目標：本研究旨在藉由時間序列分析模式，探討1999年8月實施之新制部分負擔與2000年1月實施之醫院合理門診量的實施，對於全民健保門診人次、門診申報費用及總醫療申報費用的影響。方法：主要採用時間序列模式，以1995年3月至2001年12月的健保按月申報之使用量急醫療費用進行分析。結果：時間序列經選擇適合模式後，除門診人次外，R square都達九成以上；評估兩個重要政策介入點對時間序列趨勢之影響，其中新制部分負擔實施有遞延效應。結論：在描述性資料方面大抵與文獻結果一致，而時間序列模式在2002年的預測點估計亦與實際值接近。從介入模式分析結果發現，單純就政策介入發生該月而言，節制醫療提供者之醫院合理門診量，比節制消費者之部分負擔，對於控制醫療費用之效果較為顯著，不過一年的推估結果則有不同的效果。

Objective: The purpose of this study was to use time series analysis methods to examine the impact on cases of ambulatory care, claims of ambulatory care and the grand total of medical claims when co-payment and regressive physician payment policies were implemented. Methods: Based on the data from the NHI from Mar. 1995 to Dec. 2001, time series models were used to formulate the time series. Results: By selecting the appropriate time series models, the R squares of these models were over 90 percent except for the model of ambulatory care cases. The two interventions including co-payment and regressive physician payment policy were used in the intervention models. The result showed that co-payment intervention had a delayed effect of one month. Conclusions: The results were similar with that of existing papers in terms of descriptive analyses. The outcomes of our models were also reliable in the point estimation in the forecast of 2002 expenditure. The outcome of intervention models found that regressive physician payment policy had a greater impact than the co-payment policy in the month when the policy was implemented, but the predication of one year using the time series models produced different results.

時間序列分析 ； 全民健保 ； 部分負擔 ； 合理門診量

Time series analysis, NHI, co-payment, regressive physician payment