Does Payment Variation Across Insurance Types for Diagnosis-Intervention Packet Exacerbate Health Disparities?

Does Payment Variation Across Insurance Types for Diagnosis-Intervention Packet Exacerbate Health Disparities? | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Does Payment Variation Across Insurance Types for Diagnosis-Intervention Packet Exacerbate Health Disparities? Haoran Niu, Kunhe Lin, Yingbei Xiong, Tianqin Xue, Li Xiang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9057292/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 4 You are reading this latest preprint version Abstract Background China's Diagnosis-Intervention Packet (DIP) reform establishes an innovative payment mechanism that combines global budget ceilings with performance-based incentives. However, DIP uses dual-pool financing for Urban Employee Basic Medical Insurance (UEBMI) and Urban-Rural Resident Basic Medical Insurance (URRBMI), each with distinct point values, raising concerns about potential provider selection and equity distortion. This study examined whether DIP implementation influenced provider behavior differently across insurance types and whether point value disparities induced strategic selection against URRBMI patients over time. Methods We used administrative claims data from all public hospitals in City A, a nationally designated DIP pilot city in Central China. We conducted interrupted time series (ITS) analyses across three policy phases of DIP implementation (June 2021–December 2023), including 1,749,036 inpatient admissions across UEBMI and URRBMI schemes. Main outcomes included 7-day all-cause readmission (quality), low- and high-relative weight (RW) case shares (case mix), decomposed admissions (strategic behavior), and admission volume shifts by insurance type. Outcomes were aggregated monthly and modeled using ITS with interactions for insurance type and hospital level. Results Hospitals initially favored UEBMI patients, with higher high-RW case shares and lower decomposition rates. URRBMI admissions significantly increased during the first annual budget reconciliation cycle, particularly for low-RW and decomposed admissions, indicating purposeful budget absorption. Significant decreases in decomposition and readmission trends among URRBMI hospitals were noted following the second annual budget reconciliation cycle, suggesting regulatory containment. As case-mix convergence between insurance types developed over time, enhanced regulatory enforcement and standardization led to better procedural equity. Conclusions The implementation of DIP accompanied dynamic behavioral changes influenced by regulatory enforcement, policy feedback, and payment asymmetry. The observed convergence among insurance types was driven not just by equity gains, but also by institutional learning and budget optimization. Policymakers should consider regional capability and monitoring intensity to ensure that future payment models promote equity and efficiency. Figures Figure 1 1 Introduction Healthcare systems globally face the challenge of balancing cost containment with equitable access to care 1,2 . The shift from fee-for-service (FFS) to case-based payment models, such as Diagnosis-Related Groups (DRGs) 3–8 , aims to improve efficiency but can incentivize risk selection 9–1112–14 , particularly when payment rates differ across population groups 10,15,16 .. China’s ongoing payment reform, the Diagnosis-Intervention Packet (DIP) system. Introduced nationally in 2021, DIP replaces FFS with a data-driven classification system based on ICD-10/ICD-9-CM code 14,17 s, A Relative Weight (RW) is assigned to each group. Final payment is calculated as RW × Point Value (PV), where PV is determined annually based on the total budget and service volume 18 . This design aims to blend a global budget ceiling with performance-based incentives to guide provider behavior. However, China faces a unique equity challenge rooted not in race or ethnicity 13,19 , but in its dual insurance structure. The Urban Employee Basic Medical Insurance (UEBMI) serves working urban populations, while the Urban-Rural Resident Basic Medical Insurance (URRBMI) covers urban non-employed, rural, and low-income residents. Despite uniform case grouping, DIP administers UEBMI and URRBMI funds separately. Critically, the PV differs across the two insurance pools, resulting in identical treatments receiving systematically lower hospital payment under URRBMI 20 . This structural payment asymmetry gives rise to competing theoretical expectations. On one hand, providers may prioritize higher-paying UEBMI patients to maximize revenue, leading to selective under-treatment of URRBMI patients. On the other hand, DIP’s embedded regulatory controls—such as risk audits and outcome-linked penalties—may constrain such opportunistic behaviors. While studies have examined the effects of DRG-like reforms, few have focused on how provider behavior evolves over repeated feedback cycles 13 . As Liao et al. (2020) noted, providers often adjust their behavior not at the point of policy announcement, but in response to financial feedback during budget reconciliation 21 . Departing from conventional policy timing definitions, we anchor our analysis to these annual reconciliation cycles, when payment incentives fully materialize for providers. The equity implications of DIP, particularly how provider incentives and selection behaviors evolve over these repeated cycles, remain underexplored 22,23 . Using administrative data from City A, a national DIP pilot city, we applied interrupted time series analysis to examine: (1) whether DIP improved system efficiency and care quality; (2) whether institutional responses differed across insurance types; and (3) whether payment asymmetries induced selective treatment toward URRBMI patients—and if so, whether these effects changed over time. 2 Study design 2.1 Study Context and Reform Structure City A, a prefecture-level administrative region in Central China ,with a regional GDP exceeding 420 billion RMB (65 billion USD) in 2020, and a population of 4 million, implemented the DIP payment system in September 2021 as a national pilot. The reform covers all 48 public hospitals in the jurisdiction, spanning tertiary hospitals, secondary community hospitals, and primary care centers. The city's dual-pool insurance structure—with UEBMI and URRBMI—creates a natural experiment to examine how payment disparities shape hospital behavior. 2.2 DIP Payment Mechanism In response to the growing demand for cost control and efficiency in hospital finance, China introduced DIP, a global budget payment system designed as an alternative to DRG. While DRG categorizes patients based on objective information such as primary diagnosis, DIP employs a more rule-based, data-driven approach. It classifies patients by combining ICD-10 diagnosis codes and ICD-9-CM3 procedure codes, yielding a substantially more comprehensive set of categories. This technique creates around 10,000 separate case groups, giving payers more detail for pricing and monitoring hospital behavior than DRG systems, which typically contain only a few hundred groups. Each case group is assigned a RW reflecting its historical average cost relative to a baseline condition. However, and critically for this study, DIP operates under dual-pool budgeting, one pool for UEBMI and another for URRBMI. Each pool has its own annual fixed budget \(\:{\:\text{F}}_{\text{k}}\) (where k in UEBMI / URRBMI), and at the end of year, the monetary value per RW ( \(\:{\text{V}\text{a}\text{l}\text{u}\text{e}}_{\text{k}}\) ) is computed as: $$\:{\text{V}\text{a}\text{l}\text{u}\text{e}}_{\text{k}}\:=\:\frac{{\text{F}}_{\text{k}}}{{\text{N}}_{\text{k}}·{\stackrel{-}{\text{R}\text{W}}}_{\text{k}}}$$ Where \(\:{\text{N}}_{\text{k}}\) is the number of inpatient cases and \(\:{\stackrel{-}{\text{R}\text{W}}}_{\text{k}}\) the average case weight in the year. The final payment per case is then determined by: $$\:{Payment}_{ijk}={RW}_{ij}·{V}_{k}$$ The Final payment to hospitals is \(\:{Payment}_{ijk}\) , which covers the amount paid by insurance, while patients still bear out-of-pocket costs.the same DIP group receives different payment amounts depending on the patient’s insurance type. In City A, UEBMI’s final PV was nearly twice that of URRBMI, while URRBMI’s declined annually. As a result, identical cases receive substantially different payments. For instance, an acute appendicitis case (RW of 800) was reimbursed ¥2149 under UEBMI but only ¥1034 under URRBMI, about a 51.8% gap. Particularly in situations requiring a lot of resources, this approach may systematically bias provider preferences in favor of patients with higher-value insurance (UEBMI).2.3Regulatory Safeguards The DIP payment reform's inbuilt regulatory architecture is a critical component, with the goal of constraining hospitals' potential cost-based responses. Since ultimate payment amounts are established annually based on total budget and actual case weights, hospitals have enormous incentives to influence which patients they accept and how they offer care throughout the year. To mitigate risks such as upcoding, induced admissions, or patient selection, the Medical Security Bureau introduced a surveillance system powered by administrative big data. This system continuously evaluates hospitals on five risk dimensions: diagnosis inflation (upcoding), admission of low-complexity patients, short-interval readmissions, excessive length of stay, and mortality anomalies 24 . Each dimension is quantified through standardized indices that benchmark hospitals against historical and regional norms. Institutions identified as outliers in these indices face financial penalties, such as decreased year-end budget reconciliation, withholding quality assurance deposits, or PV downgrades in subsequent pricing cycles. Importantly, these regulatory procedures are intended not only to detect fraud, but also to improve cost-effectiveness and clinical appropriateness across both insurance groups. They serve as performance-based limitations, perhaps mitigating inequitable responses to pricing discrepancies. In theory, this makes DIP both a price-based reform and a discipline-based system. As a result, the provision and enforcement of these safeguards are critical to our second hypothesis: that administrative oversight procedures can prevent the formation of payment-induced injustice. 2.4 Policy Intervention Timing Although the DIP reform in City A was officially introduced in September 2021, its financial impact did not materialize immediately because actual reimbursements were only finalized at year-end through budget reconciliation. The first reconciliation, in May 2022, retrospectively adjusted payments for all cases from Q1 to Q4 2021, including those prior to the official rollout. These two points mark when actual price signals and financial consequences became fully concrete for hospitals, shaping their strategic decision-making. To evaluate behavioral changes around these key policy nodes, We applied an Interrupted Time Series (ITS) design with two intervention points 25 , May 2022 and May 2023, corresponding to the timing of annual budget reconciliations rather than new policy rollouts. This allows us to capture not only immediate adjustments following initial price realization but also longer-term adaptation as DIP shifted from initial rollout to a fully operational, institutionalized payment system.Furthermore, to explore whether responses varies by institutional capacity, we conducted ITS models stratified by hospital level( primary, secondary, and tertiary institutions), assessing whether pricing disparities or regulatory constraints had disproportionate effects across settings with varying clinical complexity and patient casemix . 