Drug price trends, regional disparities, and supply shortages under China’s National Volume-Based Procurement: a nationwide analysis (2018–2024)

Abstract

Background

The National Volume-Based Procurement (NVBP) policy was implemented in China in 2018 to address escalating pharmaceutical expenditures. This study aimed to describe long-term price trends, regional variations in bid-winning drug prices, and supply stability during the 2018–2024 implementation period.

Methods

Data were obtained from the Shanghai Sunshine Pharmaceutical Procurement Platform and the Comprehensive Service Platform for NVBP. Price trends of bid-winning and alternative drugs were analyzed using monthly procurement data for Batches 1–7, excluding the insulin-specific sixth batch.Laspeyres, Paasche, and Fisher price indices were calculated using January 2018 as a common index anchor to standardize long-term price comparisons, rather than as a batch-specific pre-policy baseline. Regional disparities in bid-winning drug prices were examined using official bid-winning results for Batches 1–5 and Batches 7–10, with purchasing power parity indices estimated by the national product dummy method. Supply shortages were measured as monthly regional order–delivery gaps during the first post-implementation year of each included batch and classified into five severity levels.

Results

Bid-winning drugs showed substantial price declines after implementation, with Laspeyres, Paasche, and Fisher indices decreasing by approximately 55%-85%, 60%–85%, and 50%-80%, respectively. Prices remained stable at low levels without obvious rebound. Alternative drugs showed smaller and more fluctuating declines of approximately 5%-40%, 5%-40%, and 10%-40%, respectively. Supplementary interrupted time-series analysis of the Fisher index showed significant negative immediate level changes for bid-winning drugs across all included batches, whereas alternative drugs showed smaller and less consistent immediate changes. Regional PPP point estimates were generally lower in economically developed regions than in the western region, although many individual batch-region comparisons were not statistically significant. First-year supply shortages ranged from approximately 17% to 33% across batches. Lower-priced drugs were more frequently observed in higher shortage-severity categories, with drugs priced below 0.5 CNY accounting for 77.6% of severe shortages.

Conclusion

This nationwide descriptive study documented sharp and sustained price declines for bid-winning drugs, smaller and heterogeneous declines for alternative drugs, persistent regional price disparities, and supply shortages under China’s NVBP. These findings suggest a policy tension between price reduction, regional equity, and supply security.If these descriptive associations reflect underlying procurement mechanisms, future policy refinements may consider regional price-differential monitoring, dynamic price-adjustment mechanisms for extremely low-priced drugs, supply guarantee requirements, and reward-penalty mechanisms for procurement participants.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12913-026-15065-1.

Introduction

Numerous countries worldwide face the challenge of rising pharmaceutical expenditures [1, 2], with the global pharmaceutical market forecasted to reach a valuation of USD 35289.0 billion by 2035 [3]. In particular, China’s pharmaceutical market is forecast to reach USD 7996.3 billion by 2035, reflecting a 11.13% compound annual growth rate from 2025 to 2035 [4]. To curb escalating drug costs, the Chinese government implemented nationwide pharmaceutical tendering in 2009, achieving full provincial-level bidding coverage by 2010 [5]. However, as of approximately 2016,China’s pharmaceutical expenditure remained substantial, accounting for 40% of total health spending [6], compared with less than 15% in Fiji, Australia, and Singapore [7]. Even in the United States, prescription drug spending accounted for less than 10% of total health spending from 2017 to 2023 [8]. ‌This unsatisfactory outcome was largely driven by two deep-rooted problems in the traditional bidding system: the widespread “separation between bidding and procurement” and the persistent “decoupling of volume and price” [9].

To overcome this dilemma, the Chinese government introduced the National Volume-Based Procurement (NVBP) policy in 2018, adopting a “volume-for-price” mechanism to rationalize drug prices [9]. Public hospitals account for more than 70% of national drug sales and represent the primary distribution channel in China [10]. As most drugs are used in public hospitals, the policy consolidates hospital drug demand and awards supply contracts to winning bidders to achieve substantial price reductions. The implementation includes five stages: drug listing, hospital volume reporting, enterprise bidding, winner selection, and product distribution [11]. Drugs supplied by winning enterprises are hereafter referred to as bid-winning drugs, which account for the majority of volume in public medical institutions within each province.Bid-winning drug prices may differ across provinces [9, 12]. In this study, clinically interchangeable non-winning drugs are defined as alternative drugs.This policy represents China’s first nationwide volume-based procurement initiative to control drug expenditure growth through market mechanisms [13].

