MINISTRY OF EDUCATION AND TRAINING UNIVERSITY OF ECONOMICS HO CHI MINH CITY IMPACTS OF HEALTH INSURANCE IN VIETNAM ON HEALTHCARE UTILIZATION, SELF-REPORTED HEALTH, AND FINANCIAL CHOICES DOCTORAL THESIS IN ECONOMICS Ho Chi Minh City, 2022 MINISTRY OF EDUCATION AND TRAINING UNIVERSITY OF ECONOMICS HO CHI MINH CITY IMPACTS OF HEALTH INSURANCE IN VIETNAM ON HEALTHCARE UTILIZATION, SELF-REPORTED HEALTH, AND FINANCIAL CHOICES DOCTORAL THESIS IN ECONOMICS Major: Economic development Code: 9310105 Academic advisors: Dr. Le Thanh Loan Assoc. Pham Khanh Nam Ho Chi Minh City, 2022 i Declaration I hereby declare that this thesis is entirely my own work, with the exception of quotations and citations, which have been appropriately acknowledged. This thesis has not been submitted to any other academic or non-academic institution for any degree or certification.
Ho Chi Minh City, December 2022 Truong Anh Tuan ii Table of Contents DECLARATION. I TABLE OF CONTENTS. VI LIST OF TABLES .VII LIST OF FIGURES. VIII CHAPTER 1: INTRODUCTION .1 HEALTH INSURANCE AND HEALTHCARE SCHEME IN VIETNAM.
Development of health insurance in Vietnam. Healthcare scheme in Vietnam .7 OBJECTIVES OF THESIS .12 SCOPE OF THE THESIS .12 STRUCTURE OF THE THESIS .13 CHAPTER 2: THEORY AND LITERATURE REVIEW. Model of demand for health. The behavioral model of health services utilization.
Moral hazard and Cost-sharing. Literature relevant to the heterogeneous impact of health insurance coverage on healthcare utilization. Theory of precautionary savings. Program impact measures.
Program impact evaluation methods .31 EMPIRICAL LITERATURE REVIEW AND RESEARCH GAPS. Review of key findings. Review of methodologies .46 JUSTIFICATION OF ANALYTICAL METHODOLOGIES.50 CHAPTER 3: IMPACTS OF HEALTH INSURANCE ON HEALTHCARE UTILIZATION AND OUT-OF-POCKET EXPENDITURES. Data and variables.
Impacts of the health insurance program for the elderly on the probability of being insured. Impacts of health insurance on healthcare utilization outcomes .67 MODEL SIGNIFICANCE TESTING. Validity of RDD .76 CHAPTER 4: IMPACTS OF HEALTH INSURANCE ON HOUSEHOLDS’ FINANCIAL CHOICES: EVIDENCE FROM VIETNAM .78 DATA AND VARIABLES. Variable definitions and descriptive statistics .89 iv RESEARCH HYPOTHESES AND METHODS.
Impacts of covariates. Impact of health insurance on financial choices. Correlations between the error terms .106 CHAPTER 5: IMPACTS OF CO-PAYMENTS ON SELF-REPORTED HEALTH AMONG THE HEALTH INSURANCE INSURED IN RURAL VIETNAM .108 CONCEPTUAL FRAMEWORK AND HYPOTHESIS .113 DATA AND VARIABLES. Treatment and outcome variables.
Propensity score estimation. Covariate balance check. The impact of co-payment exemption on self-reported health. Sensitive test for hidden bias.
