The burdens of low-value care in hysterectomy attributable to hospital ownership in China

The burdens of low-value care in hysterectomy attributable to hospital ownership in China | 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 Article The burdens of low-value care in hysterectomy attributable to hospital ownership in China Jay Pan, Huijia Luo, Tianjiao Lan, Peter Coyte, Ke Ju This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3639662/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Scholarly attention has been dedicated to the identification of low-value care (care that is not expected to provide a net benefit). Despite a consensus on the importance of hospital characteristics in explaining the use of low-value care, the precise influence of hospital ownership, herein the distinction between public and private ownership, remains unclear. This study included 38,865 hospital discharge records with hysterectomy procedures in China from 2016 to 2020 to describe the effect of public and private hospital ownership on the provision of low-value care and estimate the attributable risk ratio and corresponding attributable burden. Private hospitals were more likely to provide low-value hysterectomies, with the average incremental effect of 33.7% (95% CI, 23.5–42.5%). Potential interventions in private hospitals could reduce this a maximum of 9.7% (95% eCI, 8.7–10.4%) of low-value hysterectomy cases, corresponding to 48,375 (95% eCI, 43,254, to 51,706) annual cases and 1.82 (95% eCI, 1.63 to 1.94) billion USD costs nationally. For the first time, we identified the potential intervention target and estimated the maximum effectiveness of interventions to eliminate excessive risk of low-value care. Health sciences/Health care Health sciences/Health care/Health services Figures Figure 1 Introduction Low-value care describes services where no evidence of patient benefit exists or where the harms or costs of care outweigh potential benefits. 1 However, despite the limited experience of measuring such services in low- to middle-income countries like China, 2 based on empirical evidence, these services, albeit necessitating context-specific measurement methodologies, can maintain standardized definitions for the evaluation of the extent of specific inappropriate healthcare services 3 , 4 . In the past few decades, several initiatives have focused on reducing the use of low-value care. Despite these efforts, the prevalence of low-value care and its associated spending remains high. 5 , 6 Recently, studies have focused on de-implementation of low-value care. The majority of these studies proposed effective interventions targeting specific service providers (e.g., teaching hospitals 7 ) by identifying drivers of low-value care delivery, using either quantitative or qualitative methods. 8 – 10 However, these studies often lacked ex-ante estimates of the maximum effectiveness of potential interventions targeting such subjects. Therefore, it becomes challenging to pre-assess the potential value of implementing interventions on a specific type of provider. Conversely, the availability of such ex-ante estimates would enable researchers to conduct a cost-effectiveness analysis before implementing interventions, as they would possess a general understanding of the anticipated costs associated with the interventions. Prior studies appealed for interventions at the hospital level to reduce low-value care provision. 11 , 12 Several studies have examined variations in low-value care across diverse hospital settings. 6 , 13 , 14 Nevertheless, there is still a lack of evidence about the association between hospital ownership and low-value care delivery. Differences between public and private hospitals may manifest in varying availability of appropriate equipment 15 , 16 (attributed to distinct funding sources, resource allocation, and business purpose), the experience levels of healthcare personnels 17 , 18 (due to divergent working conditions and career progression prospects), and physician's decision-making 19 (influenced by different patient volumes, patient reimbursement capacity, and regulatory frameworks). These discrepancies have the potential to lead to the delivery of inappropriate medical services, consequently reflecting variations in the provision of low-value care. To explore the above issue, we chose a representative low-value service, low-value hysterectomy, for illustration. This choice was primarily predicated on its well-established definition, 20 , 21 high prevalence, 2 , 22 and potential for intervention. The current definition relies heavily on clinical evidence, while there is an ongoing need for further enhancement and supplementation with economic evidence. 21 Based on hospital discharge records from a populated province with 83 million residents in China during 2016 and 2020, we employed the instrumental variable method (to address potential unmeasured bias in association identification) and took hospital ownership as a case study to estimate the attributable risk ratio and corresponding contribution degree associated with its influence on the provision of low-value hysterectomies. Results Study sample characteristics We identified 34,661 qualified episodes in public hospitals and 4,204 qualified episodes in private hospitals. Table 1 provides the characteristics associated with these qualified episodes. 343 public hospitals provide a total of 5,940 (17.1%; 95% CI, 16.7–17.5%) low-value hysterectomy episodes, and 210 private hospitals provide a total of 961 (22.9%; 95% CI, 21.6–24.1%) low-value hysterectomy episodes. The distribution of covariates is not balanced between episodes in public and private hospitals, for example, in employed inpatients or inpatients with CCI over 0, rates of low-value hysterectomy episodes in private hospitals are higher than public ones. This implies that the endogeneity problem indeed exists. Besides, the average differential distance in public hospitals is 3.1 kilometers, which suggests the patient in public hospitals has to travel 3.1 additional kilometers to the nearest private hospitals relative to the nearest public hospital, supporting the association between our IV and patients’ choice of hospital. Table 1 Descriptive statistics a, b Characteristics Public Hospital Private Hospital Overall Qualified episodes Low-value episodes PLVE (95% CI), % Qualified episodes Low-value episodes PLVE (95% CI), % Qualified episodes Low-value episodes PLVE (95% CI), % Patient characteristics Age, mean (SD) 50 (8.4) 50 (8.7) 50 (8.6) 50 (8.6) 50 (8.4) 50 (8.7) Minority, n (%) Yes 703 (2.0) 175 (2.9) 24.9 (21.7, 28.1) 83 (2.0) 25 (2.6) 30.1 (20.3, 40.0) 786 (2.0) 200 (2.9) 25.5 (22.4, 28.5) No 33,958 (98.0) 5,765 (97.1) 17.0 (16.6, 17.4) 4,121 (98.0) 936 (97.4) 22.7 (21.4, 24.0) 38,079 (98.0) 6,701 (97.1) 17.6 (17.2, 18.0) Marital status, n (%) Unmarried 1,308 (3.8) 189 (3.2) 14.5 (12.4, 16.4) 128 (3.0) 28 (2.9) 21.9 (14.7, 29.0) 1,436 (3.7) 217 (3.1) 15.1 (13.3, 17.0) Married 32,516 (93.8) 5,538 (93.2) 17.0 (16.6, 17.4) 4,001 (95.2) 911 (94.8) 22.8 (21.5, 24.1) 36,517 (94.0) 6449 (93.5) 17.7 (17.3, 18.1) Others 837 (2.4) 213 (3.6) 25.5 (22.5, 28.4) 75 (1.8) 22 (2.3) 29.3 (19.0, 39.6) 912 (2.3) 235 (3.4) 25.6 (22.9, 28.6) Occupation, n (%) Employed 16,899 (48.8) 3,535 (59.5) 20.9 (20.3, 21.5) 2,697 (64.2) 637 (66.3) 23.6 (22.0, 25.2) 19,596 (50.4) 4,172 (60.5) 21.3 (20.7, 21.9) Retired 724 (2.0) 68 (1.2) 9.4 (7.3, 11.5) 92 (2.1) 15 (1.5) 16.3 (8.8, 23.9) 816 (2.1) 83 (1.2) 10.2 (8.1, 12.3) Others 17,038 (49.2) 2,337 (39.3) 13.7 (13.2, 14.2) 1,415 (33.7) 309 (32.2) 21.8 (19.7, 24.0) 18,453 (47.5) 2,646 (38.3) 14.3 (13.8, 14.8) Insurance type, n (%) UEBMI 6,271 (18.1) 735 (12.4) 11.7 (10.9, 12.5) 765 (18.2) 111 (11.6) 14.5 (12.0, 17.0) 7,036 (18.1) 846 (12.3) 12.0 (11.3, 12.8) URBMI 12,403 (35.8) 2,236 (37.6) 18.0 (17.4, 18.7) 1,686 (40.1) 466 (48.5) 27.6 (25.5, 29.8) 14,089 (36.3) 2,702 (39.2) 19.2 (18.5, 19.8) NCMS 8,501 (24.5) 2,127 (35.8) 25.0 (24.1, 25.9) 1,176 (28.0) 259 (27.0) 22.0 (19.7, 24.4) 9,677 (24.9) 2,386 (34.6) 24.7 (23.8, 25.5) Fully self-paid 2,661 (7.7) 375 (6.3) 14.1 (12.8, 15.4) 343 (8.2) 68 (7.0) 19.8 (15.6, 24.0) 3,004 (7.7) 443 (6.4) 14.8 (13.5, 16.0) Others 4,825 (13.9) 467 (7.9) 9.7 (8.8, 10.5) 234 (5.5) 57 (5.9) 24.4 (18.9, 29.9) 5,059 (13.0) 524 (7.5) 10.4 (9.5, 11.2) Hospital stay, mean (SD) 10 (4.6) 11 (4.9) 11 (7.0) 11 (4.8) 10 (4.9) 11 (4.9) Admission pathway, n (%) ED 2,659 (7.7) 586 (9.9) 22.0 (20.5, 23.6) 354 (8.4) 31 (3.2) 8.8 (5.8, 11.7) 3,013 (7.8) 617 (8.9) 20.5 (19.0, 21.9) OPD 30,936 (89.3) 5,221 (87.9) 16.9 (16.5, 17.3) 3,790 (90.2) 927 (96.5) 24.5 (23.1, 25.8) 34,726 (89.4) 6,148 (89.1) 17.7 (17.3, 18.1) Others 1,066 (3.0) 133 (2.2) 12.5 (10.5, 14.5) 60 (1.4) 3 (0.3) 5 (-0.5, 10.5) 1,126 (2.8) 136 (2.0) 12.1 (10.2, 14.0) CCI, n (%) 0 34,223 (98.7) 5,818 (97.9) 17.0 (16.6, 17.4) 4,166 (99.1) 942 (98.0) 22.6 (21.3, 23.9) 38,389 (98.8) 6,760 (98.0) 17.6 (17.2, 18.0) > 0 438 (1.3) 122 (2.1) 27.9 (23.7, 32.1) 38 (0.9) 19 (2.0) 50.0 (34.1, 65.9) 476 (1.2) 141 (2.0) 29.6 (25.5, 33.7) Hospital characteristics Hospital level, n (%) Tertiary 27,186 (78.4) 3,950 (66.5) 14.5 (14.1, 15.0) 259 (6.2) 12 (1.2) 4.6 (2.1, 7.2) 27,445 (70.6) 3,962 (57.4) 14.4 (14.0, 14.9) Secondary 7,446 (21.5) 1,985 (33.4) 26.7 (25.7, 27.7) 2,536 (60.3) 584 (60.8) 23.0 (21.4, 24.7) 9,982 (25.7) 2,569 (37.2) 25.7 (24.9, 26.6) Primary or others 29 (0.1) 5 (0.1) 17.2 (3.5, 31.0) 1,409 (33.5) 365 (38.0) 25.9 (23.6, 28.2) 1,438 (3.7) 370 (5.4) 25.7 (23.5, 28.0) Medical services type, n (%) Western 30,098 (86.8) 4,715 (79.4) 15.7 (15.3, 16.1) 3,952 (94.0) 851 (88.6) 21.5 (20.3, 22.8) 34,050 (87.6) 5,566 (80.7) 16.4 (16.0, 16.7) Traditional Chinese 4,563 (13.2) 1,225 (20.6) 26.9 (25.6, 28.1) 252 (6.0) 110 (11.4) 43.7 (37.5, 49.8) 4,815 (12.4) 1,335 (19.3) 27.7 (26.5, 29.0) Hospital location, n (%) Rural 11,070 (31.9) 2,862 (48.2) 25.9 (25.0, 26.7) 1,696 (40.3) 510 (53.1) 30.1 (27.9, 32.3) 12,766 (32.8) 3,372 (48.9) 26.4 (25.7, 27.2) Urban 23,591 (68.1) 3,078 (51.8) 13.1 (12.6, 13.5) 2,508 (59.7) 451 (46.9) 18.0 (16.5, 19.5) 26,099 (67.2) 3,529 (51.1) 13.5 (13.1, 13.9) Area economic characteristics GDP per capita, 1000 USD mean (SD) 9.30 (5.36) 7.10 (3.86) 8.31 (4.74) 7.29 (4.19) 9.19 (5.30) 7.13 (3.91) UR, mean (SD) 63.0 (22.0) 53.0 (17.6) 56.5 (19.0) 52.1 (17.7) 62.3 (21.8) 52.9 (17.6) Year, n (%) 2016 7,378 (21.3) 1,391 (23.4) 18.9 (18.0, 19.8) 720 (17.1) 131 (13.6) 18.2 (15.4, 21.0) 8,098 (20.8) 1,522 (22.1) 18.8 (17.9, 19.7) 2017 6,626 (19.1) 1,494 (25.2) 22.6 (21.5, 23.6) 985 (23.4) 255 (26.5) 25.9 (23.2, 28.6) 7,611 (19.6) 1,749 (25.3) 23.0 (22.0, 23.9) 2018 7,204 (20.8) 1,271 (21.4) 17.6 (16.8, 18.5) 960 (22.8) 225 (23.4) 23.4 (20.8, 26.1) 8,164 (21.0) 1,496 (21.7) 18.3 (17.5, 19.2) 2019 6,975 (20.1) 1,020 (17.2) 14.6 (13.8, 15.5) 817 (19.4) 196 (20.4) 24.0 (21.1, 26.9) 7,792 (20.0) 1,216 (17.6) 15.6 (14.8, 16.4) 2020 6,478 (18.7) 764 (12.8) 11.8 (11.0, 12.6) 722 (17.3) 154 (16.1) 21.3 (18.3, 24.3) 7,200 (18.6) 918 (13.3) 12.8 (12.0, 13.5) DD, mean (SD) -3.1 (14.4) -3.4 (12.9) -0.8 (11.8) -0.5 (14.8) -2.9 (14.1) -3.0 (13.2) No. of episodes 34,661 5,940 17.1 (16.7, 17.5) 4,204 961 22.9 (21.6, 24.1) 38,865 6,901 17.8 (17.4, 18.1) Abbreviations: UEBMI, Urban Employment Basic Medical Insurance; URBMI, Urban Residents Basic Medical Insurance; NCMS, New Cooperative Medical Scheme; ED, emergency department; OPD, outpatient department; CCI, Charlson comorbidity index; UR, urbanization rate; DD, the differential distance, was calculated as the distance from the current address to the nearest public hospital minus the distance from the current address to the nearest private hospital. a PLVE = low-value episodes/qualified episodes. See eTable 1 for the definitions of qualified episodes and low-value episodes. b The comparison of every variable is between public and private hospitals; in addition, means (SDs) were used for continuous variables, and numbers of episodes (percentages) were used for classification variables. Association identification A partial F statistic exceeding 10 is suggestive of a strong IV (Table 2 ), verifying the relevance assumption. The following two tests suggest that the IV is not associated with the outcome other than through ownership, verifying the exclusion restriction. Initially, absolute standardized differences in means were all less than 0.25, suggesting no statistically significant differences between comparison groups ( eTable 3 in the Supplementary Materials). Besides, for inpatients who visit hospitals far away from their current residence, there is an insignificant correlation between the IV and hospital ownership ( eTable 4 in the Supplementary Materials). Both show that the IV is not correlated with those observed factors that affect the second-stage error term, suggesting no correlation between the IV and the second-stage error term. Table 2 presents the average incremental effects of the classical model and the IV model. After adjusted by our IV, private hospitals are more likely to provide low-value hysterectomies, with a risk difference of 33.70% (95% CI, 23.46–42.47%). Moreover, the choice of hospital ownership is endogenous, and then the IV approach is necessary. Table 2 Average incremental effects of the classical logistic model and the IV model a Classical logistic model b IV model c Average incremental effect (95% CI) a , % 0.40 (-1.22, 1.83) 33.70 (23.46, 42.47) Partial F statistics NA 169.21 Endogeneity test (95% CI) d -0.10 (-0.11, -0.08) Abbreviations: NA, not applicable. a Average incremental effect was calculated as the average predicted probability of receiving low-value care treated by private hospitals minus the predicted probability of receiving low-value care treated by public hospitals for the qualified episodes. b Models were adjusted for patient characteristics, hospital characteristics, local area economic characteristics, and year. c The predicted probability of choosing private hospitals, which was incorporated in the second stage logistic regression along with covariates, was calculated via logistic regression modeling of hospital ownership indicator as a function of provision of low-value care, where DD served as an instrument. d Refers to 100 times correlation coefficient value between hospital ownership and residual. Attributable excessive risk Table 3 shows that some of the excess risk of low-value hysterectomies is indeed attributable to private hospitals through three metrics. Population attributable fraction is 9.73% (95% CI, 8.70–10.40%), which means 9.73% low-value episodes are due to excessive risks embedded in private hospitals in comparison to public hospitals, corresponding to 48,375 (95% CI, 43,254, to 51,706) low-value episodes, thereby bringing a significant economic burden (1.82 billion USD [95% CI, 1.63 to 1.94]) nationwide annually. Table 3 Attributable excessive risk between private and public hospitals IV model a Population attributable fraction (95% eCI) b , % 9.73 (8.70, 10.40) Annual national attributable visits (95% eCI) c 48,375 (43,254, 51,706) Annual national total attributable costs (95% eCI) d , 2022 USD (billions) 1.82 (1.63, 1.94) a The predicted probability of choosing private hospitals, which was incorporated in the second stage logistic regression along with covariates, was calculated via logistic regression modeling of hospital ownership indicator as a function of provision of low-value care, where DD served as an instrument. b Population attributable fraction was calculated as the amount of low-value episodes attributable to hospital ownership divided by total number of low-value episodes, which is the estimated fraction of all qualified patients that would not have received low-value care if they had not gone to the private hospitals. c Annul national attributable visits was calculated as population attributable fraction times the amount of national low-value episodes per year. d Annual national total attributable costs were national costs of receiving low-value services due to visiting private hospitals, the sum of total costs of annul national attributable visits. Secondary analysis Secondary analyses were generally consistent with our primary results ( eAppendix 6–8 in the Supplementary Materials). Based on public hospitals as the control group, private for-profit hospitals showed greater differences in AIE and indicators related to attributable excess risk than private not-for-profit hospitals ( eTable 14–17 in the Supplementary Materials). Discussion In this retrospective analysis with a sample size of over 30 thousand, compared to public hospitals, private hospitals, regardless of whether not-for-profit or for-profit in nature were more likely to provide low-value hysterectomies. By eliminating the excess risks associated with private hospitals, we can anticipate a noteworthy reduction of 9.73% in the amount of low-value hysterectomy episodes, corresponding to 48,375 fewer low-value episodes and a total social burden of 1.82 billion USD costs nationwide annually. Among the above attributable low-value episodes, there is a potential for 1,793 more cases to face the risk of major surgical complications compared to using the alternative procedure ( eTable 5 ). Apart from the clinical burden, we have further computed the economic burden (1.82 billion USD) associated with these attributed cases. Compared to the alternative procedure (laparoscopic hysterectomy), low-value hysterectomy (abdominal hysterectomy for benign gynecological disease) incurs higher social costs and exhibits poorer clinical performance, resulting in poor cost-effectiveness, consistent with the conclusions of the majority of previous literature. 