| Effective control of the unwarranted increase in healthcare expenditures remains a major challenge confronted by healthcare delivery systems worldwide.One effective strategy is to conduct an in-depth analysis to find out factors associated with higher expenditures in particular regions via the adoption of a "natural experiment" produced by variations embedded in healthcare expenditures among different regions within a particular country to provide evidence-based implications to inform decision-making procedures under such a context.For our country,the same cost control challenge has been facing the rising healthcare expenditures.Relevant research has been carried out,especially the analysis of the causes of healthcare spending growth.However,in-depth analysis regarding variations embedded in healthcare expenditures among different regions has been relatively limited.As an attempt to bridge such research gap as well as to provide evidencebased implications to inform the implementation of effective strategies aimed at curtailing unwarranted increases in healthcare expenditures,this study is proposed to achieve the following goals under the context of China’s healthcare system,guided by the theoretic regional variation analysis framework of medical services,an indepth investigation was conducted to identify factors associated with inter-regional variations in health care expenditures as well as the underlying mechanism functioning behind such phenomenon based on detailed information retrieved from China’s nationwide hospital service system regarding nationwide hospitals’ organizational information and individuals’ hospitalizations.This paper focuses on three research questions: 1)Considering patients’ medical service-seeking behaviors are typically not confined by administrative boundaries,there may be bias when trying to account for variations among regions;relevant research proposes to delineate a map of hospital service areas to accurately connect demand and supply side,how it can be applied to the health care system in China? 2)Are there any regional variations of healthcare expenditures in China? If there are,what causes such variations? How do relative factors contribute to the variations? 3)Which variations in healthcare expenditures among regions are unwarranted? How associated factors and mechanisms affect the unwarranted variations?This study was conducted in Sichuan Province,China,using multi-source data from individuals,hospitals,and regions,mainly collecting hospitalization records of individuals over 10 million in the period year of 2017 to 2019,combined with the annual report data for 2,493 hospitals,as well as data at the regional level,such as the administrative boundaries,traffic network,and social,economic development in corresponding years.Firstly,considering patients’ medical service-seeking behaviors are typically not confined by administrative boundaries,a map that delineates hospital service areas across China was produced to accurately match patients and healthcare providers within the same regions by using hospital service areas as the units for analysis instead of simply using administrative boundaries.This study used hospital discharge data in the fourth quarter of 2017 to 2019 collected in Sichuan Province to capture flows of patients’ actual health-seeking behaviors.Towns/subdistricts of Sichuan Province are regarded as the basic units,and administrative boundary method,the network community detection method and the Dartmouth method were applied to delineate hospital service areas.The study calculated a localization index to evaluate the efficiency of supply-demand matching using different delineation methods.Finally,this study delineated 183,183,and 151 hospital service areas,respectively.The median local medical index for the three delineation methods was 72.23%,69.07%,and 71.22%,respectively.Although the administrative boundary method had the highest median localization index,it did not guarantee that each hospital service area had a high localization index.Therefore,the study chose the result of the Dartmouth method with the highest supply-demand matching efficiency as the basic geographic spatial unit for subsequent analysis of regional variation and causes of health care expenditures.Secondly,the delineated hospital service areas by the Dartmouth method was as units of analysis,combined with discharge records,detailed hospital administrative reports and regional data were used to investigate variations in per capita inpatient expenditures,log-linear regression models were applied to investigate correlated factors and the Shapley value decomposition method was used to what extent each potential factor contributes.This study found that the variations in per capita inpatient expenditures adjusted by age and gender among hospital service areas in Sichuan Province fluctuated between 2.76 to 4.42 times from 2017 to 2019.Among the demand-side factors that contributed to the variations in per capita inpatient expenditures,the proportion of unemployed individuals was negatively correlated with per capita inpatient expenditures,while per capita GDP and urbanization rate were positively correlated.Among the supply-side factors,each increase of 1% in the number of beds per thousand people resulted in a 3.8% increase in per capita inpatient expenditures.The Shapley value decomposition method revealed that demand-side factors explained most of the variations in per capita inpatient expenditures among