The impacts of health insurance on healthcare utilization and out-of- pocket expenditures. The impact of health insurance on households’ financial choices. The impacts of co-payments on self-reported health .134 CONTRIBUTIONS OF THE THESIS .134 POLICY IMPLICATIONS OF THE THESIS FINDINGS .136 LIMITATIONS AND AREAS FOR FUTURE RESEARCH .138 LIST OF THE AUTHOR'S PUBLICATIONS .163 vi Abbreviations ATE: Average Treatment Effect ATT: Average Treatment Effect of the Treated BMI: Body Mass Index DID: Difference–in–Differences ITT: Intention-to-treat effect IV: Instrumental variable LATE: Local average treatment effect LMICs: Low- and middle-income countries OLS: Ordinary least squares PSM: Propensity score matching RCTs: Randomized controlled trials RD: Regression discontinuity RDD: Regression discontinuity design TVSEP: Thailand-Vietnam Socioeconomic Panel UHC: Universal health coverage US: United States VHLSS: Vietnam Household Living Standard Survey VSS: Vietnam Social Security Agency 2SLS: Two-stage least squares vii List of tables Table 1.1: Health insurance reform in Vietnam .1: Overview of econometric methods for program impact evaluation .1: Descriptive statistics of control variables .2: Descriptive statistics of outcome variables .3: Impact of the health insurance program for the elderly on the probability of being insured .4: Regression model for the probability of being insured when reaching 80 years old, using pre-policy data .5: Impacts of health insurance on healthcare utilization outcomes .6: RD estimates of covariates .7: Robustness checks with different bandwidths .8: Intention-to-treat effects.1: Definitions of dependent variables .2: Definitions of control variables .3: Descriptive statistics of variables (mean, standard deviation).4: Pairwise tetrachoric correlations between dependent variables .5: Univariate probit coefficients .6: Multivariate probit coefficients .7: Correlation between error terms in multivariate probit regressions .8: IV random-effects two-stage least-squares regressions .9: Bivariate probit models .1: Variable definitions and descriptive statistics .2: Covariate balance check and bias reduction .3: Statistical tests pre-matching and post-matching .4: Impact of co-payment exemption on self-reported health .5: Heterogeneous effects of co-payment exemption on self-reported health .6: Mantel-Haenszel test .129 viii List of figures Figure 1.1 Organisational chart of the Vietnamese healthcare system .2: Payment mechanism in health insurance .1: Andersen’s Behavioral Model of Healthcare Utilization .2 Conventional ex-post moral hazard effect .3: Conceptual framework for the thesis .1: Conceptual framework for the impacts of health insurance on healthcare utilization and out-of-pocket expenditures .2 Insurance coverage, aged 70 – 89 years old .3 Histogram of age around the cutoff point (960 months) .4 Estimated density of age around the cutoff point (960 months) .5 RD plots of covariates .1 Map of Vietnam's surveyed provinces.2: Conceptual framework for the impacts of health insurance on financial choices .1: Conceptual framework for the impact of co-payment on self-reported health .2: Histogram of propensity score distribution for treated and control individuals .3: Standardised percent bias of covariates prior and post matching .4: The density distribution of propensity scores before and after the PSM for treated and control groups .125 ix ABSTRACT As in many other low- and middle-income countries, Vietnam is implementing a social health insurance program as one method of reaching universal health coverage. In addition to serving as a health financing mechanism for universal health coverage, this program undoubtedly impacts the beneficiaries' healthcare behaviors.
This thesis examines the impacts of health insurance in Vietnam on a range of healthcare behaviors, including healthcare usage, out-of-pocket spending, financial decisions, and self-reported health. Based on a conceptual framework derived from the theoretical literature, the present thesis includes three sub-studies in the form of three essays to investigate the impacts of health insurance on healthcare utilization, out-of- pocket expenditure, and households' financial choices; and the effect of the co- payment scheme on self-reported health. A regression discontinuity design was used to study the effects of healthcare use and out-of-pocket spending with data from the Vietnam Household Living Standard Survey. Impacts on households' financial choices and self-reported health were investigated using data from the Thailand-Vietnam Socio-economic Panel surveys.
A recursive multivariate probit model was used to examine the effects of health insurance on households' financial choices, while propensity score matching was employed to analyze how the removal of co-payments affected patients' assessments of their own health. The results of the first sub-study demonstrated that insurance helps individuals reduce the expenditures per outpatient visit but has no effect on the outcomes, including the likelihood of an outpatient visit, the probability of an inpatient visit, the number of outpatient visits, the number of inpatient visits, and the expenditures per inpatient visit. The second sub-study revealed that, although health insurance has no influence on private health insurance, it has positive effects on savings and investment and a negative impact on credit choice. The third sub-study discovered that co-payment exemption has a favorable impact on the self-reported health of the insured.
