Goldberg M S, Carpenter M, Thériault G, Fair M
Département de santé communautaire, Hôpital du Sacré-Coeur de Montréal, Québec.
Can J Public Health. 1993 May-Jun;84(3):201-4.
Vital status of a cohort of 10,211 Quebec, synthetic textiles workers was ascertained through a probabilistic record linkage to the Canadian Mortality Data Base (CMDB); 5,033 of these workers were also traced using other sources. There was agreement in the vital status of all but 60 of the subjects traced jointly through the CMDB and the alternate sources. 41 subjects were declared 'deceased' from the CMDB but 'alive' from the alternate sources; it is likely that these subjects were indeed deceased. 19 subjects, declared 'deceased' with a fair degree of certainty from the alternate sources, were not identified from the computer search of the CMDB; 17 were found manually on the microfiche death records and two died outside of Canada. The probability of identifying deceased and living subjects from the CMDB was therefore estimated to be 98.2% (95% confidence interval: 97.5-98.7%) and about 100%, respectively. Estimates of cost are also presented, and it is concluded that use of the CMDB is the method of choice for tracing moderate to large cohorts.
通过与加拿大死亡率数据库（CMDB）进行概率性记录链接，确定了10211名魁北克合成纺织品工人队列的生命状态；其中5033名工人还通过其他来源进行了追踪。通过CMDB和其他来源共同追踪的所有受试者中，除60名外，其余受试者的生命状态均一致。41名受试者在CMDB中被判定为“死亡”，但在其他来源中被判定为“存活”；这些受试者很可能确实已经死亡。19名受试者在其他来源中被相当确定地判定为“死亡”，但在对CMDB的计算机搜索中未被识别；其中17名在缩微胶片死亡记录中手动找到，另外两名在加拿大境外死亡。因此，从CMDB中识别死亡和存活受试者的概率估计分别为98.2%（95%置信区间：97.5 - 98.7%）和约100%。文中还给出了成本估计，并得出结论，使用CMDB是追踪中大型队列的首选方法。
