Statistical Methods for Complex Dependent Dental Data

描述（由申请人提供）：尽管有大量的统计分析来评估初级牙列龋齿的模式，但仍有一些基本问题尚未得到解答。例如，牙医普遍认为口腔内的龋齿具有空间对称性，但这从未在统计学意义上得到证实。比利时法兰德斯的Signal Tandmobiel项目提供了一个独特的数据库，它提供了牙齿层面有关龋齿的详细数据，可以帮助回答这些问题。这个建议是一个非常广泛的研究，包括横向和纵向调查的口腔卫生在佛兰德学童。其目的是评估1989年出生的儿童口腔卫生的决定因素，并评估每年在随机选择的学校进行口腔卫生教育的效果。由于口腔内表面的固有空间关联以及同一表面随时间产生的重复测量，对这些数据进行牙齿表面水平分析带来了许多困难。本提案的目的是发展一种统计方法，用于从广泛的流行病学研究中产生的龋齿数据的二次分析。具体而言，我们建议(1)建立时空二值数据的统计模型，并建立应用于乳牙龋齿的多元时间-事件数据模型，用于回答牙齿表面对龋齿经验的易感性、口腔中龋齿的对称性以及不同龋齿风险群体表面易感性的差异等相关问题；（二）监督拟建统计模型的质量；(3)研究缺失数据机制对所提出回归模型的影响。这项统计研究的一个关键目标是提高我们设计、分析和解释牙科研究中临床试验和流行病学研究中产生的复杂依赖数据的能力。

DESCRIPTION (provided by applicant): Despite numerous statistical analyses to assess the pattern of dental caries in primary dentition, there are still some fundamental questions that remain unanswered. As an example, it is well believed among dentist that there are spatial symmetries in the mouth with respect to caries, but this has never been shown in a statistical sense. A unique database that provides detailed tooth-level data with respect to dental caries, which could help answer some of these questions, is the Signal Tandmobiel project in Flanders (Belgium). This proposal is for a very extensive study that comprises both cross-sectional and longitudinal investigations of oral hygiene among Flemish schoolchildren. Its objectives are to assess determinants of oral hygiene among children born in 1989 and to assess the effect of yearly oral hygiene education conducted at randomly selected schools. A tooth-surface level analysis of this data poses a number of difficulties due to inherent spatial association of surfaces in the mouth and repeated measurements generated from the same surface over time. The objective of this proposal is to develop a statistical methodology for secondary analysis of dental caries data generated from this extensive epidemiologic study. Specifically, we propose (1) to develop statistical models for spatial-temporal binary data and develop models for multivariate time-to-event data with applications to dental caries in primary teeth, which will be used to answer questions related to tooth surface susceptibility to caries experience, symmetries in the mouth with respect to dental caries, and differences in surface susceptibility according to caries risk groups; (2) to monitor the quality of the proposed statistical models; and (3) to investigate the influence of the missing data mechanism on the proposed regression models. A key goal of this statistical study is to improve our ability to design, analyze, and interpret complex dependent data generated from clinical trials and epidemiologic studies in dental research.