2.5 Outcome Variable We used outcome indicators to evaluate how hospitals responded to DIP-induced price differences. These covered both efficiency and strategic behavior. 2.5.1 Efficiency and Quality Outcomes Efficiency was proxied by 7-day all-cause readmissions, widely used in DRG research 26,27 as a signal of premature discharge or poor care quality. Volume shifts by insurance type were also tracked monthly to detect strategic changes in case mix .This reflects institutional responses to budget caps and PV disparities, and has been used in several Chinese DIP studies as an indicator of strategic case mix adjustment 14 . 2.5.2 Behavior and Strategic Response Outcomes To examine potential provider manipulation or selective behavior, we included three additional outcome indicators. The first was the proportion of high-RW admissions, defined as cases with RW values above the 75th percentile within each insurance group. This measure reflects hospitals’ potential preference for high-revenue case types, often associated with cream-skimming under DRG-like systems 28 .We also examined the share of low-score admissions to detect inflation of minor cases, especially among URRBMI patients. Finally, Decomposed admissions are defined as split hospitalizations within 3 days for same patient and diagnosis, capturing potential case fragmentation 13 . All outcome variables were constructed at the patient level and aggregated monthly by insurance type and hospital tier. Definitions and coding thresholds were held constant throughout the study period. 2.6 Descriptive Trend Analysis Using Segmented Regression To describe temporal patterns in patient-level utilization behaviors during the implementation of the DIP payment reform, we employ an Interrupted Time Series (ITS) design with two pre-specified policy breakpoints: the first annual budget reconciliation in May 2022 and the second in May 2023. This approach allows us to characterize changes in the level (intercept shift) and trajectory (slope change) of outcome indicators coinciding with these recalibration points. Separate ITS models are fitted for each insurance scheme (UEBMI and URRBMI) to reflect their distinct benefit structures and baseline utilization profiles. It should be noted that this analysis is observational in nature; the estimated shifts should be interpreted as associations temporally aligned with policy milestones rather than causal effects, particularly given the concurrent influence of external factors such as the evolving pandemic context. $$\:{Y}_{it}={\beta\:}_{0}+{\beta\:}_{1}time+{{\beta\:}_{2}Policy}_{2}+{{\beta\:}_{3}Policy}_{3}+{{\beta\:}_{4}Policy}_{2}·time+{{\beta\:}_{5}Policy}_{3}·time+{\beta\:}_{7}hospita{l}_{level}+\epsilon\:$$ \(\:{Y}_{it}\) is the behavioral variable of patients with different medical insurance types i at time t (such as high score rate, low score rate, hospitalization breakdown rate, seven day readmission rate), \(\:{Policy}_{2}\) / \(\:{Policy}_{3}\) is the binary indicator variable (0/1), capturing the immediate level changes after the implementation of \(\:{Policy}_{2}·time\) / \(\:{Policy}_{3}·time\) is the interaction term between policy and time, capturing the changes in time trend after policy implementation. \(\:hospital\_level\) represents the level of the hospital and controls for the influence of confounding factors in classification. To further identify policy heterogeneity under different hospital levels, this paper estimates the grouping of \(\:hospital\_level\:\) and constructs GLM models for first, second, and third level medical institutions. potentially adjusted for covariates Vₖ, though not included in final models due to data constraints Additional Statistical Tests To ensure the robustness and validity of our findings, we perform several additional statistical tests: We run a pre-intervention trend stability test to determine that the outcome indicators were stable prior to the policy interventions. This is critical for developing a credible counterfactual scenario. Given the time series structure of our data, we use the Durbin-Watson test to detect autocorrelation in the residuals. This test ensures that significant autocorrelation does not violate our model assumptions. We use the Ljung-Box test to investigate the autocorrelation of residuals over multiple delays. This test gives a more comprehensive examination of autocorrelation in the residuals and contributes to the validation of our model's suitability. To investigate potential institutional heterogeneity, we run run stratified ITS across hospital tiers(primary, secondary, tertiary). This approach allows us to test whether different classes of providers—characterized by distinct technical capacity, bargaining power, and patient mix—respond differently to incentive structures and regulatory oversight. Additionally, we perform placebo tests using pre-reform pseudo-intervention dates to check for anticipatory effects or underlying trends. 3 Results 3.1 Basic information Table 1 presents changes in key indicators for UEBMI and URRBMI patients across three periods: June 2021–May 2022, May 2022–May 2023, and after May 2023. The decomposed admission rate for URRBMI remained at 0.1. UEBMI stayed at 0.0 across all stages. The 7-day readmission rate was stable. UEBMI ranged from 0.0 to 0.1. URRBMI stayed at 0.1.The average DIP RW decreased over time. UEBMI dropped from 1305.3 (SD 1789.4) to 1144.8 (SD 1528.3). URRBMI dropped from 1055.7 (SD 1464.9) to 927.3 (SD 1216.5). Insurance payments also declined. UEBMI decreased from 6393.9 yuan to 5316.2 yuan. URRBMI declined from 3797.6 yuan to 3239.1 yuan. For UEBMI, the high-RW case share declined from 0.3 to 0.2. Low-RW cases rose from 0.2 to 0.3. URRBMI showed fluctuation in low-RW cases (0.3 → 0.2 → 0.3), while high-RW cases declined from 0.3 to 0.2. URRBMI’s visits-per-patient increased to 1.30 in Stage 2, then returned to 1.24. UEBMI remained stable at around 1.15–1.16. Table 1 Characteristics of Decomposed Admissions, Medical Insurance Payments, High-RW and Low-RW Proportions Among Employees’ and Residents’ Basic Medical Insurance Enrollees During the DIP Reform in City A, 2021–2023 Insurance indicator June 2021–May 2022 May 2022–May 2023 after May 2023 UEBMI decomposed3 235 227 144 Points 1305.3(1789.4) 1185.6(1642.1) 1144.8(1528.3) high_RW 39642 44113 29672 insurance_pay 6393.9(9421.5) 6068.7(8876.5) 5316.2(7965.7) low_RW 35310 49853 34438 readmit7 3108 3675 3039 visits_per_patient 1.15 1.16 1.13 Patient number 144998 184856 126409 URRBMI decomposed3 2328 1474 905 Score 1055.7(1464.9) 993.5(1383.2) 927.3(1216.5) high_RW 114056 120740 81140 insurance_pay 3797.6(5655.6) 3712.7(5577.8) 3239.1(4627.1) low_RW 107704 122709 90106 readmit7 244990 28544 18740 visits_per_patient 1.24 1.30 1.24 Patient number 410031 504358 35764 Table 2 a: Main ITS Estimates: Impact of the DIP Reform on Patient-Level Behavioural Outcomes by Insurance Type, City A, 2021–2023 variable high_RW (UEBMI) high_RW (URRBMI) low_RW (UEBMI) low_RW (URRBMI) (Intercept) -2.621*** (0.046) -1.394*** (0.025) 0.155*** (0.043) 0.092** (0.024) time -0.008*** (0.0004) -0.017*** (0.0002) 0.003*** (0.0004) -0.004*** (0.0002) policy2 -0.684*** (0.065) -1.753*** (0.040) 0.257*** (0.067) -0.397*** (0.039) policy3 1.507*** (0.155) 1.890*** (0.097) -0.949*** (0.155) -1.817*** (0.092) policy2_time 0.006*** (0.0005) 0.016*** (0.0003) -0.002*** (0.0006) 0.003*** (0.0003) policy3_time -0.008*** (0.0009) -0.010*** (0.0005) 0.005*** (0.0009) 0.010*** (0.0005) hospital_level = 2 2.130*** (0.026) 2.262*** (0.009) -1.270*** (0.011) -0.846*** (0.005) hospital_level = 3 2.682*** (0.026) 2.860*** (0.009) -2.241*** (0.010) -1.743*** (0.007) Table 2 b Main ITS Estimates: Impact of the DIP Reform on Patient-Level Behavioural Outcomes by Insurance Type, City A, 2021–2023 variable decomposed (UEBMI) decomposed (URRBMI) readmit7 (UEBMI) readmit7 (URRBMI) (Intercept) -5.481*** (0.409) -5.392*** (0.136) -3.972*** (0.126) -3.994*** (0.047) time -0.001(0.004) 0.006*** (0.001) 0.016*** (0.001) 0.022*** (0.0005) policy2 -1.262 (0.744) -2.786*** (0.289) 0.601** (0.203) 0.175* (0.076) policy3 1.462 (1.898) 4.503*** (0.764) 0.821 (0.430) 6.443*** (0.177) policy2*time 0.006 (0.005) 0.012*** (0.002) -0.010*** (0.002) -0.010*** (0.001) policy3*time -0.009 (0.011) -0.029*** (0.004) -0.005 (0.002) -0.038*** (0.001) hospital_level = 2 -0.95*** (0.119) -0.781*** (0.034) -1.545*** (0.026) -1.575*** (0.009) hospital_level = 3 -0.729*** (0.099) -0.214*** (0.035) -1.896*** (0.024) -2.172*** (0.015) 3.2 ITS result Table 2 -a and 2 -b presents regression estimates for the four key outcomes: high-RW admissions, low-RW admissions, decomposed admissions, and 7-day readmissions, stratified by insurance type (UEBMI & URRBMI). For high-RW cases, both UEBMI and URRBMI groups showed significant declining trends over time. The coefficient for time was − 0.008 for UEBMI and − 0.017 for URRBMI. Policy 2 was associated with a sharp level decrease (β 2 = − 0.684 for UEBMI; − 1.753 for URRBMI), followed by a positive level shift after Policy 3. Both groups showed mild positive trend reversals following Policy 2, and moderate negative slopes after Policy 3. Hospitals at higher levels were significantly more likely to admit high-RW cases, as shown by the large positive coefficients for level 2 and 3 hospitals. For low-RW cases, the patterns were mixed. UEBMI patients showed an increasing trend over time (β 1 = 0.003), whereas URRBMI showed a slight decline (β 1 = − 0.004). Policy 2 increased the low-RW admission rate for UEBMI (β 2 = 0.257) but decreased it for URRBMI (β 2 = − 0.397). Policy 3 led to further declines in both groups. Time trends after Policy 3 indicated continued growth for URRBMI (β 5 = 0.010). Regarding decomposed admissions, the URRBMI group consistently showed higher rates than UEBMI. Both groups exhibited positive trends over time (UEBMI: β 4 = 0.006; URRBMI: β 4 = 0.012), with a strong level increase after Policy 3, especially for URRBMI (β 5 = 4.503). Post-policy slopes became negative (β =–0.029 for URRBMI), suggesting regulatory dampening of previously rising decomposition behaviors. For 7-day readmissions, baseline rates were similar between UEBMI and URRBMI (β 2 =-3.972 for UEBMI,-3.994 for URRBMI). Both groups exhibited significant increasing pre-policy trends (β 1 = 0.016 for UEBMI, 0.022 for URRBMI). Following Policy 2, both groups experienced significant immediate increases in rates, with a substantially larger spike observed for UEBMI (β 2 = 0.601 for UEBMI, 0.175 for URRBMI). After Policy 3, only URRBMI showed a significant immediate increase in levels (β 3 = 6.443), while UEBMI showed no significant change. Subsequently, URRBMI exhibited a significant negative trend during the post-Policy 3 period (β 5 = -0.038), indicating a gradual reduction following the initial Policy 3 surge." 