23 , 24 In both international low-value care catalogs and the existing domestic method for identifying low-value care, the definition and identification of low-value hysterectomy procedures primarily rely on their suboptimal clinical outcomes. Our research undoubtedly expands the economic evidence for its definition, aligning with the mainstream research on adding cost-effectiveness to define low-value care. 25 Public hospitals were more inclined to provide high-value hysterectomies (laparoscopic hysterectomies), which is consistent with previous research. 26 Underlying factors that explain why private hospitals were more likely to provide low-value hysterectomies (abdominal hysterectomies) are not clear. However, we speculate that the unevenly distributed healthcare professionals, the different patient demands, and the profit motives between the two types of hospitals could partly explain the disparities in the provision of low-value hysterectomies. First, despite the swift proliferation of private hospitals in the past decade, the progress of these newly established medical facilities is still at an embryonic stage of development. This phase is characterized by inadequate human resource management systems that struggle to ensure the availability of sufficient experienced and skilled medical professionals. Meanwhile, limited training, technical difficulty, and lack of personal surgical experience as substantial barriers to performing minimally invasive hysterectomy (high-value hysterectomy). 27 Therefore, in China, private hospitals might not be able to offer higher-value minimally invasive hysterectomy like public hospitals do. Instead, they might have to resort to offering lower-value abdominal hysterectomy. Second, private hospitals, to retain patients in a fiercely competitive market, may tend to offer services that have lower short-term economic costs for patients. 28 Public hospitals, in contrast, emphasize patient safety and long-term affordability, and prefer to provide laparoscopic hysterectomy that offer comprehensive long-term benefits. 28 , 29 Third, hysterectomy might serve as a way for hospitals to generate profits 30 and private for-profit hospitals may have more motivations to provide low-value hysterectomy with lower consumable costs to maximize their profits. Additionally, public hospitals are required to consider the public welfare, adhere to stricter policy regulations, and comply with more rigorous oversight. Furthermore, in China, private non-profit hospitals that fall between these two categories are still on their way towards profit-oriented goals, 31 but they also bear certain social responsibilities like public hospitals. Consequently, their performance in providing low-value hysterectomies falls somewhere between the above two types ( eTable 14 ). We found that private hospitals may be more inclined to provide low-value care, suggested hypotheses about possible mechanisms, and further calculated the attributable burdens, which can inform low-value care reduction efforts more broadly. Initially, the establishment of a robust human resource management system is imperative to ensure an ample and well-qualified workforce of healthcare professionals. This necessitates increased investment in healthcare personnel, accompanied by efforts to elevate the overall technical proficiency of physicians through a hybrid approach of online and offline training methods. Secondly, there is a critical need to bolster the cultivation of medical ethics and professionalism to elevate the moral and ethical standards of healthcare practitioners. Furthermore, ensuring seamless information exchange, particularly between public and private hospitals, is essential and requires the strengthening of technological collaboration. Last but not least, the utilization of low-value care hinges on the collaborative efforts of both physicians and patients. 32 Therefore, it is imperative to strive for transparency in the diagnostic and treatment processes and to facilitate enhanced communication between these two pivotal stakeholders. Any action derived from these findings should be subject to meticulous deliberation to prevent the erosion of hospital and clinician goodwill. Our study has three strengths. First, we have made advancements in the method by incorporating causal inference techniques that mimic randomized controlled trials based on observational data. 33 , 34 Comparison with a majority of the retrospective studies 35 , 36 exploring differences in low-value provision based on correlation analysis, which is straightforward but suffers from bias by the impact of residual confounders, 37 our measure could yield more accurate effect adjustment. Second, we apply a means of evaluating the risk factor and quantifying the impact of preventive health strategies for the first time and validate the feasibility of utilizing indirect indicators to determine the harmful effects of risk factors on the provision of low-value care for assessing and comparing burden levels associated with different risk factors for the population. Third, the study provides a way to estimate the maximum effectiveness of potential interventions for low-value care, i.e., the avoidable cost waste (5.81 million USD). Our study had six limitations. First, the major threat to the external validity of our findings is our IV. The IV analysis results of this study apply specifically to statistically marginal patients 38 who were admitted (or not admitted) based on their proximity to a given hospital and such patients represent those with a borderline or uncertain need for a public or private hospital. The IV analysis failed to identify individual patients who are “marginal”, but we were able to identify certain characteristics associated with these patients ( eTable 3 in the Supplementary Materials). This suggests that the population of patients in this study with those characteristics was likely to be the primary population for proximity to medical care. Therefore, our findings, focused on inpatients, might not be generalizable to populations who benefit from public (private) hospitals, younger populations undergoing non-hospitalization, or populations outside China. Nevertheless, our study demonstrates the feasibility of using instrumental variables to study disparities in low-value care at the institutional level, in the absence of large-scale and long-term RCTs. Second, it is imperative to highlight the pivotal role of physicians as the primary arbiters of medical decision-making, which significantly influences the utilization of low-value care. 39 While our study presumes that the majority of physicians can render judicious decisions based on their patient's medical conditions, it is important to acknowledge the challenge unable to adjust in quantifying certain attributes, such as physician practice type and compensation structures. These intricacies may introduce certain limitations to our research. Third, it is crucial to recognize that the overall equipment and facilities available within a hospital can exert a considerable influence on whether the institution is compelled to provide low-value care. To mitigate this influence, we have incorporated hospital grade as a variable within our models. Fourth, we calculated IV based on the Euclidean distance, where the coordinates were obtained by querying Amap, similar to Google Maps in the United States. Since the annual road network data between 2016–2020 is not available, there is a discrepancy between our calculation and the actual situation, which is the limitation of most similar IV empirical research methods. 40 Fifth, our measures of low-value care may still be susceptible to measurement error due to the limited clinical information available in inpatient discharge records and the lack of gold standards of clinical appropriateness. Although the aforementioned limitations can lead to disparities between our study results and the empirical reality, these variances pose no significant implications for the credibility and validity of the findings. Sixth, due to the differences in the structure of medical insurance and the provision of low-value care in different regions, there are some biases in our estimated national excess costs, and data from other provinces need to be further included. In consequent studies, it is crucial to investigate other risk factors related to the provision of low-value care to identify appropriate targets for interventions. Furthermore, it is important to delve deeper into the drivers behind the provision of low-value care in private hospitals and to enable the implementation of timely, appropriate, individualized, and multi-component interventions and treatment measures. 41 , 42 These measures will help to alleviate the mounting pressures of low-value care on patients and healthcare institutions, enhance the efficacy of healthcare resource allocation, exert control over healthcare expenditure, elevate the caliber of healthcare services, and ultimately advance the implementation of value-based care. Methods Data sources and study population In the context of data availability, we assessed hospital discharge records from 2016 to 2020, the fourth quarter of each year (October 1 to December 31) with more than 22 million inpatient episodes contained in the dataset. The administrative dataset, provided by the Health Commission of Sichuan Province, documents every inpatient admission across diverse hospitals within Sichuan and encompasses valuable information, including hospital identifier, expenditures during hospitalization, inpatients’ demographic details (e.g., age, sex, and occupation), as well as clinical information. Clinical information recorded by the dataset includes procedures performed coded by the International Classification of Diseases, 9th Revision, Clinical Modification, Volume 3 (ICD-9-CM3), and diagnoses assigned coded by International Classification of Diseases, 10th Revision (ICD-10). In addition, we extracted hospital characteristics including ownership from hospital annual reports provided by Health Commission of Sichuan Province. We also extracted socioeconomic characteristics (e.g., GDP per capita) of the county where the hospitals were located from the statistical yearbook reports. 43 Three sets of datasets were merged using unique hospital identifiers and county identifiers. The large population and hospital system in Sichuan Province, as well as the larger proportionate share of private hospitals and degree of development of private hospitals is similar to the national average, 44 provide a sufficient and reliable research sample and corresponding variance to identify the influence of hospital ownership on low-value care. West China Fourth Hospital and West China School of Public Health, Sichuan University Ethics Committee approved this study (Grant number: Gwll2022064). All females aged 18 years or older who underwent hysterectomy for benign indications (based on all the hysterectomy episodes) were eligible for inclusion in the study (i.e., qualified episodes in eTable 1 in the Supplementary Materials). We subsequently excluded some samples to ensure the soundness and reliability of our data. Ultimately, a refined study population of 38,865 sample size was identified (Fig. 1 ). [Insert Fig. 1 about here] Measure of low-value hysterectomy We devised an operational definition (abdominal hysterectomy for benign indications) of low-value hysterectomy episodes based on specific metrics extracted from the inpatient discharge records, including age, sex, ICD-9-CM3, and ICD-10 and then ought the expertise of a gynecology physician and a health information manager to ensure the accuracy and validity of our operational definition. Specifically, female patients 18 years of age and older underwent abdominal hysterectomy, with no mention of malignant tumors of female reproductive organs, endometriosis, or female pelvic peritoneal adhesions. The comprehensive details of the operational definition can be found in eAppendix 1 in the Supplementary Materials. The identification approach can be applied to the Chinese population, as mentioned in our previous study 2 and it was developed based on Choosing Wisely lists 3 , 45 commonly used in foreign countries to identify low-value care. Association identification issue and solution Two problems might interfere with our association identification and consequently, bias the estimated effects: the reverse causality and the missing variables problem. In China, the cost of low-value hysterectomies (abdominal hysterectomies) is generally lower compared to high-value hysterectomies (laparoscopic and vaginal hysterectomies). This difference in cost may attract patients to visit a particular type of hospital (either public or private) that predominately prescribes low-value procedures, creating a potential issue of reverse causality. Additionally, Chinese patients with more severe illnesses often exhibit a stronger inclination to seek care in public hospitals due to the perceived higher healthcare quality offered by these hospitals. 46 Meanwhile, more severe patients are more likely to undergo low-value hysterectomy. 