Previous studies have revealed short-term changes in drug prices following the implementation of China’s NVBP. In the pilots, 25 drugs were successfully bid with price reductions ranging from 25% to 96% [14]. Lan et al. (2022) reported a significant reduction in average monthly outpatient and emergency drug expenditures after NVBP implementation, which was driven by lower per-patient per-day drug costs rather than reductions in non-drug medical service costs [15]. Yang et al. examined antihypertensive drugs in Shenzhen observing significant price decreases for NVBP drugs and reduced total costs [16]. Zhao et al. analyzed NVBP drugs in Tianjin, identifying a 61.55% average price reduction for bid-winning drugs during first procurement batch with substantial cost savings [17]. Globally, centralized procurement has effectively reduced drug costs. In Jordan, joint procurement processes achieved estimated savings of 2.4%, which increased to 8.9% after the exclusion of one project [18]. In Greece, hospital-based drug bidding notably decreased medical expenses [19]. In the United Kingdom, centralized competitive procurement for generic medicines has significantly lowered prices and improved National Health Service (NHS) funding efficiency [20]. In Brazil, centralized joint pharmaceutical procurement has driven substantial price reductions, with approximately 47% of drugs achieving price decreases of more than 10% and 32% falling by over 20% [21]. Research on China’s insulin market estimated first-year total drug expenditure savings of USD 2.85 billion, greatly improving the affordability and accessibility of insulin [22]. Data released by the National Healthcare Security Administration indicate that by February 2022, six batches of centralized drug procurement had been conducted in China, saving the nation over 260 billion CNY in drug costs [23]. In addition to price reductions, centralized procurement may also generate supply-side risks [24]. When large market demand is concentrated among a limited number of winning suppliers, manufacturers may face increased pressure to ensure continuous production and distribution. Researchers noted that when vast market demand is concentrated among only a few suppliers, drug shortages may ensue. In India, the purchasing prices in the public sector range from 0.27 to 0.48 times the international reference price, yet these drugs are in short supply, with median availability in the public sector between 0% and 30% [24]. Due to factors such as insufficient raw materials, the Netherlands has also encountered drug shortages in centralized procurement [25]. In China, most bid-winning drugs face shortages during procurement cycles, leading to patients being unable to access necessary medications in a timely manner, which in turn results in discontinuities in clinical treatment, particularly in remote areas [26].

In conclusion, although research indicates that China’s NVBP policy reduces drug prices, most existing studies have been limited to analyses based on data from a single hospital, region, or therapeutic field [17, 27–29]. Current research lacks nationwide analyses of procurement prices for drugs in public hospitals and long-term trend studies on NVBP-related medicines. Moreover, considering China’s regional economic disparities, the fairness and rationality of the regional distribution of NVBP drug prices have not been adequately explored. Meanwhile, the majority of investigations have centered on changes in drug consumption, pricing, and market concentration under the NVBP [14, 22, 30, 31], while detailed studies of the supply status of bid-winning drugs are sparse.

Conceptually, the volume-for-price mechanism may affect prices, regional price distribution, and supply stability through the same procurement logic. By aggregating hospital demand and linking contracted volume to bid prices, NVBP can strengthen purchasers’ bargaining power and produce sharp price reductions. However, when large volumes are concentrated among a limited number of winning suppliers, very low prices may compress profit margins and reduce supply flexibility. In addition, province-selection rules and differences in regional demand, logistics costs, and purchasing capacity may allow a nominally national procurement system to generate uneven regional price outcomes.

This study sets out three research objectives. First, examine the long-term trends in drug prices under the NVBP, thereby providing new evidence for the global implementation of volume-based pharmaceutical procurement. Second, generate evidence regarding whether the distribution of bid-winning drug prices under the NVBP is fair and reasonable, presenting valuable lessons for other countries with uneven economic development adopting similar policies. Third, analyze supply shortages in volume-based procurement drugs to provide a reference for ensuring stable drug supply in other countries implementing comparable policies.