3.3 Heterogeneity Analysis To assess institutional differences in response to DIP reform, we stratified key outcomes by hospital level and insurance type (UEBMI vs. URRBMI). Results are summarized below (full models in Additional Table 2 –5). Decomposed admissions were consistently higher among URRBMI patients, particularly in primary and secondary hospitals. A sharp rise was observed after Policy 3 (e.g., URRBMI secondary: β = 5.056, p < 0.01), followed by significantly negative time trends, suggesting regulatory suppression after initial expansion. Readmissions rose steadily across most tiers, but the largest level increases occurred among URRBMI patients in lower-tier facilities (primary: β = 7.965), with post-policy declines indicating later correction. High-RW case shares declined across all settings, with URRBMI hospitals showing steeper drops (e.g., tertiary: β = − 1.908). Time slopes post-policy were uniformly negative, suggesting continued de-selection of high-RW cases over time. Low-RW admissions showed divergent patterns: UEBMI hospitals had stable or mildly fluctuating trends, while URRBMI hospitals experienced an initial drop followed by a rebound (e.g., URRBMI secondary: β = − 2.924 then upward trend), possibly reflecting strategic adjustment to budget space. Overall, provider behavior under DIP varied by institutional tier, with more pronounced strategic and corrective dynamics observed in URRBMI-serving primary and secondary hospitals. 4 Discussion Our analysis revealed distinct phase-specific responses to DIP implementation. The initial response to DIP reform aligns with classic agency theory, where hospitals, acting as rational economic agents, responded directly to price signals. Both UEBMI and URRBMI patients showed declining trends in high-RW admissions, with steeper reductions among URRBMI patients. This indicates stronger incentives for selective under-treatment when point values are structurally lower, consistent with DRG-based reform experiences in Germany and Switzerland, where hospitals similarly avoided high-cost cases 26,29 . Such findings underscore that risk adjustment remains a critical equity safeguard in prospective payment systems 15 . However, this trend reversed following the first budget reconciliation. URRBMI admissions rebounded, especially in secondary and tertiary hospitals. This suggests that providers adapted after understanding that URRBMI budgets, despite lower point values, could still be efficiently used under high payment ratios 10 . Unlike in Western DRG systems where fiscal feedback is often delayed 26,29 , DIP’s annual cycles and regional budget caps may accelerate learning and adjustment 31 . This finding extends beyond a simple "provider selection" narrative, illustrating how providers strategically optimize their service mix in response to the complete set of financial rules, not just the per-case price. We also found shifts in patient case-mix. The proportion of high-RW admissions declined, while low-RW admissions expanded 14 . Prior DRG studies also observed similar substitution effects 26 . However, the dual insurance design in DIP adds complexity, as price incentives differ by population group, an underexplored area in DRG literature 26,29 . Strategic responses, such as readmission within seven days and case splitting, were observed, especially among URRBMI patients in primary hospitals 15,28 . These behaviors peaked after initial implementation, then declined. This pattern matches the timeline of outcome-linked penalties and audit enforcement. China’s real-time budget tools and settlement oversight likely played a role in faster behavioral correction, compared to slower post-hoc monitoring in European systems 32,33 . when paired with timely feedback and performance-based regulation (P4P), these behaviors may adjust. DIP’s outcome-linked penalties, reconciliation cycles, and audit mechanisms not only constrained opportunistic practices but also promoted convergence between UEBMI and URRBMI over time 15,28,31 . International evidence suggests that P4P reforms are most effective when quality monitoring is combined with detection of gaming behaviors 34 . This highlights that DIP’s equity function cannot rely on price adjustment alone, but requires continuous monitoring of heterogeneous provider responses 31,34 . 5 limitation This study has limitations. First, our quasi-experimental design may not fully isolate the policy effects from major concurrent events, most notably the disruptions caused by the COVID-19 pandemic. Second, patient-level tracking across hospitals or switching between insurances was unavailable. Third, findings are from one well-funded city and may not generalize. Lastly, we could not directly observe provider intent or internal decision-making. 6 Conclusion In conclusion, DIP introduced pricing gaps that caused early inequities in hospital care. Over time, providers adjusted behavior in response to financial and regulatory signals. Aligning payment rules and embedding oversight mechanisms are key to ensuring that such reforms improve efficiency without sacrificing fairness. Declarations Conflict of Interest Disclosures: No Conflict of Interest. Authors' information Li Xiang [Corresponding author], Email: [email protected] Ethics approval and consent to participate Research involving human data has been performed in accordance with the Declaration of Helsinki. All methods were carried out in accordance with relevant guidelines and regulations in the declaration. The study was approved by the Biomedical Ethics Review Committee of Tongji Medical College, Huazhong University of Science and Technology (S058, April 23, 2025). The need for informed consent was waived by the ethics institutional review board of Tongji Medical College, Huazhong University of Science and Technology because of the retrospective nature of the study. All authors confirm that this research caused no harm (physical or mental) to any participants. Funding This work was supported by the National Natural Science Foundation of China (Grant No. 72474073 and 72174068). Author Contribution Niu and Lin wrote the main manuscript text and Xiong and Xue prepared figures. All authors reviewed the manuscript. Data Availability The data that support the findings of this study are available from the Healthcare Security Administration of City A. Restrictions apply to the availability of these data, which were used under license for the current study and are not publicly available due to privacy regulations governing administrative health claims data. De-identified data are available from the corresponding author upon reasonable request and with permission from the Healthcare Security Administration of City A. References Liao JM, Navathe AS, Werner RM. The Impact of Medicare’s Alternative Payment Models on the Value of Care. Annu Rev Public Health. 2020;41(1):551-565. doi:10.1146/annurev-publhealth-040119-094327 Dzau VJ, Mate K, O’Kane M. Equity and Quality—Improving Health Care Delivery Requires Both. JAMA. 2022;327(6):519-520. doi:10.1001/jama.2022.0283 Quinn K. After the Revolution: DRGs at Age 30. Ann Intern Med. 2014;160(6):426-429. doi:10.7326/M13-2115 Jackson T, Dimitropoulos V, Madden R, Gillett S. Australian diagnosis related groups: Drivers of complexity adjustment. Health Policy. 2015;119(11):1433-1441. doi:10.1016/j.healthpol.2015.09.011 Rajtar M. Health care reform and Diagnosis Related Groups in Germany: The mediating role of Hospital Liaison Committees for Jehovah’s Witnesses. Soc Sci Med. 2016;166:57-65. doi:10.1016/j.socscimed.2016.08.016 Or. Z. Implementation of DRG Payment in France: Issues and recent developments. Health Policy. 2014;117(2):146-150. doi:10.1016/j.healthpol.2014.05.006 Wu Y, Lin Y, Cheng C, Fu C, Liao C, Hsieh C. Diagnosis‐Related Group (DRG)‐Based Prospective Hospital Payment System can be well adopted for Acute Care Surgery: Taiwanese Experience with Acute Cholecystitis. World J Surg. 2021;45(4):1080-1087. doi:10.1007/s00268-020-05904-5 Meng Z, Hui W, Cai Y, Liu J, Wu H. The effects of DRGs-based payment compared with cost-based payment on inpatient healthcare utilization: A systematic review and meta-analysis. Health Policy. 2020;124(4):359-367. doi:10.1016/j.healthpol.2020.01.007 Kantarevic J, Kralj B. Risk selection and cost shifting in a prospective physician payment system: Evidence from Ontario. Health Policy. 2014;115(2-3):249-257. doi:10.1016/j.healthpol.2013.10.002 Ellis RP, McGuire TG. Hospital response to prospective payment: Moral hazard, selection, and practice-style effects. J Health Econ. 1996;15(3):257-277. doi:10.1016/0167-6296(96)00002-1 Paddock SM, Escarce JJ, Hayden O, Buntin MB. Did the Medicare Inpatient Rehabilitation Facility Prospective Payment System Result in Changes in Relative Patient Severity and Relative Resource Use? Med Care. 2007;45(2):123-130. doi:10.1097/01.mlr.0000250863.65686.bc Liu M, Sandhu S, Joynt Maddox KE, Wadhera RK. Health Equity Adjustment and Hospital Performance in the Medicare Value-Based Purchasing Program. JAMA. 2024;331(16):1387-1396. doi:10.1001/jama.2024.2440 Xiang X, Dong L, Qi M, Wang H. How does diagnosis-related group payment impact the health care received by rural residents? Lessons learned from China. Public Health. 2024;232:68-73. doi:10.1016/j.puhe.2024.04.021 Lin K, Li Y, Yao Y, Xiong Y, Xiang L. The impact of an innovative payment method on medical expenditure, efficiency, and quality for inpatients with different types of medical insurance: evidence from a pilot city, China. Int J Equity Health. 2024;23(1):115. doi:10.1186/s12939-024-02196-2 Sood N, Huckfeldt PJ, Grabowski DC, Newhouse JP, Escarce JJ. The effect of prospective payment on admission and treatment policy: Evidence from inpatient rehabilitation facilities. J Health Econ. 2013;32(5):965-979. doi:10.1016/j.jhealeco.2013.05.003 Martinussen PE, Hagen TP. Reimbursement systems, organisational forms and patient selection: Evidence from day surgery in Norway. Health Econ Policy Law. 2009;4(2):139-158. doi:10.1017/S1744133109004812 Qian M, Zhang X, Chen Y, Xu S, Ying X. The pilot of a new patient classification-based payment system in China: The impact on costs, length of stay and quality. Soc Sci Med. 2021;289:114415. doi:10.1016/j.socscimed.2021.114415 National Healthcare Security Administration. Notice on issuing the three-year action plan for DRG/DIP payment reform. 2021. Accessed October 2, 2025. [in Chinese] https://www.gov.cn/zhengce/zhengceku/2021-11/28/content_5653858.htm Kim N, Jacobson M. Outcomes by Race and Ethnicity Following a Medicare Bundled Payment Program for Joint Replacement. JAMA Netw Open. 2024;7(9):e2433962. doi:10.1001/jamanetworkopen.2024.33962 Zhang X, Tang S, Wang R, Qian M, Ying X, Maciejewski ML. Hospital response to a new case-based payment system in China: the patient selection effect. Health Policy Plan. 2024;39(5):519-527. doi:10.1093/heapol/czae022 Liao JM, Navathe AS, Werner RM. The Impact of Medicare’s Alternative Payment Models on the Value of Care. Annu Rev Public Health. 2020;41:551-565. doi:10.1146/annurev-publhealth-040119-094327 Song Z, Ji Y, Safran DG, Chernew ME. Health Care Spending, Utilization, and Quality 8 Years into Global Payment. N Engl J Med. 2019;381(3):252-263. doi:10.1056/NEJMsa1813621 Schwartz AL, Chernew ME, Landon BE, McWilliams JM. Changes in Low-Value Services in Year 1 of the Medicare Pioneer Accountable Care Organization Program. JAMA Intern Med. 2015;175(11):1815. doi:10.1001/jamainternmed.2015.4525 Li Q, Yang C, Zhao Z, Chen Z, Feng Z, Huang D, Yin W. Research on the policy of Diagnosis-Intervention Packet (DIP) in China: a comparative analysis based on the national, provincial and municipal levels. Chin J Health Policy. 2022;15(7):8-15. Xie Y, Zhang H, Li W, Yan H, Duan H. Impact of Health All-in-One Machines on access to healthcare of rural areas in China: an interrupted time series analysis. BMC Health Serv Res. 2025;25(1):537. doi:10.1186/s12913-025-12710-z Kutz A, Gut L, Ebrahimi F, Wagner U, Schuetz P, Mueller B. Association of the Swiss Diagnosis-Related Group Reimbursement System With Length of Stay, Mortality, and Readmission Rates in Hospitalized Adult Patients. JAMA Netw Open. 2019;2(2):e188332. doi:10.1001/jamanetworkopen.2018.8332 Vuagnat A, Yilmaz E, Roussot A, et al. Did case-based payment influence surgical readmission rates in France? A retrospective study. BMJ Open. 2018;8(2):e018164. doi:10.1136/bmjopen-2017-018164 Ellis RP. Creaming, skimping and dumping: provider competition on the intensive and extensive margins. J Health Econ. 1998;17(5):537-555. doi:10.1016/S0167-6296(97)00042-8 Cook A, Averett S. Do hospitals respond to changing incentive structures? Evidence from Medicare’s 2007 DRG restructuring. J Health Econ. 2020;73:102319. doi:10.1016/j.jhealeco.2020.102319 Béland D. The Politics of Social Learning: Finance, Institutions, and Pension Reform in the United States and Canada. Governance. 