目標：本研究的目的有兩個：(1)估計1996年到2001年之間的全民健康保險兒童預防保健服務年度利用率以及世代利用率；(2)探討兩種估計值的差異。方法：我們利用國家衛生研究院所建構之全民健康保險承保抽樣歸人檔來進行估計，並分析兩種估計值的差異來源。結果：從1996年到2000年之間，零至一歲兒童之四次預防保健服務的年度總體利用率由28.26%上升到59.29%。從1996年到2001年之間，一至三歲之兒童和三至四歲之兒童的年度利用率則分別由35.59%上升至79.16%，以及由10.58%上升至26.70%。世代利用率的趨勢大致和年度利用率相似。結論：全民健康保險兒童預防保健服務的利用情形從1996到2001年之間逐漸改善，利用率增加為兩倍以上，不過仍有明顯的改善空間，尤其是三至四歲之兒童預防保健服務利用水準。年度利用率對於預測尚未完整觀察到之世代利用率有很大的參考價值，不過做推估時應注意世代利用率之變化、出生人口數之變化以及不同世代使用服務時程之變化所導致的估計值誤差。

　　　　　Objectives: The two purposes of this study were: (1) to estimate the yearly utilization rates and the cohort utilization rates of preventive health services provided for children by the National Health Insurance Program (NHIP) for the period from 1996 to 2001; and (2) to explore differences between these two kinds of estimates. Methods: We estimated and analyzed the differences based on a National Health Insurance Database that the National Health Research Institutes constructed for a random sample of the population in the NHIP. Results: From 1996 to 2000, the yearly total utilization rate of preventive health services for infants increased from 28.26% to 59.29%. From 1996 to 2001, the yearly utilization rates for children aged 1-3 and 3-4 increased from 35.59% to 79.16% and from 10.58% to 26.70%, respectively. The trends of the cohort utilization rates were similar to those of the yearly utilization rates. Conclusions: The utilization level of preventive health services provided for children by the NHIP gradually improved from 1996 to 2001, with an overall increase of more than 100%. However, there is still much room for improvement, especially for the utilization level of children aged 3-4. The yearly utilization rate can be a good reference for predicting the cohort utilization rate of a cohort that has not been completely observed. Nevertheless, when making relevant inferences, we should pay attention to estimation errors due to changes in the cohort utilization rates, the number of births, and the distribution of timing for receiving preventive health services for children.