Methods

Data source

Data were derived from two sources.The first was the results of the first ten batches of NVBP up to December 2024, obtained from the Shanghai Sunshine Pharmaceutical Procurement Platform. We excluded the sixth batch (insulin-specialized procurement) because it covered all insulin products available on the market and showed no regional price differences. The data contain fields such as generic drug name, packaging specifications, pharmaceutical enterprise, supply regions, and winning bid price.These data were used to analyze regional variations in bid-winning drug prices.

The second source was actual drug procurement data from public medical institutions in China, obtained from the Comprehensive Service Platform for National Volume-Based Procurement, which is affiliated with the National Healthcare Security Administration of China (NHSA) and covers drug procurement activities of public medical institutions in all provinces of China. Owing to access restrictions on this platform, only procurement data corresponding to the first seven batches of NVBP were available for this study. Data cover public medical institutions in mainland China from January 2018 to December 2024, containing original fields such as drug name, selection status, procurement date, region name, manufacturer, formulation specifications, dosage form, unit price, and order quantity. The first seven batches (excluding the sixth batch) involved a total of 280 types of drugs (with respective counts of 25, 32, 55, 45, 62, and 61 in each batch). (See Appendix 1).

The research team received structured analytical datasets after platform-level preprocessing. Records with missing key identifiers, missing prices, missing procurement quantities, or impossible values were not included in the final analytical datasets.Drug records from the two data sources were harmonized at the generic-name, manufacturer, dosage-form, strength, and package-specification levels. Generic names were first standardized according to NVBP official bid-winning lists. Specifications and package units were then converted to the smallest comparable unit where necessary. Records with inconsistent names or specifications were reviewed manually against the official procurement documents before inclusion in the corresponding analytical dataset.After preprocessing and harmonization, the original data were organized into three analytical datasets corresponding to the price trend, regional price disparity, and supply shortage analyses, as shown in Appendix Figure S1.

This manuscript does not involve data collected from human subjects, therefore requiring no ethical approval.

Analysis of price trends

Taking the first seven batches of NVBP drugs as research subjects, this study adopted the Laspeyres, Paasche, and Fisher price indices to analyze drug price changes after NVBP implementation.These three complementary indices were used together to improve the robustness of results.The Laspeyres index reflects price changes using fixed base-period quantities.The Paasche index accounts for changes in consumption quantities between the base and reporting periods [32]. The Fisher index, as the geometric mean of the previous two, is widely recognized as the “ideal” index for reducing bias and improving reliability [33]. Using all three indices enhances the validity of results and avoids over-reliance on a single method. The formulas are as follows:

Note

denotes Laspeyres price index, denotes Paasche price index, and denotes Fisher price index. represents the base-period price of a drug, represents the current-period price of a drug, represents the base-period procurement quantity, and represents the current-period procurement quantity.

This study characterizes the drug prices acquired in the first seven batches of NVBP from the perspectives of both bid-winning drugs and alternative drugs. All price index calculations are based on the drug procurement volume of January 2018 as the base quantity and the prices of January 2018 as the price benchmark.

To further examine whether price changes coincided with the implementation timing of each procurement batch, we conducted a supplementary single-group interrupted time-series analysis using monthly price-index data. The included batches were Batches 1–5 and Batch 7, with implementation dates of December 2019, April 2020, November 2020, April 2021, October 2021, and November 2022, respectively. The segmented regression model was specified as follows:

where represents the monthly price index at time t; time is a continuous variable indicating the number of months from the beginning of the observation period; intervention is a binary indicator coded as 0 before implementation and 1 after implementation of the corresponding procurement batch; and posttime is a continuous variable indicating the number of months after implementation, coded as 0 before the intervention. In this model, estimates the baseline monthly trend before implementation, estimates the immediate level change at the implementation point, and estimates the change in trend after implementation compared with the pre-implementation trend. Newey–West standard errors were used to account for potential autocorrelation in monthly time-series data.

