Evaluation of the proportionate mortality index in the presence of multiple comparisons.
Evaluation of the proportionate mortality index in the presence of multiple comparisons.
复制标题
在多重比较的情况下评估比例死亡率指数。
DOI：
10.1002/sim.4780060109
复制
发表时间：
1987
影响因子：
2
通讯作者：
V. Arena
中科院分区：
文献类型：
--
作者：
H. Rockette;V. Arena
我们调查了职业研究中在多重比较的情况下错误地识别显着升高的特定原因死亡率的频率。当调查一组标准死因时，使用标准化死亡率作为汇总指标会导致实验错误率过高，但可以通过使用限制第一类错误的标准程序来控制错误率。当职业队列的主要疾病类别的发生率低于预期时，使用比例死亡率作为汇总指标时，控制第一类错误的标准程序是不够的。在这种情况下，使用比例死亡率作为汇总指标可能会导致错误地识别出过高的特定原因死亡率，这种情况比以前承认的要频繁。
We investigate the frequency in occupational studies of falsely identifying significantly elevated cause-specific mortality in the presence of multiple comparisons. When one investigates a standard battery of causes of death, use of a standardized mortality ratio as a summary index results in an excessive experimental error rate, but one can control the error rate with use of standard procedures that limit the type I error. When a major disease category for an occupational cohort has a lower rate than expected, then with use of a proportionate mortality ratio as a summary index, the standard procedures for control of the type I error are inadequate. In this situation, use of a proportionate mortality ratio as a summary index is likely to result in the false identification of an excess cause-specific mortality more often than previously acknowledged.