The Effects of Pay-for-Performance and Process of Care on Health Outcomes of Diabetic Patients

• Main Authors: Chia-Hui Tan, 譚家惠
• Other Authors: Ming-Chin Yang
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/33506585416530158046

博士 === 國立臺灣大學 === 健康政策與管理研究所 === 102 === Objective: A number of studies have examined the impacts of pay-for-performance programs on quality of care, but little is known about long-term effects of these programs on the health care outcomes. Objective: The study aimed to examine the effects of the pay-for-performance program for type 2 diabetes patients on diabetes-related complications. Materials and methods: A longitudinal cohort study with 5-year follow-up was used to evaluate the impact of pay-for-performance program on diabetes-related complications. Research materials came from claims files of the Longitudinal Health Insurance Database (LHID) 2005 released from the Nation Health Research Institute. Patients newly diagnosed as diabetes (with more than 3 times of ambulatory visits diagnosed as diabetes [ICD-9-CM: 250.xx] or 1 time hospitalization with major diagnosis of diabetes) in 2004-2006 were included in the study. The case group were defined as joined the pay-for-performance program and received the comprehensive care over 12 months during 2004 to 2010. The control group were defined as never joined the pay-for-performance program during follow-up period. Patients who involved the pay-for-performance program or not is not randomization, we applied the propensity score matching (PSM) as a mean to increase the comparatives between these two groups. The incidences of six diabetes-related complications, including retinopathy, neuropathy, nephropathy, cerebrovascular, cardiovascular, peripheral vascular disease and metabolic disease) were analysed. Different propensity score methods and survival analysis were use to estimate whether involved in the pay-for-performance program was able to decrease the risk of cardiovascular events for diabetes patients. The SAS v9.3 package software was used the database process and statistical analyses. Results: Patients with newly diagnosed diabetes who participated in the pay-for-performance program for over 12 months experienced cardiovascular event significantly later than non-participants. The marginal hazard ratios of different propensity score method ranged from 0.60 to 0.63. There were significantly different risk of heart failure, myocardial infarction and stroke between those participated in the pay-for-performance program for over 12 months and those did not. Of these propensity score methods, propensity score matching and inverse probability of treatment weighting (IPTW) using propensity score resulted in the estimate with lower standardized difference of means between participant and non-participant groups. Conclusions: Cardiovascular events were significantly reduced over 5-year period among patients newly diagnosed with diabetes who participated in the pay-for-performance program for over 12 months. When estimating relative effects of interventions, propensity score matching and IPTW using the propensity score were recommended to apply to reduce the systematic bias between groups.