Williams A N, Johnson R A, Bender A P
Section of Chronic Disease and Environmental Epidemiology, Minnesota Department of Health, Minneapolis 55440.
Am J Epidemiol. 1990 Jul;132(1 Suppl):S178-82. doi: 10.1093/oxfordjournals.aje.a115780.
In spite of their limitations, mortality data are used in many epidemiologic and public health settings. In this investigation, the authors examined the extent to which community cancer mortality rates were affected by incorrect reporting or coding of residence on death certificates. Observed and expected cancer mortality for two adjacent communities in northern rural Minnesota for the periods 1970-1974 and 1980-1984 were obtained from computerized state mortality data. Using statewide rates to obtain expected values, standardized mortality ratios for total cancers for both periods combined were 138 for men (101 observed deaths) and 148 for women (86 observed deaths). These excesses were statistically significant (p less than 0.05). However, after review of data from the actual death certificates, city maps, and information from city officials, 44 of the 187 total cancer deaths (24%) were found to have had an incorrectly reported or coded residence status. After removal of these cases, the standardized mortality ratio for total cancers for males went from 138 to 107, and for females the standardized mortality ratio went from 148 to 111. No standardized mortality ratios remained statistically significant. These findings may have implications for those who use mortality data for assessing cancer rates in communities in rural areas.
尽管存在局限性，但死亡率数据仍被用于许多流行病学和公共卫生领域。在这项调查中，作者研究了死亡证明上居住地址的错误报告或编码对社区癌症死亡率的影响程度。1970 - 1974年和1980 - 1984年期间，明尼苏达州北部两个相邻社区的观察到的和预期的癌症死亡率是从计算机化的州死亡率数据中获得的。使用全州的比率来获得预期值，两个时期合并的所有癌症的标准化死亡率男性为138（观察到101例死亡），女性为148（观察到86例死亡）。这些超额死亡率具有统计学意义（p小于0.05）。然而，在审查了实际死亡证明、城市地图以及城市官员提供的信息后，发现187例癌症死亡病例中有44例（24%）的居住地址报告或编码有误。去除这些病例后，男性所有癌症的标准化死亡率从138降至107，女性的标准化死亡率从148降至111。没有标准化死亡率仍具有统计学意义。这些发现可能对那些使用死亡率数据评估农村地区社区癌症发病率的人有启示。
