| BackgroundIn recent years, with the coverage of our country’s basic medical insurance expanding and universal, choosing the optimal payment systems becomes especially important. As the core content of medical care reform, reform of medical insurance payment systems not only relates to whether the medical expenses can be controlled effectively and the medical insurance fund balance of payments, but also whether health care reform will be success or not, social stability, harmonious and sustainable development. Our government pays more and more attention to the important role of the payment systems of medical insurance nowadays, many cities also carry out the exploration and practice of payment. But in general, payment reform is still unsatisfactory, problems still haven’t been solved effectively, the doctor-patient contradiction is increasingly fierce, so the exploring the optimal payment systems is imminent for our country. Health insurance in China is the third-party payers system, the choosing of payment systems will influence the physician’s income and the patient’s vital benefits.In the medical service market, there is principal-agent relationship exists between supplier and demander. If the physician acts as "perfect agent", the physician will make the decision which is best for patient’s health and benefit, because the physician focus on the patient’s health and benefit other than their own preferences, and this meets the requirement of medical ethics. But physician is the economic man, seeking for profit is their rational behavior, when the demander interests conflict with their own interests, the doctor usually pay attention to their own interests and deviate from the "perfect agent". Combined with asymmetric information and the uncertainty of treatment, the physicians will abuse this agency relationship for self-interest and lead to unreasonable profit behavior. The purpose of the medical insurance payment system is to control the supplier unreasonable profit behavior. However, there are many research about whether the payment system affect the physician’s behavior at home and abroad, the results are still controversial. Many researchers support the payment systems affect the physician’s behavior, but some researchers found that; there is no relationship between physician’s behavior and payment systems if exclude d the national, cultural and institutional difference. After review the literature, there are some limitations:(1) from the research methods, previous studies employed the method of systematic review and empirical study, but because of the limitations of the research method itself, such as system review articles only described the literature by groups, cannot carry on the synthesis of statistics because of the heterogeneity in national background, methodology design and judgment; Empirical research could not control the confounding factors such as demography and institutional factors, there exists choice and information bias. These lead to the lack of high quality research evidence and lead to the universality of the research results has been questioned. (2) From the content of the research, current researches lack the contents that which payment system will stimulate the physicians make the choices, which can improve the patient benefits, for which types of patients and current researches lack the evidence in patient’s health status and clinical outcomes. Therefore, this topic needs further researches. In order to compensate for the limitations of current research on the research methods and content, this study employ the experimental economics method which can be controlled strictly, simulate the real decision-making environment, study whether the payment systems affect the physicians’provision behavior with controlling the confounding variables, fill the domestic blank in exploring this topic.ObjectivesThe general objective is to validate that the payment systems affect the physician’s behavior by employing the experimental economics method in order to provide more reliable and scientific evidence for health policy makers in reforming and improving the basic medical insurance payment systems and provide the scientific evidence for the new period of health system reform. Specific objectives of this study are as follows:(1) Choosing the two most widely used payment systems in our country-Fee-for-service and Capitation, building the experimental economics lab, controlling the confounding variables and collecting the experimental data in order to get the stronger and more reliable evidence to verify the assumption that payment systems affect the physician behavior; (2) To compare the different incentives of two different payment systems deeply by comparing the medical service quantity, patient benefit, physician’s profit; (3) To analyze which payment system can stimulate the physicians make the choices, which can improve the patient benefits, for which types of patients, in other words is which health status patients can gain the maximum benefit from which kind of payment system.MethodsTwo most widely used payment systems Fee-for-service and Capitation in our country were chosen in this study, and adopted economics experiment which is designed by Hennig-Schmidt Heike and et al. from Bonn Econ Lab, laboratory for Experimental