47 Thus, the case mix as a set of confounding variables should be thoroughly controlled for in the regression model. Regrettably, administrative datasets typically do not capture the entirety of such variables. To address potential unmeasured bias in association identification, we did instrumental variable analysis 40 , 48 with the difference between the distance from the patient’s current residence to the nearest public hospital and the nearest private hospital as an instrument. Patients often seek treatment at the nearest facility (with the best geographic accessibility) 49 , 50 to reduce the risk of disease progression and cost; 51 , 52 Thus, patients’ residences should be highly predictive in determining the type of initial hospital admission. While it has been a trend to use distance to the hospital as the instrumental variable for health services, 40 , 53 – 55 it was necessary to prudently examine its validity in our study. A valid instrumental variable should satisfy the relevance assumption (The IV must be correlated with the choice of hospital ownership.) and exclusion restriction (The IV cannot be associated with the provision of low-value hysterectomy except through its correlation with the choice of hospital ownership.). Statistical analysis To determine whether the IV approach is necessary, we calculated the Spearman correlation coefficient between the hospital ownership variable and the residuals in a traditional logistic regression model. We, then calculated whether the partial F statistic was below 10 to verify the relevance assumption 56 , 57 , checked whether standardized differences in means between two groups above or below the median of the IV was less than 0.25, 58 , 59 and calculated the correlation between the IV and the inpatients’ choice of hospital ownership among inpatients who sought care far away from their current residence, such as individuals who traveled and sought care off-site to examine the exclusion restriction ( see eAppendix 3 in the Supplementary Materials for test details). We subsequently used a two-stage predictor substitution IV model for association analysis 60 to estimate the provision of low-value care as a function of the predicted probability of choosing hospitals (because of differential difference) and covariates (see eAppendix 2 in the Supplementary Materials for the variables of covariates). To help with interpretation, we reported the average incremental effect which was computed as the difference in the average predicted probability of providing if every inpatient had gone to private or public hospitals, with its 95% confidence intervals estimated by bootstrapping. Specific formulas of the IV model were in eAppendix 4 in the Supplementary Materials. To quantify excessive burdens of low-value care use attributable to hospital ownership, we used three metrics, namely population attributable fraction (PAF), annul national attributable visits (ANAVs), and annual national total attributable costs (ANTACs) (see eAppendix 5 in the Supplementary Materials for formulas). PAF was calculated on the percentage of low-value care episodes received by private hospitals and RR which was transformed by the OR estimated by the IV model. Besides, ANAVs was obtained based on PAF and annual qualified hysterectomy episodes nationwide. Additionally, ANTACs, quantifying total costs of ANAVs, incorporated direct and indirect costs. The empirical confidence intervals (eCIs) of the three estimates were obtained from Monte Carlo simulations, assuming a multivariate normal distribution of the coefficients obtained from the IV model. We performed statistical analyses using R software (version 4.2.0, developed by R Core Team) (Foundation for Statistical Computing, Vienna, Austria), mainly packages ivtool . Additionally, we used ArcGIS software (version 10.7 developed by Esri, Redlands, USA, authorization number: EFL734321752) to filter the address coordinates. Secondary analysis We conducted a series of secondary analyses. To ascertain whether there was a profit-driven influence in the disparities of low-value care between public and private hospitals to explore mechanisms, we redefined the exposure variable (public v.s. non-profit private v.s. profit private). We then redefined the measure of low-value hysterectomy to capture a broader range of clinical variations ( eAppendix 1 in the Supplementary Materials), which facilitated the resolution of the trade-off between sensitivity and specificity inherent in measuring low-value hysterectomy. As the reimbursement ability and the surgery cost may have a potential impact on the results, we incorporated the variable of medical insurance reimbursement (yes or no) into models to assess the robustness of our IV model. Declarations Funding The National Natural Science Foundation of China (Grant No. 72074163 and 72374149), Sichuan Science and Technology Program (Grant No. 2022YFS0052). Competing interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a qualifying conflict of interest. Author contributions T. Lan conceptualized the study. T. Lan designed the study and developed the analysis method. J. Pan collected the data. H. Luo and T. Lan implemented the data analysis. H. Luo drafted the manuscript. P. Coyte and K. Ju suggested revisions to the manuscript. H. Luo, T. Lan, and J. Pan revised the manuscript. All authors have agreed on the journal to which the article will be submitted gave final approval of the version to be published, and agree to be accountable for all aspects of the work. Availability of data and material The data that support the findings of this study are not publicly available due to privacy or ethical restrictions. They are part of the health administrative datasets of Sichuan, China. Code availability Available upon request. Ethics approval and consent to participate As the discharge data didn't involve any experimentation or human subjects, IRB approval, and patient consent were not required for this study. Consent for publication Not applicable Acknowledgments We would like to thank Shengtao Zhou at Western China Women's and Children's Hospital for his valuable comment on the hysterectomy, thank Doctor. Yuyan Liu at Sichuan University for her valuable comment on the instrumental variable, and thank Doctor. Qingping Xue, Xiaoxing Zhang, Qingyu Wang, Jianjian Wang, Kan Wu, Yuxin Yang, Junlong Li, Xiao Liu, Xuelian Hai, Chiyu Li, Chaohui Wang, Jia Zhang, Lu Ao, Qiang Yao, Ningxuan Zhao, and Yumeng Xie at Sichuan University for their valuable comments on the writing of this article. References Oakes, A. H. & Radomski, T. R. Reducing Low-Value Care and Improving Health Care Value. JAMA 325 , 1715-1716 (2021). Lan, T. et al. Measuring low-value care in hospital discharge records: evidence from China. The Lancet Regional Health–Western Pacific 38, 100887 (2023). Foundation, A. Choosing Wisely. https://www.choosingwisely.org/. Canada, C. W. Choosing wisely Canada. https://choosingwiselycanada.org/. Shrank, W. H., Rogstad, T. L. & Parekh, N. Waste in the US Health Care System: Estimated Costs and Potential for Savings. JAMA 322 , 1501-1509 (2019). Colla, C. H., Morden, N. E., Sequist, T. D., Schpero, W. L. & Rosenthal, M. B. Choosing wisely: prevalence and correlates of low-value health care services in the United States. J. Gen. Intern. Med. 30 , 221-228 (2015). Erard, Y. et al. A multi-level strategy for a long lasting reduction in unnecessary laboratory testing: A multicenter before and after study in a teaching hospital network. Int. J. Clin. Pract. 73 , e13286 (2018). Nilsen, P., Ingvarsson, S., Hasson, H., von Thiele Schwarz, U. & Augustsson, H. Theories, models, and frameworks for de-implementation of low-value care: A scoping review of the literature. Implement. Res. Pract. 1 . https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9978702/ (2020). Alishahi Tabriz, A. et al. De-implementing low-value care in cancer care delivery: a systematic review. Implement. Sci. 17 , 24 (2022). Parker, G. et al. Using theories and frameworks to understand how to reduce low-value healthcare: a scoping review. Implement. Sci. 17 , 6 (2022). Badgery-Parker, T. et al. Exploring variation in low-value care: a multilevel modelling study. BMC. Health. Serv. Res. 19 , 345 (2019). Shin, E. et al. Assessing patient, physician, and practice characteristics predicting the use of low-value services. Health. Serv. Res. 57, 1261-1273 (2022). Badgery-Parker, T., Pearson, S. A. & Elshaug, A. G. Hospital characteristics associated with low-value care in public hospitals in New South Wales, Australia. BMC. Health. Serv. Res. 20 , 750 (2020). Weeks, W. B., Kotzbauer, G. R. & Weinstein, J. N. Using Publicly Available Data to Construct a Transparent Measure of Health Care Value: A Method and Initial Results. Milbank. Q. 94 , 314-333 (2016). Oosting, R., Wauben, L., Groen, R., Dankelman, J. J. H. & Technology. Equipment for essential surgical care in 9 countries across Africa: availability, barriers and need for novel design. Health Technol 9 , 269-275 (2019). Hutchinson, P. L., Do, M. & Agha, S. J. B. h. s. r. Measuring client satisfaction and the quality of family planning services: a comparative analysis of public and private health facilities in Tanzania, Kenya and Ghana. BMC. Health. Serv. Res. 11 , 203 (2011). Abdelhafiz, I. M., Alloubani, A. M. D. & Almatari, M. J. J. o. n. m. Impact of leadership styles adopted by head nurses on job satisfaction: a comparative study between governmental and private hospitals in Jordan. J. Nurs. Manag. 24 , 384-392 (2016). Chabrillac, E. et al. The impact of physician's characteristics on decision-making inhead and neck oncology: Results of a national survey. Oral. Oncol. 129 , 105895 (2022). Ortiz-Prado, E. et al. Cesarean section rates in Ecuador: a 13-year comparative analysis between public and private health systems. Rev. Panam. Salud. Publica. 41 , e15 (2018). Chalmers, K. et al. Developing indicators for measuring low-value care: mapping Choosing Wisely recommendations to hospital data. BMC. Res. Notes. 11 , 163 (2018). Aarts, J. W. M. et al. Surgical approach to hysterectomy for benign gynaecological disease. Cochrane. Database. Syst. Rev. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858. CD003677.pub5/full (2015). Chalmers, K. et al. Measuring 21 low-value hospital procedures: claims analysis of Australian private health insurance data (2010-2014). BMJ Open 9 , e024142 (2019). Barnett, J. C. et al. Cost Comparison Among Robotic, Laparoscopic, and Open Hysterectomy for Endometrial Cancer. Obstet. Gynecol. 116 , 685-693 (2010). Wright, K. N., Jonsdottir, G. M., Jorgensen, S., Shah, N. & Einarsson, J. I. Costs and outcomes of abdominal, vaginal, laparoscopic and robotic hysterectomies. JSLS 16 , 519-524 (2012). Pandya, A. Adding Cost-effectiveness to Define Low-Value Care. JAMA 319 , 1977-1978 (2018). Chevrot, A. et al. [Hysterectomy: Practices evolution between 2009 and 2019 in France]. Gynecol. Obstet. Fertil. Senol. 49 , 816-822 (2021). Einarsson, J. I., Matteson, K. A., Schulkin, J., Chavan, N. R. & Sangi-Haghpeykar, H. Minimally invasive hysterectomies-a survey on attitudes and barriers among practicing gynecologists. J. Minim. Invasive. Gynecol. 17 , 167-175 (2010). Yang, J. et al. Cost effectiveness analysis of total laparoscopic hysterectomy versus total abdominal hysterectomy for uterine fibroids in Western China: a societal perspective. BMC. Health. Serv. Res. 22 , 252 (2022). Graves, N., Janda, M., Merollini, K., Gebski, V. & Obermair, A. The cost-effectiveness of total laparoscopic hysterectomy compared to total abdominal hysterectomy for the treatment of early stage endometrial cancer. BMJ Open 3 , e001884 (2013). Desai, S. Pragmatic prevention, permanent solution: Women's experiences with hysterectomy in rural India. Soc. Sci. Med. 151 , 11-18 (2016). Xie, Y., Liang, D., Huang, J. & Jin, J. Hospital Ownership and Hospital Institutional Change: A Qualitative Study in Guizhou Province, China. Int. J. Environ. Res. Public. Health. 16 , 1460 (2019). Sypes, E. E. et al. Understanding the public's role in reducing low-value care: a scoping review. Implement. Sci. 15 , 20 (2020). Robins, J. M., Hernán, M. A. & Brumback, B. Marginal structural models and causal inference in epidemiology. Epidemiology 11 , 550-560 (2000). Cole, S. R. & Hernán, M. A. Constructing inverse probability weights for marginal structural models. Am. J. Epidemiol. 168 , 656-664 (2008). de Oliveira Costa, J. et al. Rates of Low-Value Service in Australian Public Hospitals and the Association With Patient Insurance Status. JAMA Netw. Open 4 , e2138543 (2021). Kini, V. et al. Differences in High- and Low-Value Cardiovascular Testing by Health Insurance Provider. J. Am. Heart. Assoc. 10 , e018877 (2021). Robinson, L. D. & Jewell, N. P. Some surprising results about covariate adjustment in logistic regression models. Int. stat. rev. 59 , 227-240 (1991). Harris, K. M. & Remler, D. K. Who is the marginal patient? Understanding instrumental variables estimates of treatment effects. Health. Serv. Res. 33 , 1337-1360 (1998). Schwartz, A. L., Jena, A. B., Zaslavsky, A. M. & McWilliams, J. M. Analysis of Physician Variation in Provision of Low-Value Services. JAMA Intern. Med 179 , 16-25 (2019). Grabowski, D. C., Feng, Z., Hirth, R., Rahman, M. & Mor, V. Effect of nursing home ownership on the quality of post-acute care: an instrumental variables approach. J. Health. Econ. 32 , 12-21 (2013). Pitt, S. C. & Dossett, L. A. Deimplementation of Low-Value Care in Surgery. JAMA Surg. 157 , 977-978 (2022). Kjelle, E., Andersen, E. R., Soril, L. J. J., van Bodegom-Vos, L. & Hofmann, B. M. Interventions to reduce low-value imaging - a systematic review of interventions and outcomes. BMC. Health. Serv. Res. 21 , 983 (2021). Sichuan Provincial Bureau of Statistics. Statistical Yearbook Reports in Sichuan Province (2016-2020). Health Commission of Sichuan Province. Health Statistical Yearbook in Sichuan Province. (2021). University of Toronto & Association, C. M. Choosing Wisely Canada. https://choosingwiselycanada.org/recommendation/obstetrics-and-gynaecology/. Tang, C., Xu, J. & Zhang, M. The choice and preference for public-private health care among urban residents in China: evidence from a discrete choice experiment. BMC. Health. Serv. Res. 16 , 580 (2016). Ma, X., liang, Z., Xiang, Y., Zhang, G. & Zhang, S. Chinese expert consensus on surgical approaches of hysterectomy for benign uterine disease (2021 edition). Chinese Journal of Practical Gynecology and Obstetrics 37, 821-825 (2021). Baiocchi, M., Cheng, J. & Small, D. S. Instrumental variable methods for causal inference. Stat Med 33 , 2297-2340 (2014). Khan-Gates, J. A., Ersek, J. L., Eberth, J. M., Adams, S. A. & Pruitt, S. L. Geographic Access to Mammography and Its Relationship to Breast Cancer Screening and Stage at Diagnosis: A Systematic Review. Womens Health Issues 25 , 482-493 (2015). Nesbitt, R. C. et al . Methods to measure potential spatial access to delivery care in low- and middle-income countries: a case study in rural Ghana. Int. J. Health Geogr. 13 , 25 (2014). Chou, S., Deily, M. E. & Li, S. Travel distance and health outcomes for scheduled surgery. Med Care 52 , 250-257 (2014). Gertler, P., Locay, L. & Sanderson, W. Are user fees regressive?: The welfare implications of health care financing proposals in Peru. J. Econom. 36, 67-88 (1987). Norton, E. C. & Staiger, D. O. How hospital ownership affects access to care for the uninsured. Rand. J. Econ. 25 , 171-185 (1994). Gowrisankaran, G. & Town, R. J. Estimating the quality of care in hospitals using instrumental variables. J. Health. Econ. 18 , 747-767 (1999). Valley, T. S., Sjoding, M. W., Ryan, A. M., Iwashyna, T. J. & Cooke, C. R. Association of Intensive Care Unit Admission With Mortality Among Older Patients With Pneumonia. JAMA 4 , 1272-1279 (2015). Staiger, D. & Stock, J. H. Instrumental Variables Regression with Weak Instruments. Econometrica 65 , 557-586 (1997). Stock, J., & Yogo, M. Testing for Weak Instruments in Linear IV Regression. In D. Andrews & J. Stock (Eds.), Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg. https://www.nber.org/system/files/working_papers/t0284/t0284.pdf (2005). Rosenbaum, P. R. & Rubin, D. B. Constructing a Control Group Using Multivariate Matched Sampling Methods That Incorporate the Propensity Score. The American Statistician 39 , 33-38 (1985). Stuart, E. A. Matching methods for causal inference: A review and a look forward. Stat.Sci. 25 , 1-21 (2010). Wan, F., Small, D., Bekelman, J. E. & Mitra, N. Bias in estimating the causal hazard ratio when using two-stage instrumental variable methods. Stat. Med. 34 , 2235-2265 (2015). Additional Declarations There is NO Competing Interest. Supplementary Files Supplyment.pdf Supplemental Online Content Cite Share Download PDF Status: Posted Version 1 posted 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-3639662","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":253638150,"identity":"502cfb2c-261c-4762-9ff3-a6550bcb4b95","order_by":0,"name":"Jay Pan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8klEQVRIiWNgGAWjYJACgwQGBmYgOgDhHiBeC1sC8VqggMeAOC3m7GcPFDyoYGDnn5Hz+cPPNgY5vhsJjJ8L8Gix7MlLMEg4w8AscSN3m2RvG4Ox5I0EZukZeLQYHMgxMEhsA/oFqIWZsY0hccONBDZmHnxazr8BavnHwCx/I+fxZ6CWesJaboBsaWBgBjIYpIFaEgwIabGcAbQl4ZgEs+GZZ2aSPeckDGeeedgsjU+LOX+OmeGPGptkuePJjz/8KLOR5zuefPAzXocBoxCIJZIZBBJAfAkgZmzAowGshfkBkLZj4D+AV+EoGAWjYBSMYAAALnxJYHxhl9kAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-9501-1535","institution":"Sichuan University","correspondingAuthor":true,"prefix":"","firstName":"Jay","middleName":"","lastName":"Pan","suffix":""},{"id":253638151,"identity":"dc4143d5-bcc1-416c-8e96-6aea063937a2","order_by":1,"name":"Huijia Luo","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Huijia","middleName":"","lastName":"Luo","suffix":""},{"id":253638152,"identity":"acef2615-3a13-4f4b-8822-69e4d441f909","order_by":2,"name":"Tianjiao Lan","email":"","orcid":"","institution":"West China School of Public Health and West China Fourth Hospital, Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Tianjiao","middleName":"","lastName":"Lan","suffix":""},{"id":253638153,"identity":"f3daa972-4175-4ff4-bff6-6563d5ee0835","order_by":3,"name":"Peter Coyte","email":"","orcid":"","institution":"University of Toronto","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Coyte","suffix":""},{"id":253638154,"identity":"825203d6-27db-4ff4-89db-40b453cf4125","order_by":4,"name":"Ke Ju","email":"","orcid":"","institution":"Climate, Air Quality Research Unit, School of Public Health and Preventive Medicine, Monash University","correspondingAuthor":false,"prefix":"","firstName":"Ke","middleName":"","lastName":"Ju","suffix":""}],"badges":[],"createdAt":"2023-11-20 13:52:32","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3639662/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3639662/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":47224722,"identity":"022d9cdc-73fc-47eb-a8d6-b29555cc3e5a","added_by":"auto","created_at":"2023-11-28 19:21:16","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":53582,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlowchart for dataset for study population\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-3639662/v1/d36728fc026832c41e510fa4.png"},{"id":49516965,"identity":"53008d63-07e4-4605-b775-1536952ce5bc","added_by":"auto","created_at":"2024-01-12 08:33:54","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":834737,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3639662/v1/e1769b08-b1be-402a-82d4-23323cc3f752.pdf"},{"id":47224724,"identity":"26a369be-d4ac-4558-aee6-6c1117757ee4","added_by":"auto","created_at":"2023-11-28 19:21:17","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1212791,"visible":true,"origin":"","legend":"\u003cp\u003eSupplemental Online Content\u003c/p\u003e","description":"","filename":"Supplyment.