Greenwald P, Rose J S, Daitch P B
Am J Epidemiol. 1979 Aug;110(2):162-77. doi: 10.1093/oxfordjournals.aje.a112801.
A case-control study of clustering through acquaintanceship among lymphoma and leukemia patients was conducted for the years 1967 through 1972 in Orleans County, New York. Twenty lymphoma and 17 leukemia cases met criteria for inclusion in the study. Data on acquaintanceship linkage were gathered from the source cases and controls, and from their acquaintances yielding a data base of 13,409 unique individuals linked by acquaintance. Three different analyses were carried out: a statistical analysis of linkage via intermediaries of case pairs in comparison with control pairs; a computer simulation of disease transmission from selected source cases to selected targets and controls based on the acquaintance data; and a secondary attack rate type analysis. The first two types of analysis yielded statistically significant case-control differences at the .05 level. The third method also yielded a positive result but was not subject to quantitative hypothesis testing. An additional conclusion is that these epidemiologic methods for disease with long induction periods merit further study.
1967年至1972年期间，在纽约奥尔良县对淋巴瘤和白血病患者通过熟人关系形成的聚集现象进行了一项病例对照研究。20例淋巴瘤患者和17例白血病患者符合纳入该研究的标准。关于熟人关系的数据是从源病例和对照以及他们的熟人那里收集的，从而形成了一个由13409个通过熟人关系联系起来的独特个体组成的数据库。进行了三种不同的分析：对病例对与对照对通过中介进行联系的统计分析；基于熟人数据对选定源病例到选定目标和对照的疾病传播进行计算机模拟；以及二代发病率类型分析。前两种分析类型在0.05水平上产生了具有统计学意义的病例对照差异。第三种方法也得出了阳性结果，但未进行定量假设检验。另一个结论是，这些针对具有长诱导期疾病的流行病学方法值得进一步研究。