Economics, Department of Economics, University of Bonn. This research recruited 180 medical students as the subjects by posting the posters and assigned them to the six experimental groups randomly, but two subjects were refused to take part in the experiment because of being late,178 subjects participated in the experiments finally. Experiments were conducted in health economics laboratory of Shandong University, there are 40 computers for subjects and 1 computer for experimenters in the laboratory, and employed z-Tree to program and design experimental process. There are no real patients in the lab; patients are abstract, only for physician’s prescriptions. The subjects’task in the experiment is making decisions about the amount of medical service, at the same time, their decisions also determine subjects’s profit and patient benefit. Subjects made decisions under the guidance of the supervisor after entered the lab. After answered all the decisions, subjects got their salary according to their decisions, at the same time, the patient benefit was transferred to the RMB and was donated to real patients. Every subject’s experimental task is to provide a certain number of service quantity, q ∈[0,1,2,3,4...10], for the patients. Patients are furthered characterized by health status-moderate, good and poor which is corresponding to j=1,2,3. And every health status has five abstract illness k=A, B, C, D, E, so there will be 15 decisions needed to be made in every payment system. So every subject need to make 30 decisions in all, and all the decisions will be transmitted to the experimenter’s computer directly by z-Tree in the form of Excel.This study used software Stata/SE 12.0 for analyzing the decisions from the lab. Data analysis methods mainly include descriptive statistics analysis, single factor analysis and multiple-factor analysis. Because of the data not obey the normal distribution, this research employed Wilcoxon signed-rank test and Mann-Whitney U test to analyze the numerical variable data, using Pearson chi-square test to analyze the categorical variable. The preferences of subjects is viewed as the latent variable and varies between different subjects because the subjects’preferences of profits, patients benefit and other factors are unknown and their preferences may influence their decision choice. This study structured the ordered Logit and Probit models under the structure of GLLAMMs (Generalized linear latent and mixed models) to analyze whether the payment systems affect the quantity of medical service and whether the medical care quantity is different between two payment systems or not; also structured Logit, Probit and Complentary log-log models to analyze whether the payment systems incentive subjects to choose optimal quantity decisions, optimal patient benefit decisions and maximum physicians’profit decisions or not, and explore the difference of behavior in choosing optimal quantity and Pareto optimal decisions between the two payment systems.Results(1) Quantity of medical service:Firstly, patients are over-served at the aggregate level under Fee-for-service. The mean quantity for patients with medium and good health status is larger than optimal quantity of medical service. With the need of medical service getting higher, the deviation from optimal quantity of medical service is getting smaller, that means the degree of over-served will decrease with the need of medical service getting higher. The mean deviation between real quantity and optimal quantity of 86.11% subjects is larger than zero. 43.45% decisions are optimal quantity decisions, with the health status getting worse, the proportion of optimal quantity decisions increase.43.45% of medical decisions are optimal quantity decisions, with the patient’s health status getting worse, the proportion of optimal quantity decisions increase.Secondly, patients are under-served at the aggregate level under Capitation. The mean quantity for patients with medium and bad health status is less than optimal quantity of medical service. With the need of medical service getting higher, the deviation from optimal quantity of medical service is getting larger, that means the degree of under-served will get worse with the need of medical service getting higher. The mean deviation between real quantity and optimal quantity of 75.84% subjects is less than zero.61.99% decisions are optimal quantity decisions, with the health status getting worse, the proportion of optimal quantity decisions decrease.Thirdly, the mean quantity of medical service under Fee-for-service is larger 35.98% than Capitation. With the patient’s health status getting worse, the deviation between two payment systems is getting smaller. The amount of optimal quantity decisions under Fee-for-service is larger 42.67% than Capitation.