死亡回顾调查的质量及其评价

利用死因登记报告资料进行模拟的死亡回顾调查以研究评价其误差来源及质量。结果表明，回顾调查的死因（无系统误差）、死龄（误差≤1岁）及死亡日期（＜1年）的符合率分别为79．69％，其它疾病为21．05％，意外死亡为26．32％，由于误差的正负抵消作用，这四种分类的最终统计错报率为4．65％、15．64％、14．47％及10．53％；归类统计数、年龄组分布及时间分布与登记资料基本一致，其差异均无显著性。本研究表明死亡回顾调查误差甚大，但群体资料仍能大致反映疾病的分布规律、流行趋势与程度。死亡回顾调查可作为获取疾病历史资料的手段之一，但对于资料的应用及评价须留有余地。

Quality of Retrospective Survey of Deaths and Its Evaluation

An imitative retrospective survey (RS) of deaths was carried out to evaluate the errors and quality of the retrospective survey by using the data on Qidong Death-Cause Registry. Results showed that the agreement rates for death-cause, death-age and death-date of RS were 79.69%, 67.97% and 58.98%, respectively. The total agreement rate for the three items was 31.95%, in which 52.71% in liver cancer, 25.00%, other cancers, 21.05%, other diseases, and 26.32%, accidents. Because some of the errors were cancelled out each other, the mis-report rates for the above diseases were 4.65%, 15.63%, 14.47%, and 10.53%, respectively, and the statistical breakdown of data, age-group and death-time distributions met basically with the data on registration (P greater than 0.05). This study indicates that the errors of the data from RS are rather higher compared with data on registration, yet the data on the whole can reflect the population distribution, time trends and epidemic intensity of diseases investigated, RS might be considered as a mean of getting historical data, but its application and evaluation should be cautious.