2006;19(4):559-583. doi:10.1111/j.1468-0491.2006.00340.x Chang J, Chen S, Li A, et al. Facilitators and barriers to the implementation of DIP payment methodology reform in a public hospital in Guangzhou: a qualitative study based on the implementation of the meta-framework for research (CFIR) framework. Front Public Health. 2025;13. doi:10.3389/fpubh.2025.1569855 Zhang Y, Xu S yi, Tan G ming. Unraveling the effects of DIP payment reform on inpatient healthcare: insights into impacts and challenges. BMC Health Serv Res. 2024;24(1):887. doi:10.1186/s12913-024-11363-8 Zhang Y, Xu S, Tan G. Unraveling the effects of DIP payment reform on inpatient healthcare: insights into impacts and challenges. BMC Health Serv Res. 2024;24:887. doi:10.1186/s12913-024-11363-8 Van Herck, P., De Smedt, D., Annemans, L. et al. Systematic review: Effects, design choices, and context of pay-for-performance in health care. BMC Health Serv Res 10, 247 (2010). https://doi.org/10.1186/1472-6963-10-247 Additional Declarations No competing interests reported. Supplementary Files AdditionalTable.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 31 Mar, 2026 Editor assigned by journal 26 Mar, 2026 Submission checks completed at journal 26 Mar, 2026 First submitted to journal 07 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9057292","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":616825830,"identity":"c13378db-9786-4c22-9e7f-73210c24d4cc","order_by":0,"name":"Haoran Niu","email":"","orcid":"","institution":"Huazhong University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Haoran","middleName":"","lastName":"Niu","suffix":""},{"id":616825831,"identity":"00c52a90-0e04-4e8a-878c-74f573fd75a3","order_by":1,"name":"Kunhe Lin","email":"","orcid":"","institution":"Huazhong University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Kunhe","middleName":"","lastName":"Lin","suffix":""},{"id":616825832,"identity":"2f48dee6-175b-45b8-9ce5-cd05d4d4127b","order_by":2,"name":"Yingbei Xiong","email":"","orcid":"","institution":"Huazhong University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Yingbei","middleName":"","lastName":"Xiong","suffix":""},{"id":616825833,"identity":"91c9ef4d-1146-484a-9167-bd3c70b1d726","order_by":3,"name":"Tianqin Xue","email":"","orcid":"","institution":"Huazhong University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Tianqin","middleName":"","lastName":"Xue","suffix":""},{"id":616825834,"identity":"e7805f31-b1fd-45ba-a673-b89161cc01c0","order_by":4,"name":"Li Xiang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAz0lEQVRIie3RMQrCMBSA4Rci7aJ0TSnSK7Q46OBhzCWkg0MgkFNUe4VOBbeUQFyqs4ODk5NCnV1M7QFSN8H8Swi8j0cIgMv1i/mYQZbJ/jIeRDBi0DSGeF+Ro/iGJAoJUu8utCgYuj4ExHMbCXlHqhstNeB0KyDdMwsJMBLRs1K09MCLJgJWibQQD3dbckULAf5rEAk+hCnKtOGDiHkLX0itZqWmPMxPJC1tJDnw+iw3alpwVbf39TK2bjGNSH+aDwIg9nkTbgeNuVwu1//2BiuIQtxndBpfAAAAAElFTkSuQmCC","orcid":"","institution":"Huazhong University of Science and Technology","correspondingAuthor":true,"prefix":"","firstName":"Li","middleName":"","lastName":"Xiang","suffix":""}],"badges":[],"createdAt":"2026-03-07 09:39:02","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9057292/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9057292/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106404073,"identity":"0ea2ef11-d6bb-436c-ab52-c4f4cc35e957","added_by":"auto","created_at":"2026-04-08 09:15:26","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":197446,"visible":true,"origin":"","legend":"\u003cp\u003eUnnumbered image in Results \u0026nbsp;section.\u0026nbsp;\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9057292/v1/89fe369a70aec307f8aad951.png"},{"id":106405702,"identity":"333ff47d-7583-4578-8a08-94b2c43c3a46","added_by":"auto","created_at":"2026-04-08 09:28:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":984649,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9057292/v1/c38208e9-9cf9-4ce5-a7cc-e91696613772.pdf"},{"id":106300965,"identity":"c0145873-0f48-444d-9e94-621c366d9a3f","added_by":"auto","created_at":"2026-04-07 09:17:02","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":26913,"visible":true,"origin":"","legend":"","description":"","filename":"AdditionalTable.docx","url":"https://assets-eu.researchsquare.com/files/rs-9057292/v1/563cd9217880d8c707609a53.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Does Payment Variation Across Insurance Types for Diagnosis-Intervention Packet Exacerbate Health Disparities?","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eHealthcare systems globally face the challenge of balancing cost containment with equitable access to care\u003csup\u003e1,2\u003c/sup\u003e. The shift from fee-for-service (FFS) to case-based payment models, such as Diagnosis-Related Groups (DRGs) \u003csup\u003e3\u0026ndash;8\u003c/sup\u003e, aims to improve efficiency but can incentivize risk selection\u003csup\u003e9\u0026ndash;1112\u0026ndash;14\u003c/sup\u003e, particularly when payment rates differ across population groups\u003csup\u003e10,15,16\u003c/sup\u003e..\u003c/p\u003e \u003cp\u003eChina\u0026rsquo;s ongoing payment reform, the Diagnosis-Intervention Packet (DIP) system. Introduced nationally in 2021, DIP replaces FFS with a data-driven classification system based on ICD-10/ICD-9-CM code\u003csup\u003e14,17\u003c/sup\u003es, A Relative Weight (RW) is assigned to each group. Final payment is calculated as RW \u0026times; Point Value (PV), where PV is determined annually based on the total budget and service volume\u003csup\u003e18\u003c/sup\u003e. This design aims to blend a global budget ceiling with performance-based incentives to guide provider behavior.\u003c/p\u003e \u003cp\u003eHowever, China faces a unique equity challenge rooted not in race or ethnicity\u003csup\u003e13,19\u003c/sup\u003e, but in its dual insurance structure. The Urban Employee Basic Medical Insurance (UEBMI) serves working urban populations, while the Urban-Rural Resident Basic Medical Insurance (URRBMI) covers urban non-employed, rural, and low-income residents. Despite uniform case grouping, DIP administers UEBMI and URRBMI funds separately. Critically, the PV differs across the two insurance pools, resulting in identical treatments receiving systematically lower hospital payment under URRBMI\u003csup\u003e20\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThis structural payment asymmetry gives rise to competing theoretical expectations. On one hand, providers may prioritize higher-paying UEBMI patients to maximize revenue, leading to selective under-treatment of URRBMI patients. On the other hand, DIP\u0026rsquo;s embedded regulatory controls\u0026mdash;such as risk audits and outcome-linked penalties\u0026mdash;may constrain such opportunistic behaviors.\u003c/p\u003e \u003cp\u003eWhile studies have examined the effects of DRG-like reforms, few have focused on how provider behavior evolves over repeated feedback cycles\u003csup\u003e13\u003c/sup\u003e. As Liao et al. (2020) noted, providers often adjust their behavior not at the point of policy announcement, but in response to financial feedback during budget reconciliation\u003csup\u003e21\u003c/sup\u003e. Departing from conventional policy timing definitions, we anchor our analysis to these annual reconciliation cycles, when payment incentives fully materialize for providers. The equity implications of DIP, particularly how provider incentives and selection behaviors evolve over these repeated cycles, remain underexplored\u003csup\u003e22,23\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eUsing administrative data from City A, a national DIP pilot city, we applied interrupted time series analysis to examine: (1) whether DIP improved system efficiency and care quality; (2) whether institutional responses differed across insurance types; and (3) whether payment asymmetries induced selective treatment toward URRBMI patients\u0026mdash;and if so, whether these effects changed over time.\u003c/p\u003e"},{"header":"2 Study design","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study Context and Reform Structure\u003c/h2\u003e \u003cp\u003eCity A, a prefecture-level administrative region in Central China ,with a regional GDP exceeding 420\u0026nbsp;billion RMB (65\u0026nbsp;billion USD) in 2020, and a population of 4\u0026nbsp;million, implemented the DIP payment system in September 2021 as a national pilot. The reform covers all 48 public hospitals in the jurisdiction, spanning tertiary hospitals, secondary community hospitals, and primary care centers. The city's dual-pool insurance structure\u0026mdash;with UEBMI and URRBMI\u0026mdash;creates a natural experiment to examine how payment disparities shape hospital behavior.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 DIP Payment Mechanism\u003c/h2\u003e \u003cp\u003eIn response to the growing demand for cost control and efficiency in hospital finance, China introduced DIP, a global budget payment system designed as an alternative to DRG. While DRG categorizes patients based on objective information such as primary diagnosis, DIP employs a more rule-based, data-driven approach. It classifies patients by combining ICD-10 diagnosis codes and ICD-9-CM3 procedure codes, yielding a substantially more comprehensive set of categories. This technique creates around 10,000 separate case groups, giving payers more detail for pricing and monitoring hospital behavior than DRG systems, which typically contain only a few hundred groups.\u003c/p\u003e \u003cp\u003eEach case group is assigned a RW reflecting its historical average cost relative to a baseline condition. However, and critically for this study, DIP operates under dual-pool budgeting, one pool for UEBMI and another for URRBMI.