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3639662/v1/ebde3f112d5f801e4af7eff0.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"The burdens of low-value care in hysterectomy attributable to hospital ownership in China","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLow-value care describes services where no evidence of patient benefit exists or where the harms or costs of care outweigh potential benefits.\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e However, despite the limited experience of measuring such services in low- to middle-income countries like China,\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e based on empirical evidence, these services, albeit necessitating context-specific measurement methodologies, can maintain standardized definitions for the evaluation of the extent of specific inappropriate healthcare services\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. In the past few decades, several initiatives have focused on reducing the use of low-value care. Despite these efforts, the prevalence of low-value care and its associated spending remains high.\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eRecently, studies have focused on de-implementation of low-value care. The majority of these studies proposed effective interventions targeting specific service providers (e.g., teaching hospitals\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e) by identifying drivers of low-value care delivery, using either quantitative or qualitative methods.\u003csup\u003e\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e However, these studies often lacked ex-ante estimates of the maximum effectiveness of potential interventions targeting such subjects. Therefore, it becomes challenging to pre-assess the potential value of implementing interventions on a specific type of provider. Conversely, the availability of such ex-ante estimates would enable researchers to conduct a cost-effectiveness analysis before implementing interventions, as they would possess a general understanding of the anticipated costs associated with the interventions.\u003c/p\u003e \u003cp\u003ePrior studies appealed for interventions at the hospital level to reduce low-value care provision.\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e Several studies have examined variations in low-value care across diverse hospital settings.\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e Nevertheless, there is still a lack of evidence about the association between hospital ownership and low-value care delivery. Differences between public and private hospitals may manifest in varying availability of appropriate equipment\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e (attributed to distinct funding sources, resource allocation, and business purpose), the experience levels of healthcare personnels\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e (due to divergent working conditions and career progression prospects), and physician's decision-making\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e (influenced by different patient volumes, patient reimbursement capacity, and regulatory frameworks). These discrepancies have the potential to lead to the delivery of inappropriate medical services, consequently reflecting variations in the provision of low-value care. To explore the above issue, we chose a representative low-value service, low-value hysterectomy, for illustration. This choice was primarily predicated on its well-established definition,\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e high prevalence,\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e and potential for intervention. The current definition relies heavily on clinical evidence, while there is an ongoing need for further enhancement and supplementation with economic evidence.\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e Based on hospital discharge records from a populated province with 83\u0026nbsp;million residents in China during 2016 and 2020, we employed the instrumental variable method (to address potential unmeasured bias in association identification) and took hospital ownership as a case study to estimate the attributable risk ratio and corresponding contribution degree associated with its influence on the provision of low-value hysterectomies.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n\u003ch2\u003eStudy sample characteristics\u003c/h2\u003e\n\u003cp\u003eWe identified 34,661 qualified episodes in public hospitals and 4,204 qualified episodes in private hospitals. Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e provides the characteristics associated with these qualified episodes. 343 public hospitals provide a total of 5,940 (17.1%; 95% CI, 16.7\u0026ndash;17.5%) low-value hysterectomy episodes, and 210 private hospitals provide a total of 961 (22.9%; 95% CI, 21.6\u0026ndash;24.1%) low-value hysterectomy episodes. The distribution of covariates is not balanced between episodes in public and private hospitals, for example, in employed inpatients or inpatients with CCI over 0, rates of low-value hysterectomy episodes in private hospitals are higher than public ones. This implies that the endogeneity problem indeed exists. Besides, the average differential distance in public hospitals is 3.1 kilometers, which suggests the patient in public hospitals has to travel 3.1 additional kilometers to the nearest private hospitals relative to the nearest public hospital, supporting the association between our IV and patients\u0026rsquo; choice of hospital.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab1\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eDescriptive statistics\u003csup\u003ea, b\u003c/sup\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth rowspan=\"2\" align=\"left\"\u003e\n\u003cp\u003eCharacteristics\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"3\" align=\"left\"\u003e\n\u003cp\u003ePublic Hospital\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"3\" align=\"left\"\u003e\n\u003cp\u003ePrivate Hospital\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"3\" align=\"left\"\u003e\n\u003cp\u003eOverall\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eQualified episodes\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eLow-value episodes\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ePLVE\u003c/p\u003e\n\u003cp\u003e(95% CI), %\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eQualified episodes\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eLow-value episodes\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ePLVE\u003c/p\u003e\n\u003cp\u003e(95% CI), %\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eQualified episodes\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eLow-value episodes\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ePLVE\u003c/p\u003e\n\u003cp\u003e(95% CI), %\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ePatient characteristics\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eAge, mean (SD)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e50 (8.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e50 (8.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e50 (8.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e50 (8.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e50 (8.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e50 (8.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eMinority, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eYes\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e703 (2.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e175 (2.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e24.9 (21.7, 28.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e83 (2.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e25 (2.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e30.1 (20.3, 40.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e786 (2.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e200 (2.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.5 (22.4, 28.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eNo\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e33,958 (98.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5,765 (97.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.0 (16.6, 17.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,121 (98.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e936 (97.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e22.7 (21.4, 24.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e38,079 (98.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,701 (97.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.6 (17.2, 18.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eMarital status, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eUnmarried\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,308 (3.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e189 (3.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.5 (12.4, 16.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e128 (3.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e28 (2.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e21.9 (14.7, 29.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,436 (3.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e217 (3.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e15.1 (13.3, 17.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eMarried\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e32,516 (93.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5,538 (93.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.0 (16.6, 17.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,001 (95.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e911 (94.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e22.8 (21.5, 24.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e36,517 (94.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6449 (93.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.7 (17.3, 18.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eOthers\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e837 (2.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e213 (3.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.5 (22.5, 28.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e75 (1.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e22 (2.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29.3 (19.0, 39.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e912 (2.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e235 (3.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.6 (22.9, 28.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eOccupation, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEmployed\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e16,899 (48.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,535 (59.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e20.9 (20.3, 21.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,697 (64.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e637 (66.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e23.6 (22.0, 25.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e19,596 (50.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,172 (60.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e21.3 (20.7, 21.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eRetired\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e724 (2.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e68 (1.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e9.4 (7.3, 11.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e92 (2.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e15 (1.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e16.3 (8.8, 23.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e816 (2.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e83 (1.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e10.2 (8.1, 12.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eOthers\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e17,038 (49.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,337 (39.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e13.7 (13.2, 14.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,415 (33.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e309 (32.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e21.8 (19.7, 24.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e18,453 (47.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,646 (38.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.3 (13.8, 14.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eInsurance type, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eUEBMI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,271 (18.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e735 (12.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e11.7 (10.9, 12.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e765 (18.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e111 (11.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.5 (12.0, 17.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7,036 (18.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e846 (12.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e12.0 (11.3, 12.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eURBMI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e12,403 (35.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,236 (37.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e18.0 (17.4, 18.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,686 (40.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e466 (48.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e27.6 (25.5, 29.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e14,089 (36.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,702 (39.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e19.2 (18.5, 19.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eNCMS\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e8,501 (24.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,127 (35.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.0 (24.1, 25.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,176 (28.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e259 (27.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e22.0 (19.7, 24.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e9,677 (24.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,386 (34.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e24.7 (23.8, 25.