Analysis of regional distribution of prices

Pharmaceutical products from the first ten procurement batches were selected for analysis. Regional purchasing power parity (PPP) indices were computed using the national product dummy method. The concept of PPP originates from the “Law of One Price”, referring to the currency ratio required to purchase identical products between two regions, essentially representing regional price indices [34]. The indices obtained in this study essentially reflect regional PPP, indicating the price levels of the bid-winning drugs and the purchasing power of CNY across the economic regions of China.

The calculation formula for the national product virtual method is as follows [35]:

Note

denotes the arithmetic mean price of drug i in region j. denotes regional dummy variable, taking the value 1 if the price observation is from region j and 0 otherwise. denotes drug dummy variable, taking value 1 for observations of drug i and 0 otherwise. refers to independently distributed random disturbance term.

The above equation can be simplified as follows:

Note

denotes the purchasing power level of region j, indicates the number of monetary units in region j that possess the same purchasing power as one unit of currency in the reference region, that is the purchasing power parity index.

Based on the economic regional classifications of China released by the National Bureau of Statistics, a standardized and widely used framework in academic and policy research, this study quantitatively computes PPP indices for NVBP drugs across four regions (western, central, northeastern, and eastern) to compare regional price disparities among bid-winning drugs. The eastern region is the most developed with the highest economic level; the northeastern and central regions come next; the western region is underdeveloped with the lowest economic level. The correspondence between economic regions and provinces is provided in the Appendix 2. In addition, to control for price variations caused by different packaging specifications among drugs sharing the same generic name, all drug prices were standardized to the smallest available packaging unit.(See Appendix 3).

Analysis of drug supply shortage

Drug supply shortages were quantified by the discrepancy between the ‘ordered quantity (tablets)’ and the ‘delivered quantity (tablets)’, using monthly procurement data from the first seven NVBP batches during the first post-implementation year across provincial units. Ordered quantity refers to monthly platform orders by medical institutions, and delivered quantity represents actual received shipments. When delivered quantity exceeded ordered quantity within a province–drug–month record, the shortage rate was set to zero for shortage classification, because such records indicated no unmet order demand.

The shortage rate was calculated as the percentage gap between ordered and delivered quantities relative to ordered quantity:

Shortage rate = (ordered quantity−delivered quantity) / ordered quantity × 100%.

Ordered quantity was used as a proxy for procurement demand. Because monthly orders may reflect stockpiling, repeated ordering, or timing differences between ordering and delivery, this metric should be interpreted as a procurement-platform supply fulfillment gap rather than a direct measure of unmet clinical demand.

Because no standardized classification currently exists for procurement-based drug shortage severity, we used a descriptive five-level classification. Observations with a shortage rate of 0% were classified as “No shortage.” Positive shortage rates were divided into four equal 25-percentage-point intervals: “Mild shortage” (0%–25%, excluding 0%), “Moderate shortage” (25%–50%, excluding 25%), “Shortage” (50%–75%, excluding 50%), and “Severe shortage” (75%–100%, excluding 75%),to reflect the relative severity of supply shortfalls compared with total demand.The correlation between drug prices and the supply shortage status of bid-winning drugs was analyzed using Spearman rank correlation, a standard method for ordinal categorical relationships.

Data analysis was performed using Structured Query Language (SQL) for data retrieval and Python for statistical analysis and figure visualization. The main Python libraries employed included pandas for data processing and statsmodels for statistical analysis.

Results

Price trends of the drugs

The price trends of bid-winning drugs were largely consistent across the Laspeyres, Paasche, and Fisher indices (Fig. 1). After implementation, bid-winning drugs in all included batches showed a marked “cliff-like” decline, with the Laspeyres, Paasche, and Fisher indices decreasing by approximately 55%–85%, 60%–85%, and 50%–80%, respectively, within one year before and after implementation. After the sharp decline, the price indices did not show an obvious rebound, indicating that bid-winning drug prices remained stable at low levels over the longer observation period. In contrast, alternative drugs showed smaller and more fluctuating declines, with the corresponding indices decreasing by approximately 5%–40%, 5%–40%, and 10%–40%, respectively (Fig. 2).The range of reductions across batches reflects batch-specific product composition, not a systematic decline in NVBP effectiveness over time.

