Chen R
Department of Applied Mathematics, Israel Institute for Biological Research, Ness-Ziona, Israel.
Stat Med. 1996;15(7-9):807-16. doi: 10.1002/(sici)1097-0258(19960415)15:7/9<807::aid-sim250>3.0.co;2-6.
A suspicion that the incidence of cancer is rising in a residential or workplace community is often a cause of alarm among members of that community. The seemingly excessive number of diagnoses encountered may he due to: (i) a known explanatory variable, such as ageing of the community members; (ii) random occurrence of an event which is rare under stable conditions; (iii) exposure to an undefined carcinogenic agent. An exploratory analysis is suggested in order to consider the support indicated by the data with respect to each of the two last possibilities while controlling the first. The analysis involves several procedures which differ with respect to the targeted temporal pattern of the diagnoses. These procedures are mainly based on the a priori probability that a diagnosis is made within the time interval observed between consecutive diagnoses. When the number of diagnoses is found to be excessively large, the temporal pattern of the diagnoses may indicate whether the event is likely to be an incidental occurrence or due to the effect of a carcinogen.
怀疑某居住社区或工作场所社区的癌症发病率正在上升，这往往会引起该社区成员的恐慌。所遇到的诊断数量看似过多可能是由于：（i）一个已知的解释变量，例如社区成员老龄化；（ii）在稳定条件下罕见事件的随机发生；（iii）接触不明致癌物质。建议进行探索性分析，以便在控制第一个因素的同时，考虑数据对后两种可能性中每一种的支持情况。该分析涉及几个程序，这些程序因诊断的目标时间模式而异。这些程序主要基于在连续诊断之间观察到的时间间隔内做出诊断的先验概率。当发现诊断数量过多时，诊断的时间模式可能表明该事件是偶然发生还是致癌物作用的结果。