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eFully self-paid\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,661 (7.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e375 (6.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.1 (12.8, 15.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e343 (8.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e68 (7.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e19.8 (15.6, 24.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,004 (7.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e443 (6.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.8 (13.5, 16.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eOthers\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,825 (13.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e467 (7.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e9.7 (8.8, 10.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e234 (5.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e57 (5.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e24.4 (18.9, 29.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5,059 (13.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e524 (7.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e10.4 (9.5, 11.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eHospital stay, mean (SD)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e10 (4.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e11 (4.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e11 (7.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e11 (4.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e10 (4.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e11 (4.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eAdmission pathway, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eED\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,659 (7.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e586 (9.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e22.0 (20.5, 23.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e354 (8.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e31 (3.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e8.8 (5.8, 11.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,013 (7.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e617 (8.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e20.5 (19.0, 21.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eOPD\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e30,936 (89.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5,221 (87.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e16.9 (16.5, 17.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,790 (90.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e927 (96.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e24.5 (23.1, 25.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e34,726 (89.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,148 (89.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.7 (17.3, 18.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eOthers\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,066 (3.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e133 (2.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e12.5 (10.5, 14.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e60 (1.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3 (0.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e5 (-0.5, 10.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,126 (2.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e136 (2.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e12.1 (10.2, 14.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eCCI, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e34,223 (98.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5,818 (97.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.0 (16.6, 17.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,166 (99.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e942 (98.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e22.6 (21.3, 23.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e38,389 (98.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,760 (98.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.6 (17.2, 18.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u0026gt;\u0026thinsp;0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e438 (1.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e122 (2.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e27.9 (23.7, 32.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e38 (0.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e19 (2.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e50.0 (34.1, 65.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e476 (1.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e141 (2.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29.6 (25.5, 33.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eHospital characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eHospital level, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eTertiary\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e27,186 (78.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,950 (66.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.5 (14.1, 15.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e259 (6.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e12 (1.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e4.6 (2.1, 7.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e27,445 (70.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,962 (57.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.4 (14.0, 14.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSecondary\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7,446 (21.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,985 (33.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e26.7 (25.7, 27.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,536 (60.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e584 (60.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e23.0 (21.4, 24.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e9,982 (25.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,569 (37.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.7 (24.9, 26.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePrimary or\u003c/p\u003e\n\u003cp\u003eothers\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e29 (0.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5 (0.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.2 (3.5, 31.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,409 (33.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e365 (38.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.9 (23.6, 28.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,438 (3.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e370 (5.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.7 (23.5, 28.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eMedical services type, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eWestern\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e30,098 (86.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,715 (79.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e15.7 (15.3, 16.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,952 (94.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e851 (88.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e21.5 (20.3, 22.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e34,050 (87.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5,566 (80.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e16.4 (16.0, 16.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eTraditional Chinese\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,563 (13.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,225 (20.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e26.9 (25.6, 28.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e252 (6.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e110 (11.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e43.7 (37.5, 49.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,815 (12.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,335 (19.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e27.7 (26.5, 29.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eHospital location, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eRural\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e11,070 (31.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,862 (48.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.9 (25.0, 26.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,696 (40.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e510 (53.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e30.1 (27.9, 32.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e12,766 (32.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,372 (48.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e26.4 (25.7, 27.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eUrban\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e23,591 (68.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,078 (51.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e13.1 (12.6, 13.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2,508 (59.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e451 (46.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e18.0 (16.5, 19.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e26,099 (67.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3,529 (51.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e13.5 (13.1, 13.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eArea economic characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eGDP per capita, 1000 USD\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003emean (SD)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e9.30 (5.36)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7.10 (3.86)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e8.31 (4.74)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7.29 (4.19)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e9.19 (5.30)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7.13 (3.91)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eUR, mean (SD)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e63.0 (22.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e53.0 (17.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e56.5 (19.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e52.1 (17.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e62.3 (21.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e52.9 (17.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eYear, n (%)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2016\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7,378 (21.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,391 (23.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e18.9 (18.0, 19.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e720 (17.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e131 (13.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e18.2 (15.4, 21.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e8,098 (20.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,522 (22.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e18.8 (17.9, 19.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2017\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,626 (19.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,494 (25.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e22.6 (21.5, 23.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e985 (23.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e255 (26.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e25.9 (23.2, 28.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7,611 (19.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,749 (25.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e23.0 (22.0, 23.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2018\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7,204 (20.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,271 (21.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.6 (16.8, 18.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e960 (22.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e225 (23.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e23.4 (20.8, 26.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e8,164 (21.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,496 (21.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e18.3 (17.5, 19.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2019\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,975 (20.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,020 (17.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14.6 (13.8, 15.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e817 (19.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e196 (20.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e24.0 (21.1, 26.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7,792 (20.0)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1,216 (17.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e15.6 (14.8, 16.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2020\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,478 (18.7)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e764 (12.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e11.8 (11.0, 12.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e722 (17.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e154 (16.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e21.3 (18.3, 24.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e7,200 (18.6)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e918 (13.3)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e12.8 (12.0, 13.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eDD, mean (SD)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-3.1 (14.4)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-3.4 (12.9)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-0.8 (11.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-0.5 (14.8)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-2.9 (14.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-3.0 (13.2)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eNo. of episodes\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e34,661\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5,940\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.1 (16.7, 17.5)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4,204\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e961\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e22.9 (21.6, 24.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e38,865\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e6,901\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17.8 (17.4, 18.1)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"10\"\u003eAbbreviations: UEBMI, Urban Employment Basic Medical Insurance; URBMI, Urban Residents Basic Medical Insurance; NCMS, New Cooperative Medical Scheme; ED, emergency department; OPD, outpatient department; CCI, Charlson comorbidity index; UR, urbanization rate; DD, the differential distance, was calculated as the distance from the current address to the nearest public hospital minus the distance from the current address to the nearest private hospital.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"10\"\u003e\u003csup\u003ea\u003c/sup\u003e PLVE\u0026thinsp;=\u0026thinsp;low-value episodes/qualified episodes. See \u003cstrong\u003eeTable 1\u003c/strong\u003e for the definitions of qualified episodes and low-value episodes.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"10\"\u003e\u003csup\u003eb\u003c/sup\u003e The comparison of every variable is between public and private hospitals; in addition, means (SDs) were used for continuous variables, and numbers of episodes (percentages) were used for classification variables.