\u003c/p\u003e \u003cp\u003eEach pool has its own annual fixed budget\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\:\\text{F}}_{\\text{k}}\\)\u003c/span\u003e\u003c/span\u003e (where k in UEBMI / URRBMI), and at the end of year, the monetary value per RW (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{V}\\text{a}\\text{l}\\text{u}\\text{e}}_{\\text{k}}\\)\u003c/span\u003e\u003c/span\u003e) is computed as:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:{\\text{V}\\text{a}\\text{l}\\text{u}\\text{e}}_{\\text{k}}\\:=\\:\\frac{{\\text{F}}_{\\text{k}}}{{\\text{N}}_{\\text{k}}\u0026middot;{\\stackrel{-}{\\text{R}\\text{W}}}_{\\text{k}}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{N}}_{\\text{k}}\\)\u003c/span\u003e\u003c/span\u003e is the number of inpatient cases and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\stackrel{-}{\\text{R}\\text{W}}}_{\\text{k}}\\)\u003c/span\u003e\u003c/span\u003e the average case weight in the year. The final payment per case is then determined by:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:{Payment}_{ijk}={RW}_{ij}\u0026middot;{V}_{k}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThe Final payment to hospitals is \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Payment}_{ijk}\\)\u003c/span\u003e\u003c/span\u003e, which covers the amount paid by insurance, while patients still bear out-of-pocket costs.the same DIP group receives different payment amounts depending on the patient\u0026rsquo;s insurance type. In City A, UEBMI\u0026rsquo;s final PV was nearly twice that of URRBMI, while URRBMI\u0026rsquo;s declined annually. As a result, identical cases receive substantially different payments. For instance, an acute appendicitis case (RW of 800) was reimbursed \u0026yen;2149 under UEBMI but only \u0026yen;1034 under URRBMI, about a 51.8% gap. Particularly in situations requiring a lot of resources, this approach may systematically bias provider preferences in favor of patients with higher-value insurance (UEBMI).2.3Regulatory Safeguards\u003c/p\u003e \u003cp\u003eThe DIP payment reform's inbuilt regulatory architecture is a critical component, with the goal of constraining hospitals' potential cost-based responses. Since ultimate payment amounts are established annually based on total budget and actual case weights, hospitals have enormous incentives to influence which patients they accept and how they offer care throughout the year. To mitigate risks such as upcoding, induced admissions, or patient selection, the Medical Security Bureau introduced a surveillance system powered by administrative big data. This system continuously evaluates hospitals on five risk dimensions: diagnosis inflation (upcoding), admission of low-complexity patients, short-interval readmissions, excessive length of stay, and mortality anomalies\u003csup\u003e24\u003c/sup\u003e. Each dimension is quantified through standardized indices that benchmark hospitals against historical and regional norms.\u003c/p\u003e \u003cp\u003eInstitutions identified as outliers in these indices face financial penalties, such as decreased year-end budget reconciliation, withholding quality assurance deposits, or PV downgrades in subsequent pricing cycles. Importantly, these regulatory procedures are intended not only to detect fraud, but also to improve cost-effectiveness and clinical appropriateness across both insurance groups. They serve as performance-based limitations, perhaps mitigating inequitable responses to pricing discrepancies. In theory, this makes DIP both a price-based reform and a discipline-based system. As a result, the provision and enforcement of these safeguards are critical to our second hypothesis: that administrative oversight procedures can prevent the formation of payment-induced injustice.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Policy Intervention Timing\u003c/h2\u003e \u003cp\u003eAlthough the DIP reform in City A was officially introduced in September 2021, its financial impact did not materialize immediately because actual reimbursements were only finalized at year-end through budget reconciliation. The first reconciliation, in May 2022, retrospectively adjusted payments for all cases from Q1 to Q4 2021, including those prior to the official rollout. These two points mark when actual price signals and financial consequences became fully concrete for hospitals, shaping their strategic decision-making.\u003c/p\u003e \u003cp\u003eTo evaluate behavioral changes around these key policy nodes, We applied an Interrupted Time Series (ITS) design with two intervention points\u003csup\u003e25\u003c/sup\u003e, May 2022 and May 2023, corresponding to the timing of annual budget reconciliations rather than new policy rollouts. This allows us to capture not only immediate adjustments following initial price realization but also longer-term adaptation as DIP shifted from initial rollout to a fully operational, institutionalized payment system.Furthermore, to explore whether responses varies by institutional capacity, we conducted \u003cb\u003eITS models\u003c/b\u003e stratified by hospital level( primary, secondary, and tertiary institutions), assessing whether pricing disparities or regulatory constraints had disproportionate effects across settings with varying \u003cb\u003eclinical complexity\u003c/b\u003e and \u003cb\u003epatient casemix\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Outcome Variable\u003c/h2\u003e \u003cp\u003eWe used outcome indicators to evaluate how hospitals responded to DIP-induced price differences. These covered both efficiency and strategic behavior.\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.5.1 Efficiency and Quality Outcomes\u003c/h2\u003e \u003cp\u003eEfficiency was proxied by 7-day all-cause readmissions, widely used in DRG research\u003csup\u003e26,27\u003c/sup\u003e as a signal of premature discharge or poor care quality.\u003c/p\u003e \u003cp\u003eVolume shifts by insurance type were also tracked monthly to detect strategic changes in case mix .This reflects institutional responses to budget caps and PV disparities, and has been used in several Chinese DIP studies as an indicator of strategic case mix adjustment\u003csup\u003e14\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e2.5.2 Behavior and Strategic Response Outcomes\u003c/h2\u003e \u003cp\u003eTo examine potential provider manipulation or selective behavior, we included three additional outcome indicators. The first was the proportion of high-RW admissions, defined as cases with RW values above the 75th percentile within each insurance group. This measure reflects hospitals\u0026rsquo; potential preference for high-revenue case types, often associated with cream-skimming under DRG-like systems\u003csup\u003e28\u003c/sup\u003e.We also examined the share of low-score admissions to detect inflation of minor cases, especially among URRBMI patients.\u003c/p\u003e \u003cp\u003eFinally, Decomposed admissions are defined as split hospitalizations within 3 days for same patient and diagnosis, capturing potential case fragmentation\u003csup\u003e13\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAll outcome variables were constructed at the patient level and aggregated monthly by insurance type and hospital tier. Definitions and coding thresholds were held constant throughout the study period.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Descriptive Trend Analysis Using Segmented Regression\u003c/h2\u003e \u003cp\u003eTo describe temporal patterns in patient-level utilization behaviors during the implementation of the DIP payment reform, we employ an Interrupted Time Series (ITS) design with two pre-specified policy breakpoints: the first annual budget reconciliation in May 2022 and the second in May 2023. This approach allows us to characterize changes in the level (intercept shift) and trajectory (slope change) of outcome indicators coinciding with these recalibration points. Separate ITS models are fitted for each insurance scheme (UEBMI and URRBMI) to reflect their distinct benefit structures and baseline utilization profiles. It should be noted that this analysis is observational in nature; the estimated shifts should be interpreted as associations temporally aligned with policy milestones rather than causal effects, particularly given the concurrent influence of external factors such as the evolving pandemic context.\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:{Y}_{it}={\\beta\\:}_{0}+{\\beta\\:}_{1}time+{{\\beta\\:}_{2}Policy}_{2}+{{\\beta\\:}_{3}Policy}_{3}+{{\\beta\\:}_{4}Policy}_{2}\u0026middot;time+{{\\beta\\:}_{5}Policy}_{3}\u0026middot;time+{\\beta\\:}_{7}hospita{l}_{level}+\\epsilon\\:$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{Y}_{it}\\)\u003c/span\u003e \u003c/span\u003e is the behavioral variable of patients with different medical insurance types i at time t (such as high score rate, low score rate, hospitalization breakdown rate, seven day readmission rate), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Policy}_{2}\\)\u003c/span\u003e\u003c/span\u003e/\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Policy}_{3}\\)\u003c/span\u003e\u003c/span\u003e is the binary indicator variable (0/1), capturing the immediate level changes after the implementation of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Policy}_{2}\u0026middot;time\\)\u003c/span\u003e\u003c/span\u003e/\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Policy}_{3}\u0026middot;time\\)\u003c/span\u003e\u003c/span\u003e is the interaction term between policy and time, capturing the changes in time trend after policy implementation. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:hospital\\_level\\)\u003c/span\u003e\u003c/span\u003e represents the level of the hospital and controls for the influence of confounding factors in classification. To further identify policy heterogeneity under different hospital levels, this paper estimates the grouping of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:hospital\\_level\\:\\)\u003c/span\u003e\u003c/span\u003eand constructs GLM models for first, second, and third level medical institutions. potentially adjusted for covariates Vₖ, though not included in final models due to data constraints\u003c/p\u003e \u003cp\u003eAdditional Statistical Tests\u003c/p\u003e \u003cp\u003eTo ensure the robustness and validity of our findings, we perform several additional statistical tests:\u003c/p\u003e \u003cp\u003eWe run a pre-intervention trend stability test to determine that the outcome indicators were stable prior to the policy interventions. This is critical for developing a credible counterfactual scenario.\u003c/p\u003e \u003cp\u003eGiven the time series structure of our data, we use the Durbin-Watson test to detect autocorrelation in the residuals. This test ensures that significant autocorrelation does not violate our model assumptions.\u003c/p\u003e \u003cp\u003eWe use the Ljung-Box test to investigate the autocorrelation of residuals over multiple delays. This test gives a more comprehensive examination of autocorrelation in the residuals and contributes to the validation of our model's suitability.\u003c/p\u003e \u003cp\u003eTo investigate potential institutional heterogeneity, we run run stratified ITS across hospital tiers(primary, secondary, tertiary). This approach allows us to test whether different classes of providers\u0026mdash;characterized by distinct technical capacity, bargaining power, and patient mix\u0026mdash;respond differently to incentive structures and regulatory oversight. Additionally, we perform placebo tests using pre-reform pseudo-intervention dates to check for anticipatory effects or underlying trends.\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Basic information\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents changes in key indicators for UEBMI and URRBMI patients across three periods: June 2021\u0026ndash;May 2022, May 2022\u0026ndash;May 2023, and after May 2023.\u003c/p\u003e \u003cp\u003eThe decomposed admission rate for URRBMI remained at 0.1. UEBMI stayed at 0.0 across all stages. The 7-day readmission rate was stable. UEBMI ranged from 0.0 to 0.1. URRBMI stayed at 0.1.The average DIP RW decreased over time. UEBMI dropped from 1305.3 (SD 1789.4) to 1144.8 (SD 1528.3). URRBMI dropped from 1055.7 (SD 1464.9) to 927.3 (SD 1216.5). Insurance payments also declined. UEBMI decreased from 6393.9 yuan to 5316.2 yuan. URRBMI declined from 3797.6 yuan to 3239.1 yuan.\u003c/p\u003e \u003cp\u003eFor UEBMI, the high-RW case share declined from 0.3 to 0.2. Low-RW cases rose from 0.2 to 0.3. URRBMI showed fluctuation in low-RW cases (0.3 \u0026rarr; 0.2 \u0026rarr; 0.3), while high-RW cases declined from 0.3 to 0.2. URRBMI\u0026rsquo;s visits-per-patient increased to 1.30 in Stage 2, then returned to 1.24. UEBMI remained stable at around 1.15\u0026ndash;1.16.