目標：以內政部與衛生署死亡資料連結進行比對，探討兩單位死亡資料之差異，以為研究者使用該資料庫與相關單位改善資料庫品質之參考。方法：以內政部統計與衛生署統計公布之歷年(民國71-91年)死亡人數統計比較人數差異之趨勢，並連結民國88年內政部死亡檔與衛生署死因檔比較實際差異之情形。結果：若以歷年內政部戶政司與衛生署統計室公布死亡人數統計做比較，可發現各年度檔案人數之符合率皆高達98%以上。但若以實際資料逐筆連結比對，則88年死因檔符合率由內政部與衛生署統計死亡人數的98.7%降為95.09%，內政部與衛生署檔案中未連結上之人數分別為6,203人與4,747人。進一步分析發現不同年齡別與地區別之連結率有顯著差異。結論：本研究發現內政部與衛生署公佈之死亡人數總數差異約為2%，逐筆連結比對後差異約為4%，此差異會隨死亡年齡增加而減少且有地區性差異。建議未來使用者應小心年齡與地區差異。關於造成此比對差異的原因與政策意涵，有待未來研究進一步釐清。

　　　　　Objectives: To explore the difference between the datasets of the death registration of the Department of Health (DOH) and Ministry of Interior (MOI) for the reference of the researchers in the future. Methods: We used the annual number of deaths published by the MOI and DOH during 1982-2002 to analysis the difference; and linking the datasets of the death registration of MOI and DOH, published in 1999, by ID number to verify the actual difference in the number of cases. Results: The concordant rate of number of deaths published by DOH and MOI was 98.7%. When the datasets were linked by ID number, the concordant rate was dropped to from 98.7% to 95.9%. There were 6,203 cases from the MOI dataset and 4,747 cases from the DOH dataset not matched by ID number. The rate of mismatch varied by age and area. Conclusions: The difference in the number of death published by DOH and MOI was about 2%. When the datasets were linked by each ID number, the difference in the number of death was increased to 4%. The concordant rate between the datasets of DOH and MOI varied by age and area. When using the dataset of death registration from the DOH or MOI for study, researchers should be aware of the variance. The causes of the difference and policy implication of this should be investigated in further study.