hospital service areas,accounting for 44.37% of the total R2,while supply-side factors explained 34.34%,with the number of beds per thousand people contributing the most at 30.54%.Thirdly,this study decomposed health expenditures into the price of health services multiplied by the quantity of health services.Since most hospitals in China that are included in the designated health insurance network follow a unified government-guided pricing system,thereby this chapter will focus on identifying clues of regional variations in health expenditures caused by inappropriate utilization of health services.This study chose vaginal delivery high-risk pregnant women who did not undergo cesarean section as a representative medical service of "underu se",and low-risk cesarean section delivery and hospitalization for ambulatory care sensitive conditions(ACSCs)were identified as representative medical services of "overuse." This study analyzed the "underuse" and "overuse" of certain health services using three-level logistic regression models to investigate potential factors causing unwarranted variations among regions.The results showed that the higher the proportion of private hospitals in hospital service areas,the greater the likelihood that patients are at high risk of a vaginal delivery without performing a cesarean section.Regarding the factors contributing to the "overuse" of low-risk vaginal delivery and hospitalization of ACSCs,three main findings were identified.Firstly,the hospital ownership may affect the overuse of cesarean section,with a higher proportion of private hospitals in the hospital service area being a risk factor for lowrisk vaginal delivery.Secondly,bed capacity showed a positive correlation with the overuse of these two services.Finally,this study found that the accessibility of primary health resources in hospital service areas was a protective factor for hospitalization of ACSCs.Fourth,spending more money will be okay if it produces better health outcomes.Otherwise,the variations in health expenditures could be considered “unwarranted”.This study selected three representative diseases(congestive heart failure,stroke,and acute myocardial infarction)as well as two health outcomes(in-hospital mortality and 30-day unplanned all-cause readmission)to measure health quality.First,a linear probability model was used to analyze the correlation between health expenditures and quality of care.Second,considering the potential endogeneity problem of health expenditures in the regression,the study further introduced instrumental variables and used a two-stage least squares method to identify the impact of health expenditures on quality of care.The study showed that the effect of health expenditures on quality of care varies by disease and by the quality measurement indicator used.The two-stage least squares method showed the higher the total inpatient expenditures for stroke patients,the lower the probability of patients experiencing all-cause unplanned readmissions within 30 days.For every 1%increase in total inpatient expenditures,the probability of experiencing all-cause unplanned readmissions within 30 days decreases by 0.003%.No statistically significant causal effect was found between total inpatient expenditures and inhospital mortality,and no other causal effects of inpatient expenditures on quality of care were found for other diseases.Summary: Analyzing the variations in health expenditures between regions is beneficial for identifying the causes of unwarranted growth in health expenditures.Considering the actual situation of patient’s health seeking behavior across administrative boundaries,the hospital service area based on supply-demand matching can serve as the basic spatial and geographic analysis unit for analyzing the variations in health expenditures across regions and enriching the policy toolbox for controlling the rapid growth of health expenditures.Among supply-side factors,the number of beds,which represents the hospital’s size and service capacity,contributes the most to the variations in per capita inpatient expenditures across regions.In terms of health service utilization,in addition to the number of beds,hospital ownership is a potential risk factor that may affect the overuse of representative health services,and the accessibility of primary healthcare resources is a protective factor that may affect the overuse of representative health services.Regarding quality of care,the results showed that the effect of health expenditures on quality of care varies depending on the disease and the quality measurement indicators.Based on the findings of this study,it is recommended to establish a public "variation map set" of health expenditures based on hospital service areas,to achieve timely collection and dynamic monitoring and evaluation of health expenditures information nationwide and regionally;to continuously establish guidelines based on health theory and clinical evidence,and pay special attention to the utilization of "unwarranted" of medical services to avoid waste of health resources,increase in patient expenditures,and inefficiency of the health system;to continue to pay attention to the complex relationship between health expenditures and quality of care,and especially focus on the health output(such as health quality)of high per capita inpatient expenditure regions,while controlling the health expenditures in these regions to reflect the value of health services. |
PDF Full Text Request