Hsieh Hui-Min, Tsai Shu-Ling, Shin Shyi-Jang, Mau Lih-Wen, Chiu Herng-Chia
*Department of Public Health, Kaohsiung Medical University, Kaohsiung, Taiwan †National Health Insurance Administration, Ministry of Health and Welfare, Taipei ‡Division of Endocrinology and Metabolism, Graduate Institute of Medical Genetics, College of Medicine, Kaohsiung Medical University, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan §National Marrow Donor Organization, Edina, MN ∥Department of Healthcare Administration and Medical Informatics, Kaohsiung Medical University ¶Department of Business Administration, National Sun Yat-Sen University, Kaohsiung, Taiwan.
Med Care. 2015 Feb;53(2):106-15. doi: 10.1097/MLR.0000000000000264.
Taiwan's National Health Insurance (NHI) Program implemented a diabetes pay-for-performance program (P4P) based on process-of-care measures in 2001. In late 2006, that P4P program was revised to also include achievement of intermediate health outcomes.
This study examined to what extent these 2 P4P incentive designs have been cost-effective and what the difference in effect may have been.
Analyzing data using 3 population-based longitudinal databases (NHI's P4P dataset, NHI's claims database, and Taiwan's death registry), we compared costs and effectiveness between P4P and non-P4P diabetes patient groups in each phase. Propensity score matching was used to match comparable control groups for intervention groups. Outcomes included life-years, quality-adjusted life-years (QALYs), program intervention costs, cost-savings, and incremental cost-effectiveness ratios.
QALYs for P4P patients and non-P4P patients were 2.08 and 1.99 in phase 1 and 2.08 and 2.02 in phase 2. The average incremental intervention costs per QALYs was TWD$335,546 in phase 1 and TWD$298,606 in phase 2. The average incremental all-cause medical costs saved by the P4P program per QALYs were TWD$602,167 in phase 1 and TWD$661,163 in phase 2. The findings indicated that both P4P programs were cost-effective and the resulting return on investment was 1.8:1 in phase 1 and 2.0:1 in phase 2.
We conclude that the diabetes P4P program in both phases enabled the long-term cost-effective use of resources and cost-savings regardless of whether a bonus for intermediate outcome improvement was added to a process-based P4P incentive design.
台湾地区的全民健康保险计划于2001年实施了一项基于医疗过程指标的糖尿病绩效付费计划（P4P）。2006年末，该P4P计划进行了修订，将中期健康结果的达成情况也纳入其中。
本研究考察了这两种P4P激励设计在何种程度上具有成本效益，以及效果上可能存在的差异。
我们使用3个基于人群的纵向数据库（全民健康保险的P4P数据集、全民健康保险的理赔数据库以及台湾地区的死亡登记处）进行数据分析，比较了各阶段P4P和非P4P糖尿病患者组之间的成本和效果。采用倾向得分匹配法为干预组匹配可比的对照组。结果包括生命年数、质量调整生命年数（QALY）、项目干预成本、成本节约以及增量成本效益比。
第一阶段P4P患者和非P4P患者的QALY分别为2.08和1.99，第二阶段分别为2.08和2.02。第一阶段每QALY的平均增量干预成本为新台币335,546元，第二阶段为新台币298,606元。P4P计划每QALY平均节省的全因医疗成本在第一阶段为新台币602,167元，第二阶段为新台币661,163元。研究结果表明，两个阶段的P4P计划均具有成本效益，第一阶段的投资回报率为1.8:1，第二阶段为2.0:1。
我们得出结论，无论在基于过程的P4P激励设计中是否增加了中期结果改善的奖励，两个阶段的糖尿病P4P计划都能实现资源的长期成本效益利用和成本节约。
