed to increase remuneration for professional services provided by healthcare workers, improve performance-based salaries for medical staff, and establish a contemporary hospital management system [17]. These reforms sought to reduce instances of PID by aligning incentives with professional services, rather than excessive treatments or prescriptions. However, the continued increase in drug, consumables, and examination expenses suggests that further research and policy efforts are necessary to refine pricing strategies for medical services and reduce PID incentives driven by unreasonable salaries.

Factors influencing hospitalization expenses

Regarding the demographic characteristics of patients, this study found that elderly GC patients incurred higher total hospitalization expenses, consistent with previous studies. On one hand, this is due to the higher incidence of GC among the elderly [21]. Elderly GC patients undergo complex treatments, leading to increased use of high-value consumables and adjuvant items, thereby increasing their hospitalization expenses [18]. Through univariate analysis, our study found that male GC patients incurred higher treatment costs than female ones. In multiple regression analysis, however, we found that male GC patients would have fewer costs, which is inconsistent with previous studies [22]. The inconsistency between results may be due to the unequal distribution of key factors, such as disease severity and LOS, between male and female GC patients. After adjusting for these variables in the multivariable model, the independent effect of gender reversed. This suggests that further research should include subgroup analyses and carefully select appropriate covariates to better understand these relationships [23, 24].

Regarding illness severity, our study indicated that patients with more severe conditions have higher hospitalization expenses than those with milder illnesses. Specifically, patients diagnosed with GC who undergo prolonged hospital stays and surgical interventions tend to accumulate elevated hospitalization expenses due to increased healthcare resource utilization, resulting in higher medical costs [22]. However, it has been found that inpatients have unnecessary hospitalization days, that is, patients need to stay in the hospital for observation even if they do not receive any treatment, which leads to an increase in their costs [25]. This not only increases the economic burden on patients but also leads to the waste of healthcare resources, suggesting the need to improve the efficiency of technical services while ensuring the quality of healthcare services in the future [25].

In terms of hospital characteristics, in line with prior research, our study indicates that patients incur the highest hospitalization expenses when receiving care at tertiary hospitals. This association is likely due to the higher concentration of patients with severe gastric cancer in tertiary care settings [16].

Implications for policy reforms and implementation

Our study suggests that a comprehensive policy framework is needed to collectively regulate the irrational growth of medical expenses and establish a fair remuneration system for medical services, aiming to reduce PID driven by unreasonable salaries in future policy reforms. Specifically, policymakers should focus on implementing performance-based payment mechanisms that align doctors’ incentives with quality care, rather than the quantity of services. This could involve balancing remuneration between medical services and consumables, ensuring that doctors are adequately compensated for delivering high-value care. Secondly, transparent procurement systems for consumables and standardized pricing models for examinations could help control excessive costs, while maintaining resource efficiency. Additionally, considering the characteristics of GC patients, it is essential to enhance awareness campaigns targeting older adults to improve early diagnosis rates.

Strengths and limitations

Our study is the first to investigate trends in hospitalization expenses and the impact of the UZMDP on GC patients in Shanghai. It provides valuable insights into the rational control of hospitalization expenses for GC patients and the advancement of healthcare system reforms.

The study has several limitations. First, the unavailability of other variables, such as patients’ income and history of other diseases, introduces potential confounding effects on the results. For instance, patients with higher incomes often have greater access to advanced medical treatments and high-quality healthcare services, which may lead to higher hospitalization expenses. These increased expenses are not solely attributable to drug, consumable, and examination costs, but could also be driven by the selection of more intensive or specialized care options. Addressing these confounding variables in future research, including subgroup analyses, would enhance the accuracy and reliability of the findings. Second, the absence of specific GC classifications hinder a detailed analysis of the influencing factors of hospitalization expenses among patients with diverse GC classifications. Without categorizing patients based on these subtypes, it becomes challenging to determine which factors are driving hospitalization expenses for specific groups. Future studies should incorporate specific GC classifications, such as early-stage and advanced-stage, into their analysis. Third, the intricate policy environment, marked by the concurrent implementation of multiple healthcare reform measures, complicates the measurement of the net effect of the UZMDP. The results of the ITSA and the new GRA may be influenced by various policies [12], potentially distorting the observed outcomes. Future studies could conduct sensitivity analyses to isolate the specific impact of the UZMDP. Additionally, since ITSA assumes linear relationships and clear interruption points, it may not fully capture non-linear effects, while GRA is sensitive to data scaling and does not establish causality, which could lead to biased or incomplete conclusions. To address this limitation, future research could use a more comprehensive model to explore the trends before and after the implementation of the UZMDP. Moreover, we did not account for the potential impact of inflation on the observed increase in hospitalization expenses over the years. Future studies could adjust for inflation to better isolate the effects of policy reforms and provide a more accurate representation of changes in hospitalization expenses. Furthermore, it is important to consider the potential impact of a large-scale outbreak of COVID −19 in 2020 on the hospitalization expenses of GC patients, as it could have contributed to an increase in medical across various categories.

Conclusions

Our study indicates that hospitalization expenses for GC patients in Shanghai have shown an upward trend both before and after the implementation of the UZMDP. Drug expenses, which were a primary components of total hospitalization expenses, initially decreased, but then steadily increased over time after the full implementation of the UZMDP. Meanwhile, expenses of consumables, healthcare services, and examination continued to rise post-UZMDP. Furthermore, the study found that older age, longer hospitalization stay, surgery, and receiving medical care at secondary and tertiary hospitals were key factors contributing to higher hospitalization expenses for GC patients. These findings provide a basis for more effective management of hospitalization expenses for GC patients.