Fig. 1.

Price indices for bid-winning drugs (2018–2024)

Fig. 2.

Price indices for alternative drugs (2018–2024)

To further assess the timing of price changes, a supplementary single-group interrupted time-series analysis was conducted using the Fisher index(Table 1). Bid-winning drugs showed significant negative immediate level changes across all included batches, consistent with the observed cliff-like decline and subsequent low-price stability. Alternative drugs showed smaller and less consistent immediate changes, suggesting a more gradual and heterogeneous price response.

Table 1.

ITS of fisher price indices by drug type and procurement batch

Drug Type	Procurement Batch	Immediate level change	Change in post-implementation trend
Bid-winning Drugs	Batch 1	-0.2417***	0.0214***
	Batch 2	-0.7900***	0.0012
	Batch 3	-0.6550***	0.0018*
	Batch 4	-0.5665***	-0.0010
	Batch 5	-0.5231***	0.0029***
	Batch 7	-0.4873***	0.0050***
Alternative Drugs	Batch 1	-0.0571	0.0075***
	Batch 2	-0.1434***	-0.0059***
	Batch 3	-0.0543*	-0.0016
	Batch 4	-0.0272	-0.0085***
	Batch 5	-0.0583**	0.0053***
	Batch 7	-0.0132	0.0003

Note: *p < 0.05, **p < 0.01, ***p < 0.001

Regional distribution of selected drug prices

Table 2 presents the regional PPP indices for bid-winning drugs by procurement batch, using the western region as the reference. PPP point estimates were generally below 1.000 in the eastern, central, and northeastern regions relative to the western region, suggesting a directionally lower procurement-price pattern in these regions. However, most individual batch–region coefficients did not reach conventional statistical significance, and statistically significant differences were concentrated in Batch 5 and selected later batches. Therefore, the regional pricing pattern should be interpreted as suggestive rather than statistically definitive for every batch. The first batch may have reflected pilot-stage market dynamics and a smaller set of participating products and enterprises, and therefore its regional pattern may not be directly comparable with subsequent batches.

Table 2.

Regional PPP indices of bid-winning drug prices by procurement batch

Procurement batch	Western	Central	Northeastern	Eastern
Batch 1	1.000 (Reference)	0.978	0.965	0.996
Batch 2		0.978	0.988	0.965*
Batch 3		1.004	0.998	0.971
Batch 4		0.984	0.972	0.975
Batch 5		0.953**	0.946***	0.952**
Batch 7		0.983	0.994	0.963*
Batch 8		0.988	0.965	0.953***
Batch 9		0.986	0.969	0.932***
Batch 10		1.009	0.962**	0.967***

Note: *p < 0.1, **p < 0.05, ***p < 0.01. Full regression output is provided in Appendix 5

Shortages of drug supplies

In the first year of each procurement batch, the supply consistently fails to meet the order demand, with a supply shortage of approximately 17%–33%. (See Appendix 4). Figure 3 illustrates the supply status of drugs across different procurement batches, where the non-shortage cases increase starting from the third batch, while the shortage/severe shortage proportions decrease from the fourth batch onward. For example, severe shortage rates in the included batches were 4.8%, 5.2%, 4.6%, 1.4%, 0.8%, and 1.4%, respectively, excluding the insulin-specific sixth batch.

Fig. 3.

Drug shortage status for each batch

Figure 4 demonstrates the relationship between drug shortages and prices. Under conditions of no shortage, the distribution of drugs across various price categories is relatively uniform. As the degree of shortage intensifies, the proportion of drugs priced below 0.5 CNY increases significantly, reaching 77.6% in cases of severe shortage. In contrast, the proportions of drugs priced 10–50 CNY and 50 CNY or above continually decrease with increasing shortage levels. Under shortage conditions, drugs priced 10–50 CNY account for 10.6%, and those priced at 50 CNY or above account for 1.1%. In cases of severe shortage, the samples in the 10–50 CNY range constitute 1.2% and those at 50 CNY or above only 0.1% (observed only 5 times).To formally quantify this monotonic relationship, a Spearman rank correlation analysis was performed between price level and shortage severity. The results yielded a statistically significant moderate negative association (r=-0.3345, P < 0.001), indicating a statistically significant negative association between price category and shortage severity.