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCharacteristics of Decomposed Admissions, Medical Insurance Payments, High-RW and Low-RW Proportions Among Employees\u0026rsquo; and Residents\u0026rsquo; Basic Medical Insurance Enrollees During the DIP Reform in City A, 2021\u0026ndash;2023\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInsurance\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eindicator\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eJune 2021\u0026ndash;May 2022\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMay 2022\u0026ndash;May 2023\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eafter May 2023\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"7\" rowspan=\"8\"\u003e \u003cp\u003eUEBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003edecomposed3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e235\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e227\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e144\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePoints\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1305.3(1789.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1185.6(1642.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1144.8(1528.3)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ehigh_RW\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39642\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44113\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e29672\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003einsurance_pay\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6393.9(9421.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6068.7(8876.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5316.2(7965.7)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elow_RW\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35310\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49853\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e34438\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ereadmit7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3108\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3675\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3039\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003evisits_per_patient\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.13\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePatient number\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e144998\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e184856\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e126409\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"7\" rowspan=\"8\"\u003e \u003cp\u003eURRBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003edecomposed3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2328\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1474\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e905\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eScore\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1055.7(1464.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e993.5(1383.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e927.3(1216.5)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ehigh_RW\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e114056\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e120740\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e81140\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003einsurance_pay\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3797.6(5655.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3712.7(5577.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3239.1(4627.1)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003elow_RW\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e107704\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e122709\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e90106\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ereadmit7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e244990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e28544\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e18740\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003evisits_per_patient\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.24\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePatient number\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e410031\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e504358\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e35764\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ea: Main ITS Estimates: Impact of the DIP Reform on Patient-Level Behavioural Outcomes by Insurance Type, City A, 2021\u0026ndash;2023\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003evariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ehigh_RW (UEBMI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ehigh_RW (URRBMI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003elow_RW (UEBMI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003elow_RW (URRBMI)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(Intercept)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-2.621*** (0.046)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-1.394*** (0.025)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.155*** (0.043)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.092** (0.024)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003etime\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.008*** (0.0004)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.017*** (0.0002)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.003*** (0.0004)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.004*** (0.0002)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.684*** (0.065)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-1.753*** (0.040)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.257*** (0.067)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.397*** (0.039)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.507*** (0.155)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.890*** (0.097)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.949*** (0.155)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-1.817*** (0.092)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy2_time\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.006*** (0.0005)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.016*** (0.0003)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.002*** (0.0006)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.003*** (0.0003)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy3_time\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.008*** (0.0009)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.010*** (0.0005)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.005*** (0.0009)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.010*** (0.0005)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehospital_level\u0026thinsp;=\u0026thinsp;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.130*** (0.026)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.262*** (0.009)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-1.270*** (0.011)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.846*** (0.005)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehospital_level\u0026thinsp;=\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.682*** (0.026)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.860*** (0.009)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-2.241*** (0.010)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-1.743*** (0.007)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eb Main ITS Estimates: Impact of the DIP Reform on Patient-Level Behavioural Outcomes by Insurance Type, City A, 2021\u0026ndash;2023\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003evariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003edecomposed (UEBMI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003edecomposed (URRBMI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ereadmit7 (UEBMI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ereadmit7 (URRBMI)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e(Intercept)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-5.481*** (0.409)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-5.392*** (0.136)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-3.972*** (0.126)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-3.994*** (0.047)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003etime\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.001(0.004)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.006*** (0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.016*** (0.001)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.022*** (0.0005)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-1.262 (0.744)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-2.786*** (0.289)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.601** (0.203)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.175* (0.076)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.462 (1.898)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.503*** (0.764)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.821 (0.430)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e6.443*** (0.177)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy2*time\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.006 (0.005)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.012*** (0.002)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.010*** (0.002)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.010*** (0.001)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003epolicy3*time\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.009 (0.011)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.029*** (0.004)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.005 (0.002)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.038*** (0.001)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehospital_level\u0026thinsp;=\u0026thinsp;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.95*** (0.119)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.781*** (0.034)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-1.545*** (0.026)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-1.575*** (0.009)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehospital_level\u0026thinsp;=\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.729*** (0.099)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.214*** (0.035)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-1.896*** (0.024)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-2.172*** (0.015)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.2 ITS result\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u0026lt;link rid=\"tb3\"\u0026gt;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e2\u0026lt;/link\u0026gt;\u003c/span\u003e-a and \u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e2\u003c/span\u003e-b presents regression estimates for the four key outcomes: high-RW admissions, low-RW admissions, decomposed admissions, and 7-day readmissions, stratified by insurance type (UEBMI \u0026amp; URRBMI).\u003c/p\u003e \u003cp\u003eFor high-RW cases, both UEBMI and URRBMI groups showed significant declining trends over time. The coefficient for time was \u0026minus;\u0026thinsp;0.008 for UEBMI and \u0026minus;\u0026thinsp;0.017 for URRBMI. Policy 2 was associated with a sharp level decrease (β\u003csub\u003e2\u003c/sub\u003e = \u0026minus;\u0026thinsp;0.684 for UEBMI; \u0026minus;\u0026thinsp;1.753 for URRBMI), followed by a positive level shift after Policy 3. Both groups showed mild positive trend reversals following Policy 2, and moderate negative slopes after Policy 3. Hospitals at higher levels were significantly more likely to admit high-RW cases, as shown by the large positive coefficients for level 2 and 3 hospitals.