\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\n\u003ch2\u003eAssociation identification\u003c/h2\u003e\n\u003cp\u003eA partial \u003cem\u003eF\u003c/em\u003e statistic exceeding 10 is suggestive of a strong IV (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e), verifying the relevance assumption. The following two tests suggest that the IV is not associated with the outcome other than through ownership, verifying the exclusion restriction. Initially, absolute standardized differences in means were all less than 0.25, suggesting no statistically significant differences between comparison groups (\u003cstrong\u003eeTable 3\u003c/strong\u003e in the Supplementary Materials). Besides, for inpatients who visit hospitals far away from their current residence, there is an insignificant correlation between the IV and hospital ownership (\u003cstrong\u003eeTable 4\u003c/strong\u003e in the Supplementary Materials). Both show that the IV is not correlated with those observed factors that affect the second-stage error term, suggesting no correlation between the IV and the second-stage error term.\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e presents the average incremental effects of the classical model and the IV model. After adjusted by our IV, private hospitals are more likely to provide low-value hysterectomies, with a risk difference of 33.70% (95% CI, 23.46\u0026ndash;42.47%). Moreover, the choice of hospital ownership is endogenous, and then the IV approach is necessary.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab2\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eAverage incremental effects of the classical logistic model and the IV model\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eClassical logistic model\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eIV model\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAverage incremental effect (95% CI)\u003csup\u003ea\u003c/sup\u003e, %\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0.40 (-1.22, 1.83)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e33.70 (23.46, 42.47)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePartial \u003cem\u003eF\u003c/em\u003e statistics\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eNA\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e169.21\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEndogeneity test (95% CI)\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd colspan=\"2\" align=\"left\"\u003e\n\u003cp\u003e-0.10 (-0.11, -0.08)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"3\"\u003eAbbreviations: NA, not applicable.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"3\"\u003e\u003csup\u003ea\u003c/sup\u003e Average incremental effect was calculated as the average predicted probability of receiving low-value care treated by private hospitals minus the predicted probability of receiving low-value care treated by public hospitals for the qualified episodes.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"3\"\u003e\u003csup\u003eb\u003c/sup\u003e Models were adjusted for patient characteristics, hospital characteristics, local area economic characteristics, and year.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"3\"\u003e\u003csup\u003ec\u003c/sup\u003e The predicted probability of choosing private hospitals, which was incorporated in the second stage logistic regression along with covariates, was calculated via logistic regression modeling of hospital ownership indicator as a function of provision of low-value care, where DD served as an instrument.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"3\"\u003e\u003csup\u003ed\u003c/sup\u003e Refers to 100 times correlation coefficient value between hospital ownership and residual.\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\n\u003ch2\u003eAttributable excessive risk\u003c/h2\u003e\n\u003cp\u003eTable\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e shows that some of the excess risk of low-value hysterectomies is indeed attributable to private hospitals through three metrics. Population attributable fraction is 9.73% (95% CI, 8.70\u0026ndash;10.40%), which means 9.73% low-value episodes are due to excessive risks embedded in private hospitals in comparison to public hospitals, corresponding to 48,375 (95% CI, 43,254, to 51,706) low-value episodes, thereby bringing a significant economic burden (1.82\u0026nbsp;billion USD [95% CI, 1.63 to 1.94]) nationwide annually.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab3\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eAttributable excessive risk between private and public hospitals\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eIV model\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePopulation attributable fraction (95% eCI)\u003csup\u003eb\u003c/sup\u003e, %\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e9.73 (8.70, 10.40)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAnnual national attributable visits (95% eCI)\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e48,375 (43,254, 51,706)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAnnual national total attributable costs (95% eCI)\u003csup\u003ed\u003c/sup\u003e, 2022 USD (billions)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1.82 (1.63, 1.94)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"2\"\u003e\u003csup\u003ea\u003c/sup\u003e The predicted probability of choosing private hospitals, which was incorporated in the second stage logistic regression along with covariates, was calculated via logistic regression modeling of hospital ownership indicator as a function of provision of low-value care, where DD served as an instrument.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"2\"\u003e\u003csup\u003eb\u003c/sup\u003e Population attributable fraction was calculated as the amount of low-value episodes attributable to hospital ownership divided by total number of low-value episodes, which is the estimated fraction of all qualified patients that would not have received low-value care if they had not gone to the private hospitals.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"2\"\u003e\u003csup\u003ec\u003c/sup\u003e Annul national attributable visits was calculated as population attributable fraction times the amount of national low-value episodes per year.\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"2\"\u003e\u003csup\u003ed\u003c/sup\u003e Annual national total attributable costs were national costs of receiving low-value services due to visiting private hospitals, the sum of total costs of annul national attributable visits.\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\n\u003ch2\u003eSecondary analysis\u003c/h2\u003e\n\u003cp\u003eSecondary analyses were generally consistent with our primary results (\u003cstrong\u003eeAppendix 6\u0026ndash;8\u003c/strong\u003e in the Supplementary Materials). Based on public hospitals as the control group, private for-profit hospitals showed greater differences in AIE and indicators related to attributable excess risk than private not-for-profit hospitals (\u003cstrong\u003eeTable 14\u0026ndash;17\u003c/strong\u003e in the Supplementary Materials).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this retrospective analysis with a sample size of over 30 thousand, compared to public hospitals, private hospitals, regardless of whether not-for-profit or for-profit in nature were more likely to provide low-value hysterectomies. By eliminating the excess risks associated with private hospitals, we can anticipate a noteworthy reduction of 9.73% in the amount of low-value hysterectomy episodes, corresponding to 48,375 fewer low-value episodes and a total social burden of 1.82\u0026nbsp;billion USD costs nationwide annually.\u003c/p\u003e \u003cp\u003eAmong the above attributable low-value episodes, there is a potential for 1,793 more cases to face the risk of major surgical complications compared to using the alternative procedure (\u003cb\u003eeTable 5\u003c/b\u003e). Apart from the clinical burden, we have further computed the economic burden (1.82\u0026nbsp;billion USD) associated with these attributed cases. Compared to the alternative procedure (laparoscopic hysterectomy), low-value hysterectomy (abdominal hysterectomy for benign gynecological disease) incurs higher social costs and exhibits poorer clinical performance, resulting in poor cost-effectiveness, consistent with the conclusions of the majority of previous literature.\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e In both international low-value care catalogs and the existing domestic method for identifying low-value care, the definition and identification of low-value hysterectomy procedures primarily rely on their suboptimal clinical outcomes. Our research undoubtedly expands the economic evidence for its definition, aligning with the mainstream research on adding cost-effectiveness to define low-value care.\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003ePublic hospitals were more inclined to provide high-value hysterectomies (laparoscopic hysterectomies), which is consistent with previous research.\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e Underlying factors that explain why private hospitals were more likely to provide low-value hysterectomies (abdominal hysterectomies) are not clear. However, we speculate that the unevenly distributed healthcare professionals, the different patient demands, and the profit motives between the two types of hospitals could partly explain the disparities in the provision of low-value hysterectomies. First, despite the swift proliferation of private hospitals in the past decade, the progress of these newly established medical facilities is still at an embryonic stage of development. This phase is characterized by inadequate human resource management systems that struggle to ensure the availability of sufficient experienced and skilled medical professionals. Meanwhile, limited training, technical difficulty, and lack of personal surgical experience as substantial barriers to performing minimally invasive hysterectomy (high-value hysterectomy).\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e Therefore, in China, private hospitals might not be able to offer higher-value minimally invasive hysterectomy like public hospitals do. Instead, they might have to resort to offering lower-value abdominal hysterectomy. Second, private hospitals, to retain patients in a fiercely competitive market, may tend to offer services that have lower short-term economic costs for patients.\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e Public hospitals, in contrast, emphasize patient safety and long-term affordability, and prefer to provide laparoscopic hysterectomy that offer comprehensive long-term benefits.\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e Third, hysterectomy might serve as a way for hospitals to generate profits\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e and private for-profit hospitals may have more motivations to provide low-value hysterectomy with lower consumable costs to maximize their profits. Additionally, public hospitals are required to consider the public welfare, adhere to stricter policy regulations, and comply with more rigorous oversight. Furthermore, in China, private non-profit hospitals that fall between these two categories are still on their way towards profit-oriented goals,\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e but they also bear certain social responsibilities like public hospitals. Consequently, their performance in providing low-value hysterectomies falls somewhere between the above two types (\u003cb\u003eeTable 14\u003c/b\u003e).\u003c/p\u003e \u003cp\u003e We found that private hospitals may be more inclined to provide low-value care, suggested hypotheses about possible mechanisms, and further calculated the attributable burdens, which can inform low-value care reduction efforts more broadly. Initially, the establishment of a robust human resource management system is imperative to ensure an ample and well-qualified workforce of healthcare professionals. This necessitates increased investment in healthcare personnel, accompanied by efforts to elevate the overall technical proficiency of physicians through a hybrid approach of online and offline training methods. Secondly, there is a critical need to bolster the cultivation of medical ethics and professionalism to elevate the moral and ethical standards of healthcare practitioners. Furthermore, ensuring seamless information exchange, particularly between public and private hospitals, is essential and requires the strengthening of technological collaboration. Last but not least, the utilization of low-value care hinges on the collaborative efforts of both physicians and patients.\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e Therefore, it is imperative to strive for transparency in the diagnostic and treatment processes and to facilitate enhanced communication between these two pivotal stakeholders. Any action derived from these findings should be subject to meticulous deliberation to prevent the erosion of hospital and clinician goodwill.\u003c/p\u003e \u003cp\u003eOur study has three strengths. First, we have made advancements in the method by incorporating causal inference techniques that mimic randomized controlled trials based on observational data.\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e Comparison with a majority of the retrospective studies\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e exploring differences in low-value provision based on correlation analysis, which is straightforward but suffers from bias by the impact of residual confounders,\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e our measure could yield more accurate effect adjustment. Second, we apply a means of evaluating the risk factor and quantifying the impact of preventive health strategies for the first time and validate the feasibility of utilizing indirect indicators to determine the harmful effects of risk factors on the provision of low-value care for assessing and comparing burden levels associated with different risk factors for the population. Third, the study provides a way to estimate the maximum effectiveness of potential interventions for low-value care, i.e., the avoidable cost waste (5.81\u0026nbsp;million USD).\u003c/p\u003e \u003cp\u003eOur study had six limitations. First, the major threat to the external validity of our findings is our IV. The IV analysis results of this study apply specifically to statistically marginal patients\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e who were admitted (or not admitted) based on their proximity to a given hospital and such patients represent those with a borderline or uncertain need for a public or private hospital. The IV analysis failed to identify individual patients who are \u0026ldquo;marginal\u0026rdquo;, but we were able to identify certain characteristics associated with these patients (\u003cb\u003eeTable 3\u003c/b\u003e in the Supplementary Materials). This suggests that the population of patients in this study with those characteristics was likely to be the primary population for proximity to medical care. Therefore, our findings, focused on inpatients, might not be generalizable to populations who benefit from public (private) hospitals, younger populations undergoing non-hospitalization, or populations outside China. Nevertheless, our study demonstrates the feasibility of using instrumental variables to study disparities in low-value care at the institutional level, in the absence of large-scale and long-term RCTs. Second, it is imperative to highlight the pivotal role of physicians as the primary arbiters of medical decision-making, which significantly influences the utilization of low-value care.\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e While our study presumes that the majority of physicians can render judicious decisions based on their patient's medical conditions, it is important to acknowledge the challenge unable to adjust in quantifying certain attributes, such as physician practice type and compensation structures. These intricacies may introduce certain limitations to our research. Third, it is crucial to recognize that the overall equipment and facilities available within a hospital can exert a considerable influence on whether the institution is compelled to provide low-value care. To mitigate this influence, we have incorporated hospital grade as a variable within our models. Fourth, we calculated IV based on the Euclidean distance, where the coordinates were obtained by querying Amap, similar to Google Maps in the United States. Since the annual road network data between 2016\u0026ndash;2020 is not available, there is a discrepancy between our calculation and the actual situation, which is the limitation of most similar IV empirical research methods.