Fig. 4.

Relationship between shortage level and bid-winning drug price

Discussion

This nationwide descriptive study documented long-term procurement-price trends, regional price disparities, and supply fulfillment gaps under China’s NVBP. Following NVBP implementation, bid-winning drugs showed a steep decline in prices, while alternative drugs showed smaller and more fluctuating price reductions. The low-price pattern was more pronounced and stable for bid-winning drugs than for alternative drugs. Second, bid-winning drug prices showed a paradoxical regional pattern, with lower procurement prices in economically developed regions and higher procurement prices in underdeveloped regions. This pattern suggests a potential regional equity concern. Finally, during the first post-implementation year, each procurement batch showed a procurement-platform supply fulfillment gap of approximately 17%–33%. Lower-priced drugs were more frequently observed in higher shortage-severity categories, although this relationship should be interpreted as an association rather than a causal effect.

Following the implementation of the NVBP policy, bid-winning drugs experienced a marked precipitous price reduction (50%-85%) in the short term, with this low-price effect demonstrating long-term stability. This price reduction is consistent with existing research findings. Long et al. observed an immediate 61.71% price decrease for bid-winning drugs during China’s inaugural centralized procurement, followed by a sustained 23.97% reduction in the second procurement batch [36]. Similarly, Wang et al. observed an 81.35% decline in the short-term Fisher price index of bid-winning drugs [37]. The enduring low-price pattern may be partly explained by economies of scale [38]. By leveraging the substantial procurement volumes from public medical institutions, aggregated procurement volumes may allow manufacturers to reduce unit costs and offer lower bid prices. Secondly, enterprises can plan their production capacity based on clearly defined and long-term stable procurement volumes, avoiding idle capacity or the costs associated with emergency capacity expansions due to demand fluctuations inherent in the traditional decentralized procurement mode. This mechanism may create room for price concessions.

The prices of alternative drugs have shown a fluctuating downward trend, with relatively smaller declines (10%-40%). Chen et al. found that in Shenzhen, the prices across all drug categories dropped, and the reduction in bid-winning drug prices played a positive role in controlling overall pharmaceutical costs [39]. Wang et al. found the drug price index for no bid-winning drugs dropped significantly, with the Fisher price index declining by 11.96% and that of alternatives by 4.34%, and no long-term downward trend was observed [37]. While previous studies have indicated that price adjustments among non-winning drugs are mainly driven by short‑term factors, the present study observed that alternative drug prices showed a sustained downward trend.One possible explanation is that the sharp price reduction of bid-winning drugs may have been associated with competitive pressure. Public hospitals, which account for more than 70% of national pharmaceutical sales, serve as the primary sales channel in China [40], creating incentives for manufacturers of alternative drugs to lower prices in order to capture the remaining market share and avoid obsolescence [41, 42]. However, the sustainability of this trend remains uncertain. Continuous downward price pressure may reduce manufacturers’ profit margins and could potentially affect the availability of alternative drugs, particularly for low-priced generic medicines. Because our data do not include information on drug quality, adverse events, product recalls, production costs, or product withdrawal, we were unable to assess whether the decline in alternative drug prices compromised product quality or long-term availability.

The PPP results suggest a possible inverse regional pricing pattern, but this interpretation requires caution because many individual batch–region comparisons were not statistically significant.The inverse relationship between procurement prices and regional economic development may create potential equity concerns for underdeveloped areas and could generate incentives for cross-regional arbitrage. Typically, regional price levels are expected to demonstrate a positive association with per capita income [43]; however, this positive association fails to manifest within China’s NVBP framework. Shi et al. discovered that certain low-income provinces in China even pay higher procurement prices for identical drugs compared to high-income provinces, indicating the absence of a positive correlation between income levels and prices, which may lead to disproportionately heavier pharmaceutical burdens on populations in impoverished regions [44]. Analysis of medical insurance fund utilization reveals that China’s eastern and central regions, characterized by developed economies and dense insured populations, maintain significantly larger cumulative medical insurance fund surpluses than western and northeastern regions [45]. The limited payment capacity of medical insurance funds in underdeveloped areas further weakens their bargaining power in drug procurement. Against this backdrop, inter-regional price divergence was observed under the NVBP.

