台灣健保醫療費用時間序列預測模型

本文探究影響醫療費用成長的重要因素，並探討台灣實施的三種抑制醫療費用成長措施，亦即牙醫總額支付制度、中醫總額支付制度及門診部份負擔新制對於醫療費用所造成的衝擊。結果發現：（一）在自變數的影響效果方面：前一、二期的醫療費用對總醫療費用成長率有顯著負的效應，平均投保薪資、人口老化指數和標準死亡比對總醫療費用成長率有顯著正的效應，而投保人數、特約醫療機構數和通貨膨脹率並無顯著影響。（二）在制度的衝擊方面：政府在資料期間實施的牙醫總額支付制度、中醫總額支付制度及門診部份負擔新制，對醫療費用成長率並無顯著影響。（三）預測精確度方面：本文模型所計算出之絕對相對誤差（absolute relative error）平均值較ARIMA模型計算所得之值為小。

關鍵字

醫療費用支出；時間序列預測模型

並列摘要

In this paper, we explore the key factors for the growth of National Health Insurance expenditure and the impact of three most important expenditure containment policies that have been practiced during our data period. The results of empirical research were presented as follows: First, prior one and two month health care expenditure affected the rate of current health care expenditure negatively. On the other hand, average insured wage, ratio between population 65 years of age and over and population aged 15 to 64, and standard death rate had positive effect on the rate of health care expenditure. The amount of insured persons, contracted medical care institutions, and rate of inflation had no significant effect on it. Second, three supply-side incentive mechanism including the global budget policies on dental institutions and institutions practicing Chinese medicine, the new copayment system in ambulatory are of no significant impact on the health care expenditure. Third, the average of absolute relative error calculated by our model was smaller than that of ARIMA model.