Department of Statistics, National Chengchi University, Taipei, Taiwan, Republic of China.
Department of Statistical Information and Actuarial Science, Aletheia University, New Taipei City, Taiwan, Republic of China.
PLoS One. 2024 Sep 25;19(9):e0307508. doi: 10.1371/journal.pone.0307508. eCollection 2024.
Prolonging life is a global trend, and more medical expenditure is being spent on chronic diseases owing to population aging. Diseases commonly seen in middle-aged and elderly people, such as heart disease and diabetes, have slowed mortality improvement in recent years. Diabetes is a common chronic disease and comorbidity of many serious health conditions. The total estimated cost of diabetes in the United States was $327 billion in 2017. However, many people are unaware that diabetes is common, and at least 21.4% of adults do not know that they have diabetes. The number of diabetes-related deaths has been increasing, and diabetes was the 5th cause of death in Taiwan in 2019. In this study, we explore the trend and influence of diabetes in Taiwan and apply mortality models, such as the Lee-Carter and Age-Period-Cohort models, using data from Taiwan's National Insurance to model the incidence and mortality rates of diabetes. We found that the Lee-Carter model provides fairly satisfactory estimates and that people with diabetes regularly taking diabetes medication have lower mortality rates. Moreover, we demonstrate how these results can be used to design diabetes related insurance products and prepare the insured to face the impact of incurring diabetes. In addition, we consider different criteria for judging whether people have diabetes (as there is no consensus on these criteria) and investigate the issue of moral hazard in designing diabetes insurance products.
延长寿命是一个全球性的趋势，由于人口老龄化，更多的医疗支出用于治疗慢性病。近年来，心脏病和糖尿病等中老年人常见的疾病减缓了死亡率的改善。糖尿病是一种常见的慢性病，也是许多严重健康状况的合并症。2017 年，美国糖尿病的总估计成本为 3270 亿美元。然而，许多人不知道糖尿病很常见，至少有 21.4%的成年人不知道自己患有糖尿病。与糖尿病相关的死亡人数一直在增加，2019 年糖尿病是台湾的第五大死因。在这项研究中，我们探讨了糖尿病在台湾的趋势和影响，并应用死亡率模型，如 Lee-Carter 和 Age-Period-Cohort 模型，使用台湾国家保险数据来模拟糖尿病的发病率和死亡率。我们发现，Lee-Carter 模型提供了相当令人满意的估计，并且经常服用糖尿病药物的糖尿病患者的死亡率较低。此外，我们展示了如何使用这些结果来设计与糖尿病相关的保险产品，并为被保险人做好应对患糖尿病影响的准备。此外，我们还考虑了判断人们是否患有糖尿病的不同标准（因为这些标准没有共识），并研究了在设计糖尿病保险产品时的道德风险问题。