(2) Patient benefit:Firstly, patient benefit under Fee-for-service is less than optimal patient benefit at the aggregate level. With the health status getting worse, the loss of patient benefits decrease. The mean deviation between real health benefit and optimal health benefit of 91.57% subjects is less than zero.43.45% of medical decisions are optimal health benefit decisions. With the patient’s health status getting worse, the proportion of optimal health benefit decisions increase.Secondly, patient benefit under Capitation is less than optimal patient benefit at the aggregate level. With the health status getting worse, the loss of patient benefits increase. The mean deviation between real health benefit and optimal health benefit of 86.52% subjects is less than zero.61.99% of medical decisions are optimal health benefit decisions. With the patient’s health status getting worse, the proportion of optimal health benefit decisions decrease.Thirdly, patient benefit under Fee-for-service is larger than Capitation. The mean patient benefits of good and medium health status under Fee-for-service are smaller than Capitation, and for patient with bad health status, the result is opposite, the amount of optimal patient benefit decisions under Capitation is larger 42.67% than Fee-for-service.(3) Physician’s profit:Firstly, the profit is less than maximum profit at the aggregate level. With the health status getting worse, the loss of profit decrease. The mean deviation between real profit and maximum profit of all subjects is less than zero.16.14% of medical decisions are maximum profit decisions. The proportion of maximum profit decisions for patients with mediate health status is the highest, for patients with bad health status is the lowest.Secondly, the profit is less than at the aggregate level. With the health status getting worse, the loss of profit increase. The mean deviation between real profit and maximum profit of all subjects is less than zero. Only 0.49% decisions are maximum profit decisions.Thirdly, the mean profit under Fee-for-service is smaller 12.18% than Capitation. The mean profit under Fee-for-service from patient with bad status is larger 9.71% than Capitation, the result of the left two health status, the results are opposite. The amount of maximum profit decisions under Capitation is less 96.98% than Fee-for-service.Conclusion and policy implicationBy conducting the controlled experiment, we can make the conclusion that the payment systems affect the physicians’ behavior significant, every single payment has both sides influence for physicians’ provision behavior. In specific description are:(1) Quantity of medical service:Fee-for-service incentives physician to provide excess quantity of health services, with the patient’s health status getting worse, the degree of excess provision of medical service will decrease. Capitation incentives physician to provide inadequate quantity of health services, with the patient’s health status getting worse, the degree of inadequate provision of medical service will increase. Fee-for-service incentives physician to provide more quantity of health services than Capitation. The amount of optimal quantity decision under Capitation is more than Fee-for-service. The amount of optimal quantity decision for patients with good (mediate) health status under Capitation is more than Fee-for-service, but the amount of optimal quantity decisions for patients with bad health status is less than Fee-for-service.(2) Patient benefit:Patient benefits from the physicians’decisions in both payment systems are smaller than the optimal health benefit at the aggregate level. Under Fee-for-service, with the patient’s health status getting worse, the loss of patient benefit will decrease. Under Capitation, with the patient’s health status getting worse, the loss of patient benefit will increase. Patients benefit under Fee-for-service is larger than patients benefit under Capitation. The amount of optimal patient benefit decisions under Fee-for-service is more than Capitation. The optimal patient benefit decisions for patients with good (mediate) health status under Capitation is more than Fee-for-service, but the amount of optimal patient benefit decisions for patients with bad health status is less than Fee-for-service.(3) Physician’s profit:The maximum of profit is not the only motivation for physician’s quantity decisions. Both payment systems incentive physicians to make the decisions which profits are smaller than maximum profit at the aggregate level. Under Fee-for-service, with the patient’s health status getting worse, the loss of profit will decrease. Under Capitation, with the patient’s health status getting worse, the loss of profit will increase. Compared with Fee-for-service, physician’s profit under Capitation is larger. Physicians provide more maximum profit decisions under Fee-for-service than Capitation. Physicians provide most maximum profit decisions for patients with mediate health status, and fewest maximum profit decisions for patients with bad health status.Given these results, this study gives the policy implications as follows:(1) Exploring the mixed payment systems, according to the characteristics of the single payment system, foster strengths and circumvent weaknesses to realize the advantageous complementarities; (2) Paying abundant attention and developing the positive role in providing the medical service for the patients with servious illness; (3) Strengthening the government financial input to the medical institutions, and improving the physicians’ income; (4) Establishing the prescription evaluation mechanism to evaluate and supervise the physicians’ provision behavior according to the characteristics of the different payment methods. |
PDF Full Text Request