The NVBP stipulate that bid-winning enterprises with the lowest prices receive priority in selecting two provinces as supply regions, while other bidders select supply regions sequentially in descending order of their quoted prices. Driven by profit-maximization motives, low-price bid-winning enterprises typically prioritize regions with maximal demand and minimal distribution radius [46]. Economically developed eastern regions, generally featuring superior transportation infrastructure, denser populations, and greater medication demand, are consequently more likely to be preferentially selected by enterprises [47]. Conversely, underdeveloped regions with smaller demands paradoxically tend to be allocated to high-price bid-winning enterprises. This dual imbalance between fund distribution and procurement rules may partly explain the observed inverse regional pricing pattern: developed regions leverage abundant medical insurance funds and scale advantages to secure even lower-priced drugs, while underdeveloped regions may face relatively higher procurement prices.

In the first year of NVBP, the supply shortage of bid-winning drugs was relatively severe. However, in later batches, the proportion of drugs facing shortage and severe shortage decreased after the fourth batch. Drug shortages showed a clear price-related pattern: higher-priced drug categories were less frequently observed in severe shortage categories, whereas lower-priced drugs were more frequently observed in these categories.Analysis of shortage reports from the first five procurement batches reveals increasing shortage frequencies across all batches, with bid-winning enterprises’ shortage report ratios substantially exceeding pre-NVBP levels [48]. Zhao et al. further found that due to the limited profit margins, low-priced drugs may reduce enterprises’ supply incentives, thereby facing a higher risk of shortages [49]. Studies also show heightened shortage susceptibility among low-price drugs [50]. Possible explanations include: first, bid-winning enterprises face the pressure of “volume-price binding"‌, and under a low-margin model lacking in technological upgrades and production capacity flexibility, are often unable to cope with sudden surges in hospital demand that far exceed their reported volumes [51]. Second, the complex layers in grassroots drug distribution and significant regional differences in delivery costs mean that low-priced drugs, due to their minimal marginal returns, may encounter regional “implicit supply disruptions” [52, 53]. In addition, the lack of a robust early-warning and emergency production response mechanism, insufficient penalties for enterprises that unlawfully discontinue drug supply, and inefficiencies in coordinating alternative supplies, may further contribute to this issue [54].

Several policy considerations may be drawn from these descriptive findings. First, policymakers may consider monitoring whether price reductions of bid-winning drugs are accompanied by broader price adjustments among alternative drugs. Second, regional price-differential caps for identical drug specifications could be evaluated to reduce excessive procurement-price disparities across regions. Third, dynamic price-floor or price-adjustment mechanisms may be considered for extremely low-priced drugs, particularly when low prices are accompanied by higher supply fulfillment risks. Finally, a credit evaluation and reward–penalty system for procurement participants may help strengthen supply guarantees and improve procurement performance. These policy options should be further assessed using causal designs and additional data on production costs, manufacturer behavior, drug quality, and patient-level access.

This study has several limitations. First, the analysis was based on administrative procurement data from public medical institutions and did not include retail pharmacy sales. Although public hospitals account for the majority of pharmaceutical sales in China, the exclusion of retail pharmacy data may limit the comprehensiveness of the findings, particularly for drugs with substantial out-of-hospital use. Second, this study was designed as a nationwide descriptive analysis rather than a causal evaluation of the NVBP. The price-index and PPP analyses describe observed procurement-price trends and regional price patterns, but they do not establish a counterfactual policy effect. Without an appropriate comparison group, a difference-in-differences design, or other quasi-experimental approach, we cannot fully separate NVBP-related changes from concurrent pharmaceutical reforms, market trends, manufacturer strategies, or broader macroeconomic conditions. Therefore, the observed price declines and regional patterns should be interpreted as trends under the NVBP rather than as definitive causal effects attributable solely to the policy.