\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e Fifth, our measures of low-value care may still be susceptible to measurement error due to the limited clinical information available in inpatient discharge records and the lack of gold standards of clinical appropriateness. Although the aforementioned limitations can lead to disparities between our study results and the empirical reality, these variances pose no significant implications for the credibility and validity of the findings. Sixth, due to the differences in the structure of medical insurance and the provision of low-value care in different regions, there are some biases in our estimated national excess costs, and data from other provinces need to be further included.\u003c/p\u003e \u003cp\u003eIn consequent studies, it is crucial to investigate other risk factors related to the provision of low-value care to identify appropriate targets for interventions. Furthermore, it is important to delve deeper into the drivers behind the provision of low-value care in private hospitals and to enable the implementation of timely, appropriate, individualized, and multi-component interventions and treatment measures.\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e These measures will help to alleviate the mounting pressures of low-value care on patients and healthcare institutions, enhance the efficacy of healthcare resource allocation, exert control over healthcare expenditure, elevate the caliber of healthcare services, and ultimately advance the implementation of value-based care.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eData sources and study population\u003c/h2\u003e \u003cp\u003eIn the context of data availability, we assessed hospital discharge records from 2016 to 2020, the fourth quarter of each year (October 1 to December 31) with more than 22\u0026nbsp;million inpatient episodes contained in the dataset. The administrative dataset, provided by the Health Commission of Sichuan Province, documents every inpatient admission across diverse hospitals within Sichuan and encompasses valuable information, including hospital identifier, expenditures during hospitalization, inpatients\u0026rsquo; demographic details (e.g., age, sex, and occupation), as well as clinical information. Clinical information recorded by the dataset includes procedures performed coded by the International Classification of Diseases, 9th Revision, Clinical Modification, Volume 3 (ICD-9-CM3), and diagnoses assigned coded by International Classification of Diseases, 10th Revision (ICD-10). In addition, we extracted hospital characteristics including ownership from hospital annual reports provided by Health Commission of Sichuan Province. We also extracted socioeconomic characteristics (e.g., GDP per capita) of the county where the hospitals were located from the statistical yearbook reports.\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e Three sets of datasets were merged using unique hospital identifiers and county identifiers. The large population and hospital system in Sichuan Province, as well as the larger proportionate share of private hospitals and degree of development of private hospitals is similar to the national average,\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e provide a sufficient and reliable research sample and corresponding variance to identify the influence of hospital ownership on low-value care. West China Fourth Hospital and West China School of Public Health, Sichuan University Ethics Committee approved this study (Grant number: Gwll2022064).\u003c/p\u003e \u003cp\u003eAll females aged 18 years or older who underwent hysterectomy for benign indications (based on all the hysterectomy episodes) were eligible for inclusion in the study (i.e., qualified episodes in \u003cb\u003eeTable 1\u003c/b\u003e in the Supplementary Materials). We subsequently excluded some samples to ensure the soundness and reliability of our data. Ultimately, a refined study population of 38,865 sample size was identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e[Insert Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e about here]\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eMeasure of low-value hysterectomy\u003c/h2\u003e \u003cp\u003eWe devised an operational definition (abdominal hysterectomy for benign indications) of low-value hysterectomy episodes based on specific metrics extracted from the inpatient discharge records, including age, sex, ICD-9-CM3, and ICD-10 and then ought the expertise of a gynecology physician and a health information manager to ensure the accuracy and validity of our operational definition. Specifically, female patients 18 years of age and older underwent abdominal hysterectomy, with no mention of malignant tumors of female reproductive organs, endometriosis, or female pelvic peritoneal adhesions. The comprehensive details of the operational definition can be found in \u003cb\u003eeAppendix 1\u003c/b\u003e in the Supplementary Materials. The identification approach can be applied to the Chinese population, as mentioned in our previous study\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e and it was developed based on Choosing Wisely lists\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e commonly used in foreign countries to identify low-value care.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eAssociation identification issue and solution\u003c/h2\u003e \u003cp\u003eTwo problems might interfere with our association identification and consequently, bias the estimated effects: the reverse causality and the missing variables problem. In China, the cost of low-value hysterectomies (abdominal hysterectomies) is generally lower compared to high-value hysterectomies (laparoscopic and vaginal hysterectomies). This difference in cost may attract patients to visit a particular type of hospital (either public or private) that predominately prescribes low-value procedures, creating a potential issue of reverse causality. Additionally, Chinese patients with more severe illnesses often exhibit a stronger inclination to seek care in public hospitals due to the perceived higher healthcare quality offered by these hospitals.\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e Meanwhile, more severe patients are more likely to undergo low-value hysterectomy.\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e Thus, the case mix as a set of confounding variables should be thoroughly controlled for in the regression model. Regrettably, administrative datasets typically do not capture the entirety of such variables.\u003c/p\u003e \u003cp\u003eTo address potential unmeasured bias in association identification, we did instrumental variable analysis\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e with the difference between the distance from the patient\u0026rsquo;s current residence to the nearest public hospital and the nearest private hospital as an instrument. Patients often seek treatment at the nearest facility (with the best geographic accessibility)\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e,\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e to reduce the risk of disease progression and cost;\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e,\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e Thus, patients\u0026rsquo; residences should be highly predictive in determining the type of initial hospital admission. While it has been a trend to use distance to the hospital as the instrumental variable for health services,\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan additionalcitationids=\"CR54\" citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e it was necessary to prudently examine its validity in our study. A valid instrumental variable should satisfy the relevance assumption (The IV must be correlated with the choice of hospital ownership.) and exclusion restriction (The IV cannot be associated with the provision of low-value hysterectomy except through its correlation with the choice of hospital ownership.).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eTo determine whether the IV approach is necessary, we calculated the Spearman correlation coefficient between the hospital ownership variable and the residuals in a traditional logistic regression model. We, then calculated whether the partial \u003cem\u003eF\u003c/em\u003e statistic was below 10 to verify the relevance assumption\u003csup\u003e\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e,\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e, checked whether standardized differences in means between two groups above or below the median of the IV was less than 0.25,\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e,\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e and calculated the correlation between the IV and the inpatients\u0026rsquo; choice of hospital ownership among inpatients who sought care far away from their current residence, such as individuals who traveled and sought care off-site to examine the exclusion restriction \u003cb\u003e(\u003c/b\u003esee \u003cb\u003eeAppendix 3\u003c/b\u003e in the Supplementary Materials for test details).\u003c/p\u003e \u003cp\u003eWe subsequently used a two-stage predictor substitution IV model for association analysis\u003csup\u003e\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e to estimate the provision of low-value care as a function of the predicted probability of choosing hospitals (because of differential difference) and covariates (see \u003cb\u003eeAppendix 2\u003c/b\u003e in the Supplementary Materials for the variables of covariates). To help with interpretation, we reported the average incremental effect which was computed as the difference in the average predicted probability of providing if every inpatient had gone to private or public hospitals, with its 95% confidence intervals estimated by bootstrapping. Specific formulas of the IV model were in \u003cb\u003eeAppendix 4\u003c/b\u003e in the Supplementary Materials.\u003c/p\u003e \u003cp\u003eTo quantify excessive burdens of low-value care use attributable to hospital ownership, we used three metrics, namely population attributable fraction (PAF), annul national attributable visits (ANAVs), and annual national total attributable costs (ANTACs) (see \u003cb\u003eeAppendix 5\u003c/b\u003e in the Supplementary Materials for formulas). PAF was calculated on the percentage of low-value care episodes received by private hospitals and RR which was transformed by the OR estimated by the IV model. Besides, ANAVs was obtained based on PAF and annual qualified hysterectomy episodes nationwide. Additionally, ANTACs, quantifying total costs of ANAVs, incorporated direct and indirect costs. The empirical confidence intervals (eCIs) of the three estimates were obtained from Monte Carlo simulations, assuming a multivariate normal distribution of the coefficients obtained from the IV model.\u003c/p\u003e \u003cp\u003eWe performed statistical analyses using R software (version 4.2.0, developed by R Core Team) (Foundation for Statistical Computing, Vienna, Austria), mainly packages \u003cem\u003eivtool\u003c/em\u003e. Additionally, we used ArcGIS software (version 10.7 developed by Esri, Redlands, USA, authorization number: EFL734321752) to filter the address coordinates.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eSecondary analysis\u003c/h2\u003e \u003cp\u003eWe conducted a series of secondary analyses. To ascertain whether there was a profit-driven influence in the disparities of low-value care between public and private hospitals to explore mechanisms, we redefined the exposure variable (public v.s. non-profit private v.s. profit private). We then redefined the measure of low-value hysterectomy to capture a broader range of clinical variations (\u003cb\u003eeAppendix 1\u003c/b\u003e in the Supplementary Materials), which facilitated the resolution of the trade-off between sensitivity and specificity inherent in measuring low-value hysterectomy. As the reimbursement ability and the surgery cost may have a potential impact on the results, we incorporated the variable of medical insurance reimbursement (yes or no) into models to assess the robustness of our IV model.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe National Natural Science Foundation of China (Grant No. 72074163 and 72374149), Sichuan Science and Technology Program (Grant No. 2022YFS0052).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a qualifying conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eT. Lan conceptualized the study. T. Lan designed the study and developed the analysis method. J. Pan collected the data. H. Luo and T. Lan implemented the data analysis. H. Luo drafted the manuscript. P. Coyte and K. Ju suggested revisions to the manuscript. H. Luo, T. Lan, and J. Pan revised the manuscript. All authors have agreed on the journal to which the article will be submitted gave final approval of the version to be published, and agree to be accountable for all aspects of the work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are not publicly available due to privacy or ethical restrictions. They are part of the health administrative datasets of Sichuan, China.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAvailable upon request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs the discharge data didn\u0026apos;t involve any experimentation or human subjects, IRB approval, and patient consent were not required for this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank Shengtao Zhou at Western China Women\u0026apos;s and Children\u0026apos;s Hospital for his valuable comment on the hysterectomy, thank Doctor. Yuyan Liu at Sichuan University for her valuable comment on the instrumental variable, and thank Doctor. Qingping Xue, Xiaoxing Zhang, Qingyu Wang, Jianjian Wang, Kan Wu, Yuxin Yang, Junlong Li, Xiao Liu, Xuelian Hai, Chiyu Li, Chaohui Wang, Jia Zhang, Lu Ao, Qiang Yao, Ningxuan Zhao, and Yumeng Xie at Sichuan University for their valuable comments on the writing of this article.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eOakes, A. H. \u0026amp; Radomski, T. R. Reducing Low-Value Care and Improving Health Care Value. \u003cem\u003eJAMA\u003c/em\u003e \u003cstrong\u003e325\u003c/strong\u003e, 1715-1716 (2021).\u003c/li\u003e\n\u003cli\u003eLan, T.\u003cem\u003e \u003c/em\u003eet al. Measuring low-value care in hospital discharge records: evidence from China. \u003cem\u003eThe Lancet Regional Health\u0026ndash;Western Pacific\u003c/em\u003e \u003cstrong\u003e38,\u003c/strong\u003e 100887 (2023).\u003c/li\u003e\n\u003cli\u003eFoundation, A. Choosing Wisely. https://www.choosingwisely.org/.\u003c/li\u003e\n\u003cli\u003eCanada, C. W. Choosing wisely Canada. https://choosingwiselycanada.org/. \u003c/li\u003e\n\u003cli\u003eShrank, W. H., Rogstad, T. L. \u0026amp; Parekh, N. Waste in the US Health Care System: Estimated Costs and Potential for Savings. \u003cem\u003eJAMA\u003c/em\u003e\u003cstrong\u003e 322\u003c/strong\u003e, 1501-1509 (2019).\u003c/li\u003e\n\u003cli\u003eColla, C. H., Morden, N. E., Sequist, T. D., Schpero, W. L. \u0026amp; Rosenthal, M. B. Choosing wisely: prevalence and correlates of low-value health care services in the United States. \u003cem\u003eJ. Gen. Intern. Med.\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 221-228 (2015).\u003c/li\u003e\n\u003cli\u003eErard, Y. et al. A multi-level strategy for a long lasting reduction in unnecessary laboratory testing: A multicenter before and after study in a teaching hospital network. \u003cem\u003eInt. J. Clin. Pract.\u003c/em\u003e \u003cstrong\u003e73\u003c/strong\u003e, e13286 (2018).