Schnatter A R, Acquavella J F, Thompson F S, Donaleski D, Thériault G
Dept. of Occupational Health and Epidemiology, Exxon Biomedical Sciences, Inc., East Millstone, NJ 08875-2350.
Can J Public Health. 1990 Jan-Feb;81(1):60-5.
The Canadian Mortality Data Base (MDB) was utilized in mortality follow-up of 17,446 refinery and petrochemical workers throughout Canada. The performance of the system in detecting 757 previously known deaths was evaluated. Of the deaths submitted, 93.1% (90% confidence interval (CI) 91.5-94.5) were detected, including 97.6% (90% CI 96.5-98.4) of deaths which occurred in Canada. Detection was generally unaffected by age, year of death (1964-1973), and the presence of middle initial or the remainder of first name. Slightly lower ascertainment was found for deaths occurring in Quebec (94.5%) and Newfoundland (93.3%). This could be due to several factors, including fewer MDB records with complete identifier information for these provinces, or lower accuracy of linking fields for records supplied to the Agency. Few links were made to employees presumed to be alive, indicating 99.8% specificity for these records.
加拿大死亡率数据库（MDB）被用于对加拿大各地17446名炼油和石化工人进行死亡率随访。评估了该系统检测757例已知死亡病例的性能。在提交的死亡病例中，93.1%（90%置信区间（CI）为91.5 - 94.5）被检测到，其中包括在加拿大发生的死亡病例中的97.6%（90% CI为96.5 - 98.4）。检测结果通常不受年龄、死亡年份（1964 - 1973年）以及中间名首字母或名字其余部分的影响。在魁北克省（94.5%）和纽芬兰省（93.3%）发生的死亡病例的确定率略低。这可能是由于多种因素造成的，包括这些省份具有完整标识符信息的MDB记录较少，或者提供给该机构的记录的链接字段准确性较低。与被假定为活着的员工建立的链接很少，表明这些记录的特异性为99.8%。