\u003c/p\u003e \u003cp\u003eFor low-RW cases, the patterns were mixed. UEBMI patients showed an increasing trend over time (β\u003csub\u003e1\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;0.003), whereas URRBMI showed a slight decline (β\u003csub\u003e1\u003c/sub\u003e = \u0026minus;\u0026thinsp;0.004). Policy 2 increased the low-RW admission rate for UEBMI (β\u003csub\u003e2\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;0.257) but decreased it for URRBMI (β\u003csub\u003e2\u003c/sub\u003e= \u0026minus;\u0026thinsp;0.397). Policy 3 led to further declines in both groups. Time trends after Policy 3 indicated continued growth for URRBMI (β\u003csub\u003e5\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;0.010).\u003c/p\u003e \u003cp\u003eRegarding decomposed admissions, the URRBMI group consistently showed higher rates than UEBMI. Both groups exhibited positive trends over time (UEBMI: β\u003csub\u003e4\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;0.006; URRBMI: β\u003csub\u003e4\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;0.012), with a strong level increase after Policy 3, especially for URRBMI (β\u003csub\u003e5\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;4.503). Post-policy slopes became negative (β =\u0026ndash;0.029 for URRBMI), suggesting regulatory dampening of previously rising decomposition behaviors.\u003c/p\u003e \u003cp\u003eFor 7-day readmissions, baseline rates were similar between UEBMI and URRBMI (β\u003csub\u003e2\u003c/sub\u003e=-3.972 for UEBMI,-3.994 for URRBMI). Both groups exhibited significant increasing pre-policy trends (β\u003csub\u003e1\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;0.016 for UEBMI, 0.022 for URRBMI). Following Policy 2, both groups experienced significant immediate increases in rates, with a substantially larger spike observed for UEBMI (β\u003csub\u003e2\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;0.601 for UEBMI, 0.175 for URRBMI). After Policy 3, only URRBMI showed a significant immediate increase in levels (β\u003csub\u003e3\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;6.443), while UEBMI showed no significant change. Subsequently, URRBMI exhibited a significant negative trend during the post-Policy 3 period (β\u003csub\u003e5\u003c/sub\u003e = -0.038), indicating a gradual reduction following the initial Policy 3 surge.\"\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Heterogeneity Analysis\u003c/h2\u003e \u003cp\u003eTo assess institutional differences in response to DIP reform, we stratified key outcomes by hospital level and insurance type (UEBMI vs. URRBMI). Results are summarized below (full models in Additional Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u0026ndash;5).\u003c/p\u003e \u003cp\u003eDecomposed admissions were consistently higher among URRBMI patients, particularly in primary and secondary hospitals. A sharp rise was observed after Policy 3 (e.g., URRBMI secondary: β\u0026thinsp;=\u0026thinsp;5.056, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01), followed by significantly negative time trends, suggesting regulatory suppression after initial expansion. Readmissions rose steadily across most tiers, but the largest level increases occurred among URRBMI patients in lower-tier facilities (primary: β\u0026thinsp;=\u0026thinsp;7.965), with post-policy declines indicating later correction. High-RW case shares declined across all settings, with URRBMI hospitals showing steeper drops (e.g., tertiary: β = \u0026minus;\u0026thinsp;1.908). Time slopes post-policy were uniformly negative, suggesting continued de-selection of high-RW cases over time. Low-RW admissions showed divergent patterns: UEBMI hospitals had stable or mildly fluctuating trends, while URRBMI hospitals experienced an initial drop followed by a rebound (e.g., URRBMI secondary: β = \u0026minus;\u0026thinsp;2.924 then upward trend), possibly reflecting strategic adjustment to budget space. Overall, provider behavior under DIP varied by institutional tier, with more pronounced strategic and corrective dynamics observed in URRBMI-serving primary and secondary hospitals.\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eOur analysis revealed distinct phase-specific responses to DIP implementation. The initial response to DIP reform aligns with classic agency theory, where hospitals, acting as rational economic agents, responded directly to price signals. Both UEBMI and URRBMI patients showed declining trends in high-RW admissions, with steeper reductions among URRBMI patients. This indicates stronger incentives for selective under-treatment when point values are structurally lower, consistent with DRG-based reform experiences in Germany and Switzerland, where hospitals similarly avoided high-cost cases\u003csup\u003e26,29\u003c/sup\u003e. Such findings underscore that risk adjustment remains a critical equity safeguard in prospective payment systems\u003csup\u003e15\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eHowever, this trend reversed following the first budget reconciliation. URRBMI admissions rebounded, especially in secondary and tertiary hospitals. This suggests that providers adapted after understanding that URRBMI budgets, despite lower point values, could still be efficiently used under high payment ratios \u003csup\u003e10\u003c/sup\u003e. Unlike in Western DRG systems where fiscal feedback is often delayed \u003csup\u003e26,29\u003c/sup\u003e, DIP\u0026rsquo;s annual cycles and regional budget caps may accelerate learning and adjustment \u003csup\u003e31\u003c/sup\u003e. This finding extends beyond a simple \"provider selection\" narrative, illustrating how providers strategically optimize their service mix in response to the complete set of financial rules, not just the per-case price.\u003c/p\u003e \u003cp\u003eWe also found shifts in patient case-mix. The proportion of high-RW admissions declined, while low-RW admissions expanded \u003csup\u003e14\u003c/sup\u003e. Prior DRG studies also observed similar substitution effects \u003csup\u003e26\u003c/sup\u003e. However, the dual insurance design in DIP adds complexity, as price incentives differ by population group, an underexplored area in DRG literature \u003csup\u003e26,29\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eStrategic responses, such as readmission within seven days and case splitting, were observed, especially among URRBMI patients in primary hospitals \u003csup\u003e15,28\u003c/sup\u003e. These behaviors peaked after initial implementation, then declined. This pattern matches the timeline of outcome-linked penalties and audit enforcement. China\u0026rsquo;s real-time budget tools and settlement oversight likely played a role in faster behavioral correction, compared to slower post-hoc monitoring in European systems\u003csup\u003e32,33\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003ewhen paired with timely feedback and performance-based regulation (P4P), these behaviors may adjust. DIP\u0026rsquo;s outcome-linked penalties, reconciliation cycles, and audit mechanisms not only constrained opportunistic practices but also promoted convergence between UEBMI and URRBMI over time \u003csup\u003e15,28,31\u003c/sup\u003e. International evidence suggests that P4P reforms are most effective when quality monitoring is combined with detection of gaming behaviors \u003csup\u003e34\u003c/sup\u003e. This highlights that DIP\u0026rsquo;s equity function cannot rely on price adjustment alone, but requires continuous monitoring of heterogeneous provider responses \u003csup\u003e31,34\u003c/sup\u003e.\u003c/p\u003e"},{"header":"5 limitation","content":"\u003cp\u003eThis study has limitations. First, our quasi-experimental design may not fully isolate the policy effects from major concurrent events, most notably the disruptions caused by the COVID-19 pandemic. Second, patient-level tracking across hospitals or switching between insurances was unavailable. Third, findings are from one well-funded city and may not generalize. Lastly, we could not directly observe provider intent or internal decision-making.\u003c/p\u003e"},{"header":"6 Conclusion","content":"\u003cp\u003eIn conclusion, DIP introduced pricing gaps that caused early inequities in hospital care. Over time, providers adjusted behavior in response to financial and regulatory signals. Aligning payment rules and embedding oversight mechanisms are key to ensuring that such reforms improve efficiency without sacrificing fairness.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eConflict of Interest Disclosures:\u003c/h2\u003e \u003cp\u003eNo Conflict of Interest.\u003c/p\u003e \u003c/p\u003e\u003cp\u003e \u003ch2\u003eAuthors' information\u003c/h2\u003e \u003cp\u003eLi Xiang [Corresponding author], Email: [email protected]\u003c/p\u003e \u003c/p\u003e\u003cp\u003e \u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e \u003cp\u003eResearch involving human data has been performed in accordance with the Declaration of Helsinki. All methods were carried out in accordance with relevant guidelines and regulations in the declaration. The study was approved by the Biomedical Ethics Review Committee of Tongji Medical College, Huazhong University of Science and Technology (S058, April 23, 2025). The need for informed consent was waived by the ethics institutional review board of Tongji Medical College, Huazhong University of Science and Technology because of the retrospective nature of the study. All authors confirm that this research caused no harm (physical or mental) to any participants.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis work was supported by the National Natural Science Foundation of China (Grant No. 72474073 and 72174068).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eNiu and Lin wrote the main manuscript text and Xiong and Xue prepared figures. All authors reviewed the manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data that support the findings of this study are available from the Healthcare Security Administration of City A. Restrictions apply to the availability of these data, which were used under license for the current study and are not publicly available due to privacy regulations governing administrative health claims data. De-identified data are available from the corresponding author upon reasonable request and with permission from the Healthcare Security Administration of City A.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eLiao JM, Navathe AS, Werner RM. The Impact of Medicare\u0026rsquo;s Alternative Payment Models on the Value of Care. Annu Rev Public Health. 2020;41(1):551-565. doi:10.1146/annurev-publhealth-040119-094327\u003c/li\u003e\n\u003cli\u003eDzau VJ, Mate K, O\u0026rsquo;Kane M. Equity and Quality\u0026mdash;Improving Health Care Delivery Requires Both. JAMA. 2022;327(6):519-520. doi:10.1001/jama.2022.0283\u003c/li\u003e\n\u003cli\u003eQuinn K. After the Revolution: DRGs at Age 30. Ann Intern Med. 2014;160(6):426-429. doi:10.7326/M13-2115\u003c/li\u003e\n\u003cli\u003eJackson T, Dimitropoulos V, Madden R, Gillett S. Australian diagnosis related groups: Drivers of complexity adjustment. Health Policy. 2015;119(11):1433-1441. doi:10.1016/j.healthpol.2015.09.011\u003c/li\u003e\n\u003cli\u003eRajtar M. Health care reform and Diagnosis Related Groups in Germany: The mediating role of Hospital Liaison Committees for Jehovah\u0026rsquo;s Witnesses. Soc Sci Med. 2016;166:57-65. doi:10.1016/j.socscimed.2016.08.016\u003c/li\u003e\n\u003cli\u003eOr. Z. Implementation of DRG Payment in France: Issues and recent developments. Health Policy. 