本文檢驗我國於1995年實施全民健康保險對醫療服務使用量的影響。由於在全民健保前,有許多民眾沒有健康保險,我們比較實施全民健保前後,這些原本沒有健康保險的民眾在門診次數及住院天數的變化,而為控制同時發生但與全民健保無關的因素,我們以公保為控制組使用'difference-in-differences'、採用二部模型(two-part model)以檢驗醫療服務使用的變化是存在於民眾的看診決策、醫生的治療決策,或二者同時存在。實證結果顯示全民健保實施後,醫療服務使用量大幅增加。具體而言,民眾在健保實施後,住院及門診的可能性均顯著增加,對有住院、門診紀錄的民眾而言,平均住院天數增加、但平均門診次數減少。換言之,實施全民健保後,原先沒有保險的民眾無論是住院可能性及住院天數均顯著增加;但雖然門診的機率增加,平均門診次數卻減少,顯示健保增加許多使用少量門診次數的病患。
This paper investigates the impact on medical consumption (hospitalization days and physician visits) of the 1995 introduction of Taiwan's National Health Insurance (NHI). Using the 'difference-in-differences' approach to control for unobservable effects, we examine the post-NHI rise in medical consumption among the previously uninsured. Our two-part model enables us to determine whether this increase occurred because of a rise in first-time consumption, a rise in treatment intensity, or both. We present empirical evidence that NHI led to a substantial increase in consumption. The likelihood of first-time hospitalization and physician visits increased post-NHI. Conditional on nonzero consumption, the number of hospitalization days increased, but the number of physician visits decreased. Post- NHI, a larger proportion of consumers visited physicians infrequently. Therefore, NHI seems to have had an ambiguous effect on average consumption.