\u003c/li\u003e\n\u003cli\u003eNilsen, P., Ingvarsson, S., Hasson, H., von Thiele Schwarz, U. \u0026amp; Augustsson, H. Theories, models, and frameworks for de-implementation of low-value care: A scoping review of the literature. \u003cem\u003eImplement. Res. Pract.\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9978702/ (2020).\u003c/li\u003e\n\u003cli\u003eAlishahi Tabriz, A. et al. De-implementing low-value care in cancer care delivery: a systematic review. \u003cem\u003eImplement. Sci.\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 24 (2022).\u003c/li\u003e\n\u003cli\u003eParker, G. et al. Using theories and frameworks to understand how to reduce low-value healthcare: a scoping review. \u003cem\u003eImplement. Sci.\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 6 (2022).\u003c/li\u003e\n\u003cli\u003eBadgery-Parker, T. et al. Exploring variation in low-value care: a multilevel modelling study. \u003cem\u003eBMC. Health. Serv. Res.\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 345 (2019).\u003c/li\u003e\n\u003cli\u003eShin, E. et al. Assessing patient, physician, and practice characteristics predicting the use of low-value services. \u003cem\u003eHealth. Serv. Res. \u003c/em\u003e\u003cstrong\u003e57,\u003c/strong\u003e 1261-1273 (2022).\u003c/li\u003e\n\u003cli\u003eBadgery-Parker, T., Pearson, S. A. \u0026amp; Elshaug, A. G. Hospital characteristics associated with low-value care in public hospitals in New South Wales, Australia. \u003cem\u003eBMC. Health. Serv. Res.\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 750 (2020).\u003c/li\u003e\n\u003cli\u003eWeeks, W. B., Kotzbauer, G. R. \u0026amp; Weinstein, J. N. Using Publicly Available Data to Construct a Transparent Measure of Health Care Value: A Method and Initial Results. \u003cem\u003eMilbank. Q.\u003c/em\u003e \u003cstrong\u003e94\u003c/strong\u003e, 314-333 (2016).\u003c/li\u003e\n\u003cli\u003eOosting, R., Wauben, L., Groen, R., Dankelman, J. J. H. \u0026amp; Technology. Equipment for essential surgical care in 9 countries across Africa: availability, barriers and need for novel design. \u003cem\u003eHealth Technol\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 269-275 (2019).\u003c/li\u003e\n\u003cli\u003eHutchinson, P. L., Do, M. \u0026amp; Agha, S. J. B. h. s. r. Measuring client satisfaction and the quality of family planning services: a comparative analysis of public and private health facilities in Tanzania, Kenya and Ghana. \u003cem\u003eBMC. Health. Serv. Res.\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 203 (2011).\u003c/li\u003e\n\u003cli\u003eAbdelhafiz, I. M., Alloubani, A. M. D. \u0026amp; Almatari, M. J. J. o. n. m. Impact of leadership styles adopted by head nurses on job satisfaction: a comparative study between governmental and private hospitals in Jordan. \u003cem\u003eJ. Nurs. Manag.\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 384-392 (2016).\u003c/li\u003e\n\u003cli\u003eChabrillac, E. et al. The impact of physician\u0026apos;s characteristics on decision-making inhead and neck oncology: Results of a national survey. \u003cem\u003eOral. Oncol.\u003c/em\u003e \u003cstrong\u003e129\u003c/strong\u003e, 105895 (2022).\u003c/li\u003e\n\u003cli\u003eOrtiz-Prado, E. et al. Cesarean section rates in Ecuador: a 13-year comparative analysis between public and private health systems. \u003cem\u003eRev. Panam. Salud. Publica.\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, e15 (2018).\u003c/li\u003e\n\u003cli\u003eChalmers, K. et al. Developing indicators for measuring low-value care: mapping Choosing Wisely recommendations to hospital data. \u003cem\u003eBMC. Res. Notes.\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 163 (2018).\u003c/li\u003e\n\u003cli\u003eAarts, J. W. M. et al. Surgical approach to hysterectomy for benign gynaecological disease. \u003cem\u003eCochrane. Database. Syst. Rev.\u003c/em\u003e https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858. CD003677.pub5/full (2015).\u003c/li\u003e\n\u003cli\u003eChalmers, K. et al. Measuring 21 low-value hospital procedures: claims analysis of Australian private health insurance data (2010-2014). \u003cem\u003eBMJ Open\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, e024142 (2019).\u003c/li\u003e\n\u003cli\u003eBarnett, J. C. et al. Cost Comparison Among Robotic, Laparoscopic, and Open Hysterectomy for Endometrial Cancer. \u003cem\u003eObstet. Gynecol.\u003c/em\u003e \u003cstrong\u003e116\u003c/strong\u003e, 685-693 (2010).\u003c/li\u003e\n\u003cli\u003eWright, K. N., Jonsdottir, G. M., Jorgensen, S., Shah, N. \u0026amp; Einarsson, J. I. Costs and outcomes of abdominal, vaginal, laparoscopic and robotic hysterectomies. \u003cem\u003eJSLS \u003c/em\u003e\u003cstrong\u003e16\u003c/strong\u003e, 519-524 (2012).\u003c/li\u003e\n\u003cli\u003ePandya, A. Adding Cost-effectiveness to Define Low-Value Care. \u003cem\u003eJAMA\u003c/em\u003e \u003cstrong\u003e319\u003c/strong\u003e, 1977-1978 (2018).\u003c/li\u003e\n\u003cli\u003eChevrot, A. et al. [Hysterectomy: Practices evolution between 2009 and 2019 in France]. \u003cem\u003eGynecol. Obstet. Fertil. Senol.\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e, 816-822 (2021).\u003c/li\u003e\n\u003cli\u003eEinarsson, J. I., Matteson, K. A., Schulkin, J., Chavan, N. R. \u0026amp; Sangi-Haghpeykar, H. Minimally invasive hysterectomies-a survey on attitudes and barriers among practicing gynecologists. \u003cem\u003eJ. Minim. Invasive. Gynecol.\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 167-175 (2010).\u003c/li\u003e\n\u003cli\u003eYang, J. et al. Cost effectiveness analysis of total laparoscopic hysterectomy versus total abdominal hysterectomy for uterine fibroids in Western China: a societal perspective. \u003cem\u003eBMC. Health. Serv. Res.\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 252 (2022).\u003c/li\u003e\n\u003cli\u003eGraves, N., Janda, M., Merollini, K., Gebski, V. \u0026amp; Obermair, A. The cost-effectiveness of total laparoscopic hysterectomy compared to total abdominal hysterectomy for the treatment of early stage endometrial cancer. \u003cem\u003eBMJ Open\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, e001884 (2013).\u003c/li\u003e\n\u003cli\u003eDesai, S. Pragmatic prevention, permanent solution: Women\u0026apos;s experiences with hysterectomy in rural India. \u003cem\u003eSoc. Sci. Med.\u003c/em\u003e \u003cstrong\u003e151\u003c/strong\u003e, 11-18 (2016).\u003c/li\u003e\n\u003cli\u003eXie, Y., Liang, D., Huang, J. \u0026amp; Jin, J. Hospital Ownership and Hospital Institutional Change: A Qualitative Study in Guizhou Province, China. \u003cem\u003eInt. J. Environ. Res. Public. Health.\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 1460 (2019).\u003c/li\u003e\n\u003cli\u003eSypes, E. E. et al. Understanding the public\u0026apos;s role in reducing low-value care: a scoping review. \u003cem\u003eImplement. Sci.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 20 (2020).\u003c/li\u003e\n\u003cli\u003eRobins, J. M., Hern\u0026aacute;n, M. A. \u0026amp; Brumback, B. Marginal structural models and causal inference in epidemiology. \u003cem\u003eEpidemiology\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 550-560 (2000).\u003c/li\u003e\n\u003cli\u003eCole, S. R. \u0026amp; Hern\u0026aacute;n, M. A. Constructing inverse probability weights for marginal structural models. \u003cem\u003eAm. J. Epidemiol.\u003c/em\u003e \u003cstrong\u003e168\u003c/strong\u003e, 656-664 (2008).\u003c/li\u003e\n\u003cli\u003ede Oliveira Costa, J. et al. Rates of Low-Value Service in Australian Public Hospitals and the Association With Patient Insurance Status. \u003cem\u003eJAMA Netw. Open\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, e2138543 (2021).\u003c/li\u003e\n\u003cli\u003eKini, V. et al. Differences in High- and Low-Value Cardiovascular Testing by Health Insurance Provider. \u003cem\u003eJ. Am. Heart. Assoc.\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, e018877 (2021).\u003c/li\u003e\n\u003cli\u003eRobinson, L. D. \u0026amp; Jewell, N. P. Some surprising results about covariate adjustment in logistic regression models. \u003cem\u003eInt. stat. rev.\u003c/em\u003e \u003cstrong\u003e59\u003c/strong\u003e, 227-240 (1991).\u003c/li\u003e\n\u003cli\u003eHarris, K. M. \u0026amp; Remler, D. K. Who is the marginal patient? Understanding instrumental variables estimates of treatment effects. \u003cem\u003eHealth. Serv. Res.\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 1337-1360 (1998).\u003c/li\u003e\n\u003cli\u003eSchwartz, A. L., Jena, A. B., Zaslavsky, A. M. \u0026amp; McWilliams, J. M. Analysis of Physician Variation in Provision of Low-Value Services. \u003cem\u003eJAMA Intern. Med\u003c/em\u003e \u003cstrong\u003e179\u003c/strong\u003e, 16-25 (2019).\u003c/li\u003e\n\u003cli\u003eGrabowski, D. C., Feng, Z., Hirth, R., Rahman, M. \u0026amp; Mor, V. Effect of nursing home ownership on the quality of post-acute care: an instrumental variables approach. \u003cem\u003eJ. Health. Econ.\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 12-21 (2013).\u003c/li\u003e\n\u003cli\u003ePitt, S. C. \u0026amp; Dossett, L. A. Deimplementation of Low-Value Care in Surgery. \u003cem\u003eJAMA Surg.\u003c/em\u003e \u003cstrong\u003e157\u003c/strong\u003e, 977-978 (2022).\u003c/li\u003e\n\u003cli\u003eKjelle, E., Andersen, E. R., Soril, L. J. J., van Bodegom-Vos, L. \u0026amp; Hofmann, B. M. Interventions to reduce low-value imaging - a systematic review of interventions and outcomes. \u003cem\u003eBMC. Health. Serv. Res.\u003c/em\u003e\u003cem\u003e \u003c/em\u003e\u003cstrong\u003e21\u003c/strong\u003e, 983 (2021).\u003c/li\u003e\n\u003cli\u003eSichuan Provincial Bureau of Statistics. Statistical Yearbook Reports in Sichuan Province (2016-2020).\u003c/li\u003e\n\u003cli\u003eHealth Commission of Sichuan Province. Health Statistical Yearbook in Sichuan Province. (2021).\u003c/li\u003e\n\u003cli\u003eUniversity of Toronto \u0026amp; Association, C. M. Choosing Wisely Canada. https://choosingwiselycanada.org/recommendation/obstetrics-and-gynaecology/.\u003c/li\u003e\n\u003cli\u003eTang, C., Xu, J. \u0026amp; Zhang, M. The choice and preference for public-private health care among urban residents in China: evidence from a discrete choice experiment. \u003cem\u003eBMC. Health. Serv. Res.\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 580 (2016).\u003c/li\u003e\n\u003cli\u003eMa, X., liang, Z., Xiang, Y., Zhang, G. \u0026amp; Zhang, S. Chinese expert consensus on surgical approaches of hysterectomy for benign uterine disease (2021 edition). \u003cem\u003eChinese Journal of Practical Gynecology and Obstetrics\u003c/em\u003e \u003cstrong\u003e37, \u003c/strong\u003e821-825 (2021).\u003c/li\u003e\n\u003cli\u003eBaiocchi, M., Cheng, J. \u0026amp; Small, D. S. Instrumental variable methods for causal inference. \u003cem\u003eStat Med\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 2297-2340 (2014).\u003c/li\u003e\n\u003cli\u003eKhan-Gates, J. A., Ersek, J. L., Eberth, J. M., Adams, S. A. \u0026amp; Pruitt, S. L. Geographic Access to Mammography and Its Relationship to Breast Cancer Screening and Stage at Diagnosis: A Systematic Review. \u003cem\u003eWomens Health Issues\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 482-493 (2015).\u003c/li\u003e\n\u003cli\u003eNesbitt, R. C. et al\u003cem\u003e.\u003c/em\u003e Methods to measure potential spatial access to delivery care in low- and middle-income countries: a case study in rural Ghana. \u003cem\u003eInt. J. Health Geogr.\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 25 (2014). \u003c/li\u003e\n\u003cli\u003eChou, S., Deily, M. E. \u0026amp; Li, S. Travel distance and health outcomes for scheduled surgery. \u003cem\u003eMed Care\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e, 250-257 (2014).\u003c/li\u003e\n\u003cli\u003eGertler, P., Locay, L. \u0026amp; Sanderson, W. Are user fees regressive?: The welfare implications of health care financing proposals in Peru. \u003cem\u003eJ. Econom.\u003c/em\u003e \u003cstrong\u003e36,\u003c/strong\u003e 67-88 (1987). \u003c/li\u003e\n\u003cli\u003eNorton, E. C. \u0026amp; Staiger, D. O. How hospital ownership affects access to care for the uninsured. \u003cem\u003eRand. J. Econ.\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 171-185 (1994).\u003c/li\u003e\n\u003cli\u003eGowrisankaran, G. \u0026amp; Town, R. J. Estimating the quality of care in hospitals using instrumental variables.\u003cem\u003e J. Health. Econ.\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 747-767 (1999).\u003c/li\u003e\n\u003cli\u003eValley, T. S., Sjoding, M. W., Ryan, A. M., Iwashyna, T. J. \u0026amp; Cooke, C. R. Association of Intensive Care Unit Admission With Mortality Among Older Patients With Pneumonia. \u003cem\u003eJAMA\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 1272-1279 (2015).\u003c/li\u003e\n\u003cli\u003eStaiger, D. \u0026amp; Stock, J. H. Instrumental Variables Regression with Weak Instruments. \u003cem\u003eEconometrica\u003c/em\u003e \u003cstrong\u003e65\u003c/strong\u003e, 557-586 (1997).\u003c/li\u003e\n\u003cli\u003eStock, J., \u0026amp; Yogo, M. Testing for Weak Instruments in Linear IV Regression. In D. Andrews \u0026amp; J. Stock (Eds.), Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg. https://www.nber.org/system/files/working_papers/t0284/t0284.pdf (2005).\u003c/li\u003e\n\u003cli\u003eRosenbaum, P. R. \u0026amp; Rubin, D. B. Constructing a Control Group Using Multivariate Matched Sampling Methods That Incorporate the Propensity Score. \u003cem\u003eThe American Statistician\u003c/em\u003e \u003cstrong\u003e39\u003c/strong\u003e, 33-38 (1985).\u003c/li\u003e\n\u003cli\u003eStuart, E. A. Matching methods for causal inference: A review and a look forward. \u003cem\u003eStat.Sci. \u003c/em\u003e\u003cstrong\u003e25\u003c/strong\u003e, 1-21 (2010).\u003c/li\u003e\n\u003cli\u003eWan, F., Small, D., Bekelman, J. E. \u0026amp; Mitra, N. Bias in estimating the causal hazard ratio when using two-stage instrumental variable methods. \u003cem\u003eStat. Med.\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, 2235-2265 (2015).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3639662/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3639662/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eScholarly attention has been dedicated to the identification of low-value care (care that is not expected to provide a net benefit). Despite a consensus on the importance of hospital characteristics in explaining the use of low-value care, the precise influence of hospital ownership, herein the distinction between public and private ownership, remains unclear. This study included 38,865 hospital discharge records with hysterectomy procedures in China from 2016 to 2020 to describe the effect of public and private hospital ownership on the provision of low-value care and estimate the attributable risk ratio and corresponding attributable burden. Private hospitals were more likely to provide low-value hysterectomies, with the average incremental effect of 33.7% (95% CI, 23.5\u0026ndash;42.5%). Potential interventions in private hospitals could reduce this a maximum of 9.7% (95% eCI, 8.7\u0026ndash;10.4%) of low-value hysterectomy cases, corresponding to 48,375 (95% eCI, 43,254, to 51,706) annual cases and 1.82 (95% eCI, 1.63 to 1.94) billion USD costs nationally. For the first time, we identified the potential intervention target and estimated the maximum effectiveness of interventions to eliminate excessive risk of low-value care.\u003c/p\u003e","manuscriptTitle":"The burdens of low-value care in hysterectomy attributable to hospital ownership in China","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2023-11-28 19:21:12","doi":"10.21203/rs.3.rs-3639662/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4f66c535-21db-4838-bfa0-2916e3ea6f87","owner":[],"postedDate":"November 28th, 2023","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":26770378,"name":"Health sciences/Health care"},{"id":26770379,"name":"Health sciences/Health care/Health services"}],"tags":[],"updatedAt":"2024-01-12T08:25:47+00:00","versionOfRecord":[],"versionCreatedAt":"2023-11-28 19:21:12","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3639662","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3639662","identity":"rs-3639662","version":["v1"]},"buildId":"_2-kVJe1T_tPrBINL-cwx","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}