2014;117(2):146-150. doi:10.1016/j.healthpol.2014.05.006\u003c/li\u003e\n\u003cli\u003eWu Y, Lin Y, Cheng C, Fu C, Liao C, Hsieh C. Diagnosis‐Related Group (DRG)‐Based Prospective Hospital Payment System can be well adopted for Acute Care Surgery: Taiwanese Experience with Acute Cholecystitis. World J Surg. 2021;45(4):1080-1087. doi:10.1007/s00268-020-05904-5\u003c/li\u003e\n\u003cli\u003eMeng Z, Hui W, Cai Y, Liu J, Wu H. The effects of DRGs-based payment compared with cost-based payment on inpatient healthcare utilization: A systematic review and meta-analysis. Health Policy. 2020;124(4):359-367. doi:10.1016/j.healthpol.2020.01.007\u003c/li\u003e\n\u003cli\u003eKantarevic J, Kralj B. Risk selection and cost shifting in a prospective physician payment system: Evidence from Ontario. Health Policy. 2014;115(2-3):249-257. doi:10.1016/j.healthpol.2013.10.002\u003c/li\u003e\n\u003cli\u003eEllis RP, McGuire TG. Hospital response to prospective payment: Moral hazard, selection, and practice-style effects. J Health Econ. 1996;15(3):257-277. doi:10.1016/0167-6296(96)00002-1\u003c/li\u003e\n\u003cli\u003ePaddock SM, Escarce JJ, Hayden O, Buntin MB. Did the Medicare Inpatient Rehabilitation Facility Prospective Payment System Result in Changes in Relative Patient Severity and Relative Resource Use? Med Care. 2007;45(2):123-130. doi:10.1097/01.mlr.0000250863.65686.bc\u003c/li\u003e\n\u003cli\u003eLiu M, Sandhu S, Joynt Maddox KE, Wadhera RK. Health Equity Adjustment and Hospital Performance in the Medicare Value-Based Purchasing Program. JAMA. 2024;331(16):1387-1396. doi:10.1001/jama.2024.2440\u003c/li\u003e\n\u003cli\u003eXiang X, Dong L, Qi M, Wang H. How does diagnosis-related group payment impact the health care received by rural residents? Lessons learned from China. Public Health. 2024;232:68-73. doi:10.1016/j.puhe.2024.04.021\u003c/li\u003e\n\u003cli\u003eLin K, Li Y, Yao Y, Xiong Y, Xiang L. The impact of an innovative payment method on medical expenditure, efficiency, and quality for inpatients with different types of medical insurance: evidence from a pilot city, China. Int J Equity Health. 2024;23(1):115. doi:10.1186/s12939-024-02196-2\u003c/li\u003e\n\u003cli\u003eSood N, Huckfeldt PJ, Grabowski DC, Newhouse JP, Escarce JJ. The effect of prospective payment on admission and treatment policy: Evidence from inpatient rehabilitation facilities. J Health Econ. 2013;32(5):965-979. doi:10.1016/j.jhealeco.2013.05.003\u003c/li\u003e\n\u003cli\u003eMartinussen PE, Hagen TP. Reimbursement systems, organisational forms and patient selection: Evidence from day surgery in Norway. Health Econ Policy Law. 2009;4(2):139-158. doi:10.1017/S1744133109004812\u003c/li\u003e\n\u003cli\u003eQian M, Zhang X, Chen Y, Xu S, Ying X. The pilot of a new patient classification-based payment system in China: The impact on costs, length of stay and quality. Soc Sci Med. 2021;289:114415. doi:10.1016/j.socscimed.2021.114415\u003c/li\u003e\n\u003cli\u003eNational Healthcare Security Administration. Notice on issuing the three-year action plan for DRG/DIP payment reform. 2021. Accessed October 2, 2025. [in Chinese] https://www.gov.cn/zhengce/zhengceku/2021-11/28/content_5653858.htm\u003c/li\u003e\n\u003cli\u003eKim N, Jacobson M. Outcomes by Race and Ethnicity Following a Medicare Bundled Payment Program for Joint Replacement. JAMA Netw Open. 2024;7(9):e2433962. doi:10.1001/jamanetworkopen.2024.33962\u003c/li\u003e\n\u003cli\u003eZhang X, Tang S, Wang R, Qian M, Ying X, Maciejewski ML. Hospital response to a new case-based payment system in China: the patient selection effect. Health Policy Plan. 2024;39(5):519-527. doi:10.1093/heapol/czae022\u003c/li\u003e\n\u003cli\u003eLiao JM, Navathe AS, Werner RM. The Impact of Medicare\u0026rsquo;s Alternative Payment Models on the Value of Care. Annu Rev Public Health. 2020;41:551-565. doi:10.1146/annurev-publhealth-040119-094327\u003c/li\u003e\n\u003cli\u003eSong Z, Ji Y, Safran DG, Chernew ME. Health Care Spending, Utilization, and Quality 8 Years into Global Payment. N Engl J Med. 2019;381(3):252-263. doi:10.1056/NEJMsa1813621\u003c/li\u003e\n\u003cli\u003eSchwartz AL, Chernew ME, Landon BE, McWilliams JM. Changes in Low-Value Services in Year 1 of the Medicare Pioneer Accountable Care Organization Program. JAMA Intern Med. 2015;175(11):1815. doi:10.1001/jamainternmed.2015.4525\u003c/li\u003e\n\u003cli\u003eLi Q, Yang C, Zhao Z, Chen Z, Feng Z, Huang D, Yin W. Research on the policy of Diagnosis-Intervention Packet (DIP) in China: a comparative analysis based on the national, provincial and municipal levels. Chin J Health Policy. 2022;15(7):8-15.\u003c/li\u003e\n\u003cli\u003eXie Y, Zhang H, Li W, Yan H, Duan H. Impact of Health All-in-One Machines on access to healthcare of rural areas in China: an interrupted time series analysis. BMC Health Serv Res. 2025;25(1):537. doi:10.1186/s12913-025-12710-z\u003c/li\u003e\n\u003cli\u003eKutz A, Gut L, Ebrahimi F, Wagner U, Schuetz P, Mueller B. Association of the Swiss Diagnosis-Related Group Reimbursement System With Length of Stay, Mortality, and Readmission Rates in Hospitalized Adult Patients. JAMA Netw Open. 2019;2(2):e188332. doi:10.1001/jamanetworkopen.2018.8332\u003c/li\u003e\n\u003cli\u003eVuagnat A, Yilmaz E, Roussot A, et al. Did case-based payment influence surgical readmission rates in France? A retrospective study. BMJ Open. 2018;8(2):e018164. doi:10.1136/bmjopen-2017-018164\u003c/li\u003e\n\u003cli\u003eEllis RP. Creaming, skimping and dumping: provider competition on the intensive and extensive margins. J Health Econ. 1998;17(5):537-555. doi:10.1016/S0167-6296(97)00042-8\u003c/li\u003e\n\u003cli\u003eCook A, Averett S. Do hospitals respond to changing incentive structures? Evidence from Medicare\u0026rsquo;s 2007 DRG restructuring. J Health Econ. 2020;73:102319. doi:10.1016/j.jhealeco.2020.102319\u003c/li\u003e\n\u003cli\u003eB\u0026eacute;land D. The Politics of Social Learning: Finance, Institutions, and Pension Reform in the United States and Canada. Governance. 2006;19(4):559-583. doi:10.1111/j.1468-0491.2006.00340.x\u003c/li\u003e\n\u003cli\u003eChang J, Chen S, Li A, et al. Facilitators and barriers to the implementation of DIP payment methodology reform in a public hospital in Guangzhou: a qualitative study based on the implementation of the meta-framework for research (CFIR) framework. Front Public Health. 2025;13. doi:10.3389/fpubh.2025.1569855\u003c/li\u003e\n\u003cli\u003eZhang Y, Xu S yi, Tan G ming. Unraveling the effects of DIP payment reform on inpatient healthcare: insights into impacts and challenges. BMC Health Serv Res. 2024;24(1):887. doi:10.1186/s12913-024-11363-8\u003c/li\u003e\n\u003cli\u003eZhang Y, Xu S, Tan G. Unraveling the effects of DIP payment reform on inpatient healthcare: insights into impacts and challenges. BMC Health Serv Res. 2024;24:887. doi:10.1186/s12913-024-11363-8 \u003c/li\u003e\n\u003cli\u003eVan Herck, P., De Smedt, D., Annemans, L. et al. Systematic review: Effects, design choices, and context of pay-for-performance in health care. BMC Health Serv Res 10, 247 (2010). https://doi.org/10.1186/1472-6963-10-247\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"international-journal-for-equity-in-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ijeh","sideBox":"Learn more about [International Journal for Equity in Health](http://equityhealthj.biomedcentral.com)","snPcode":"12939","submissionUrl":"https://submission.nature.com/new-submission/12939/3","title":"International Journal for Equity in Health","twitterHandle":"@equityhealthj","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9057292/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9057292/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eChina's Diagnosis-Intervention Packet (DIP) reform establishes an innovative payment mechanism that combines global budget ceilings with performance-based incentives. However, DIP uses dual-pool financing for Urban Employee Basic Medical Insurance (UEBMI) and Urban-Rural Resident Basic Medical Insurance (URRBMI), each with distinct point values, raising concerns about potential provider selection and equity distortion. This study examined whether DIP implementation influenced provider behavior differently across insurance types and whether point value disparities induced strategic selection against URRBMI patients over time.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe used administrative claims data from all public hospitals in City A, a nationally designated DIP pilot city in Central China. We conducted interrupted time series (ITS) analyses across three policy phases of DIP implementation (June 2021\u0026ndash;December 2023), including 1,749,036 inpatient admissions across UEBMI and URRBMI schemes. Main outcomes included 7-day all-cause readmission (quality), low- and high-relative weight (RW) case shares (case mix), decomposed admissions (strategic behavior), and admission volume shifts by insurance type. Outcomes were aggregated monthly and modeled using ITS with interactions for insurance type and hospital level.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eHospitals initially favored UEBMI patients, with higher high-RW case shares and lower decomposition rates. URRBMI admissions significantly increased during the first annual budget reconciliation cycle, particularly for low-RW and decomposed admissions, indicating purposeful budget absorption. Significant decreases in decomposition and readmission trends among URRBMI hospitals were noted following the second annual budget reconciliation cycle, suggesting regulatory containment. As case-mix convergence between insurance types developed over time, enhanced regulatory enforcement and standardization led to better procedural equity.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThe implementation of DIP accompanied dynamic behavioral changes influenced by regulatory enforcement, policy feedback, and payment asymmetry. The observed convergence among insurance types was driven not just by equity gains, but also by institutional learning and budget optimization. Policymakers should consider regional capability and monitoring intensity to ensure that future payment models promote equity and efficiency.\u003c/p\u003e","manuscriptTitle":"Does Payment Variation Across Insurance Types for Diagnosis-Intervention Packet Exacerbate Health Disparities?","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-07 09:16:40","doi":"10.21203/rs.3.rs-9057292/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-04-01T00:46:00+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-26T17:20:43+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-26T17:20:08+00:00","index":"","fulltext":""},{"type":"submitted","content":"International Journal for Equity in Health","date":"2026-03-07T09:24:16+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"international-journal-for-equity-in-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ijeh","sideBox":"Learn more about [International Journal for Equity in Health](http://equityhealthj.biomedcentral.com)","snPcode":"12939","submissionUrl":"https://submission.nature.com/new-submission/12939/3","title":"International Journal for Equity in Health","twitterHandle":"@equityhealthj","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"032bb3f8-16b8-4fb9-bc21-7a5b1df7f4cd","owner":[],"postedDate":"April 7th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-07T09:16:40+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-07 09:16:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9057292","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"identity":"rs-9057292","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}