In x addition, the exemption for co-payments has varied effects across income levels and locations. Key words: health insurance, healthcare utilization, financial choice, co-payment, self-reported health xi TÓM TẮT Cũng như nhiều quốc gia có thu nhập thấp và trung bình khác, Việt Nam sử dụng chương trình bảo hiểm y tế xã hội (social health insurance) như một phương thức để đạt bao phủ chăm sóc sức khỏe toàn dân (universal health coverage). Ngoài vai trò là một cơ chế tài chính y tế nhằm hướng đến bao phủ sức khỏe toàn dân, chương trình này chắc chắn tác động đến hành vi chăm sóc sức khỏe của người thụ hưởng. Luận án này nghiên cứu tác động của bảo hiểm y tế ở Việt Nam đối với một số các hành vi liên quan đến chăm sóc sức khỏe như: mức sử dụng dịch vụ chăm sóc sức khỏe, chi tiêu tự túi, quyết định tài chính của hộ gia đình và tình trạng sức khỏe của người thụ hưởng.
Sử dụng khung lý thuyết được xây dựng từ các lý thuyết kinh tế, luận án thực hiện ba nghiên cứu dưới dạng ba tiểu luận để xem xét tác động của bảo hiểm y tế đối với việc sử dụng dịch vụ chăm sóc sức khỏe, chi tiêu tự túi và lựa chọn tài chính của các hộ gia đình; và ảnh hưởng của chính sách đồng chi trả (co-payment) đến tính trạng sức khỏe của người tham gia. Phương pháp hồi quy gián đoạn (regression discontinuity design) được sử dụng để nghiên cứu tác động của bảo hiểm y tế đến mức sử dụng dịch vụ chăm sóc sức khỏe và chi tiêu tự túi, với dữ liệu từ các cuộc Khảo sát mức sống hộ gia đình Việt Nam (VHLSS). Các nghiên cứu tác động của bảo hiểm y tế đến lựa chọn tài chính của hộ gia đình và tình trạng sức khỏe sử dụng dữ liệu từ các cuộc Khảo sát dữ liệu bảng về kinh tế xã hội Thái lan- Việt Nam (Thailand-Vietnam Socio-economic Panel- TVSEP). Mô hình probit đa biến đệ quy (recursive multivariate probit) được sử dụng để nghiên cứu tác động của bảo hiểm y tế đến lựa chọn tài chính của các hộ gia đình, trong khi đó phương pháp so sánh bằng điểm xu hướng (propensity score matching) được sử dụng để nghiên cứu tác động của việc miễn đồng chi trả đến tình trạng sức khỏe của người tham gia.
Kết quả của nghiên thứ nhất chứng minh rằng bảo hiểm y tế tác động làm giảm chi phí khám bệnh ngoại trú, nhưng không tác động đến các kết quả như: xác suất khám bệnh ngoại trú, xác suất khám bệnh nội trú, số lần khám bệnh ngoại trú, số lần khám xii bệnh nội trú và chi phí cho mỗi lần khám bệnh nội trú. Nghiên cứu thứ hai cho thấy, mặc dù bảo hiểm y tế không ảnh hưởng đến bảo hiểm y tế tư nhân, nhưng nó có tác động dương đến tiết kiệm và đầu tư, và tác động âm đến tín dụng. Nghiên cứu thứ ba cho thấy việc miễn đồng chi trả có tác động tích cực đến tình trạng sức khỏe của người được bảo hiểm. Ngoài ra, tác động của việc miễn đồng chi trả thay đổi theo thu nhập và địa phương.
Từ khóa: bảo hiểm y tế, mức sử dụng dịch vụ y tế, lựa chọn tài chính, đồng chi trả, tình trạng sức khỏe. 1 Chapter 1: INTRODUCTION Background Universal health coverage (UHC) is prominent among 13 health-related targets in the United Nations Sustainable Development Goals. This target includes “financial risk protection, access to quality essential healthcare services, and access to safe, effective, quality, and affordable essential medicines and vaccines for all” (World Health Organization, 2015). In recent years, UHC has been a major policy goal across the world (Marten et al.
However, Low- and middle-income countries (LMICs) generally confront unique obstacles in attaining UHC because of their low healthcare resources, inefficient use of resources, dependence on out-of-pocket payments, and large populations (World Health Organization, 2010). Therefore, the fundamental challenge facing LMICs is how their health financing schemes can enhance people's access to healthcare services. While each country's path to UHC is unique, scaling up the population's health insurance has gained popularity as a health financing mechanism for UHC in LMICs (Barasa et al. As a result of risk sharing, health insurance protects its customers against unexpectedly large medical expenditures and increases their access to essential medical care (Qiu & Wu, 2019).