本文考慮以配對病例對照鑑取抽樣 (matched case-control proband sampling) 來探討疾病的「家族聚集性 (familial aggregation) 」現象。當沒有共變量資訊時,我們推廣 Datta & Satten (2005) 群集資料的秩和檢定 (rank-sum test for clustered data),以無母數統計法檢視疾病是否存在家族聚集性。當進一步蒐集到共變量資料時,我們推廣 Liang & Pulver (1996) 的方法,透過模型的建立,以解析家族聚集性的起因:遺傳因素或是環境因素。針對推廣後的新方法,我們亦進行有限樣本的模擬研究。最後就阻塞型睡眠呼吸中止症候群 (Obstructive sleep apnea, OSA) 的虛擬資料進行實例分析。
In this paper, some methods are proposed for the analysis of matched case-control proband data arising from familial aggregation studies. When there is no covariate information, we generalized the rank-sum test for clustered data (Datta & Satten, 2005) to detect familial aggregation. If there are some cavariate informations, we apply the generalized method of Liang & Pulver (1996) to analyze the cause of familial aggregation : genetic or environmental factors. We apply the proposed methods to obstructive sleep apnea pseudodata. Some simulation studies with finite sample demonstrate the usefulness of our approach.