Third, pharmaceutical supply shortages were categorized into graded types. This was done through predetermined criteria to enhance descriptive clarity and facilitate the analysis of the relationship between prices and shortages. As no standardized classification methodology exists for quantifying shortage severity, the current categorization serves primarily as an illustrative framework.Fourth, Since the actual clinical medication demand cannot be fully quantified, this study adopts ordered quantity as its proxy indicator, which inevitably involves a certain time-lag bias. In addition, this study did not explicitly adjust for the potential impacts of the COVID-19 pandemic. The pandemic overlapped with several procurement batches, but its influence was unlikely to be uniform across batches. Batch 2 was implemented in April 2020, during the early stage of the pandemic, when hospital utilization, logistics, and manufacturing capacity were substantially disrupted. Batches 3–5 were implemented during subsequent periods of intermittent pandemic control, while Batch 7 overlapped with the Omicron period in China. These heterogeneous contexts may have affected hospital ordering behavior, delivery performance, and observed supply gaps differently across batches. As this was a descriptive observational study, we could not isolate or quantify these batch-specific pandemic effects. In addition, we were unable to examine the potential association between low procurement prices and drug quality because quality-related indicators were not available in our dataset.

Although excessive price cuts may raise quality concerns, this study cannot provide quantitative evidence on this issue due to data limitations.Finally, the generalizability of our findings deserves caution. This study was conducted based on China’s national volume-based procurement (NVBP) policy, which operates within a distinct institutional environment characterized by dominant public hospitals, a centralized national medical insurance administration, and a state-led, near-compulsory procurement mechanism. Therefore, our conclusions may not be directly generalizable to other countries or regions with mixed public–private hospital systems, decentralized insurance structures, voluntary participation rules, or different pharmaceutical market regulatory frameworks. Caution is needed when extrapolating these results to other policy settings.

Conclusion

This nationwide descriptive study provides evidence on long-term procurement-price trends, regional price disparities, and supply shortages under China’s National Volume-Based Procurement policy. The findings show that bid-winning drugs experienced a rapid and substantial price decline after implementation and remained stable at low levels over time, while alternative drugs showed smaller and more fluctuating downward trends. These results suggest that centralized volume-based procurement can be associated with sustained low procurement prices.

Beyond price reduction, this study highlights two policy-relevant tensions. First, bid-winning drug prices were generally lower in economically developed regions than in underdeveloped regions, indicating that a national procurement framework does not automatically eliminate regional pricing inequity. Second, supply shortages remained evident during the first post-implementation year, especially among very low-priced drugs, suggesting that aggressive price competition may need to be balanced against supply stability.

For China and other countries considering centralized or volume-based pharmaceutical procurement, these findings suggest that price reduction should not be the only policy objective. If the associations observed in this study reflect underlying procurement mechanisms, future policy optimization may need to balance price containment with regional equity and supply continuity. Potential complementary mechanisms include regional price-differential monitoring, dynamic price-floor or price-adjustment rules for extremely low-priced medicines, supply-risk monitoring, supply guarantee penalties, and credit evaluation systems for participating enterprises.

China’s experience also suggests that procurement systems concentrating demand among selected suppliers should be accompanied by safeguards to prevent unintended regional and supply-side consequences. However, these lessons should be adapted to local institutional contexts, particularly in countries with decentralized insurance systems, mixed public–private delivery structures, or voluntary procurement participation. Further studies using causal designs, province-level economic indicators, manufacturer-level data, budget-impact analysis, and drug-quality information are needed to verify the mechanisms underlying the observed regional disparities and supply risks.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

NVBP

National Volume-Based Procurement

PPP

Purchasing Power Parity

CNY

Chinese Yuan

USD

United States dollar

NHS

National Health Service

NHSA

National Healthcare Security Administration of China

Author contributions

Rui Huang provided data and technical support, Li Luo and Junyu Niu carried out partial data management, Qianqian Yu and Gan Wang polished the language, Qianqian Yu provided critical comments, Biao Wang analyzed the data and wrote the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported by National Natural Science Foundation of China Scientists Fund(Project Number: 72174041) and Shanghai Center for Emerging Technologies Governance in Medicine and Public Health.

Data availability

The datasets generated or analysed during the current study are not publicly available due to confidentiality policies but are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable. This study used anonymized administrative procurement data and did not involve individual-level human subject data.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.