Examining multilevel associations in dental research

描述（由申请人提供）：在2002年的美国，大约有780万50岁及以上的女性患有髋关节骨质疏松症，另外有2180万女性患有髋关节骨量低或骨质减少症，有患骨质疏松症的风险。大部分来自横断面研究的数据表明，骨质疏松症与牙齿脱落、牙槽骨脱落和牙周附着体丧失增加有关。调查牙周病进展和牙槽骨丢失的口腔健康研究通常涉及多层次、嵌套的测量，在牙齿水平上对受试者进行纵向跟踪。疾病过程可能在受试者的口腔内或在受试者体内随时间变化而变化，并且疾病过程可能在受试者之间不同。通常，回归建模方法不能区分口腔内慢性或患者水平的关联和随时间或空间变化的急性变化的关联。更好地了解多阶段疾病过程可以导致更有效的临床随访程序和更有效的治疗。我们的长期目标是减缓绝经后牙周炎骨质减少妇女牙槽骨丢失和临床附着丧失的进展。本建议的具体目标是：1)基于广义估计方程方法开发计算宏，可在SAS和R中实现，用于拟合和统计比较相关数据回归模型中单独的口腔内，跨时间和受试者之间的效应；2)通过模拟，制定研究设计指南，包括样本量，以确保有足够的能力检测不同的口腔内、跨时间和受试者之间的效应；3)将发展的回归分析方法应用于我们最近完成的国家牙科和颅面研究所（NIDCR）资助的（R01DE012872）随机双盲、安慰剂对照临床试验的现有数据，该试验使用亚抗菌剂量的多西环素（SDD）治疗绝经后骨质减少的中晚期牙周炎妇女。拟议的分析将利用现有的口腔放射学和临床牙周测量，以及血清炎症生物标志物和全身骨密度测量，来自128名参加为期2年的临床试验的妇女。次要数据分析的目的是确定2年内与牙槽骨丢失和牙周炎进展相关的口腔和全身因素，区分慢性或患者水平的影响以及口腔内随时间或空间变化的急性变化的影响。更好地了解与牙槽骨丢失和牙周炎进展相关的口腔和全身因素之间复杂的多层次关联，可能会导致新的临床随访方案和治疗方法，从而改善这一人群的口腔健康。

DESCRIPTION (provided by applicant): In the United States in 2002, approximately 7.8 million women aged 50 and over had osteoporosis of the hip and an additional 21.8 million had low bone mass or osteopenia of the hip and were at risk for developing osteoporosis. Data, largely from cross-sectional studies, suggest that osteoporosis is associated with increased tooth loss, alveolar bone loss, and periodontal attachment loss. Oral health research studies investigating progression of periodontal disease and alveolar bone loss often involve multilevel, nested measures at the tooth-site level within subjects followed longitudinally over time. Disease processes may vary within a subject's mouth or over time within a subject and disease processes may differ between subjects. Typically, regression modeling approaches do not distinguish among effects of chronic or patient-level associations and associations with acute changes over time or spatial variation within a mouth. A better understanding of the multilevel disease processes could lead to more effective clinical follow-up procedures and more effective therapies. Our long-term goal is to slow the progression of alveolar bone loss and clinical attachment loss in post-menopausal osteopenic women with periodontitis. The specific aims of this proposal are to: 1) Develop computing macros, based on generalized estimating equation methodology, that can be implemented in SAS and R for fitting and statistically comparing separate within-mouth, across-time, and between-subject effects in correlated data regression models; 2) Through simulation, develop guidelines for study design, including sample size, to ensure adequate power to detect different within-mouth, across-time, and between-subject effects; and 3) Apply the developed regression analysis methodology to existing data from our recently-completed National Institute of Dental and Craniofacial Research (NIDCR)-funded (R01DE012872) randomized double-blind, placebo-controlled clinical trial of sub-antimicrobial dose doxycycline (SDD) in postmenopausal osteopenic women with moderate to advanced periodontitis. The proposed analyses will utilize existing oral radiographic and clinical periodontal measures, as well as serum inflammatory biomarker and systemic bone mineral density measurements, from the 128 women who were enrolled in the 2-year clinical trial. The objectives of the secondary data analyses are to identify oral and systemic factors associated with alveolar bone loss and periodontitis progression over a 2-year period, distinguishing among effects of chronic or patient-level associations and associations with acute changes over time or spatial variation within a mouth. A better understanding of the complex, multilevel associations among oral and systemic factors associated with alveolar bone loss and periodontitis progression could lead to new clinical follow-up protocols and therapeutic approaches that will improve the oral health of this population. PUBLIC HEALTH RELEVANCE: Through the planned research project, statistical analysis and research design tools will be developed to investigate factors associated with oral disease processes that may vary within a given subject's mouth from tooth to tooth, over time as a subject ages or undergoes therapy, or between subjects with different characteristics. The methods will be used to identify oral and systemic factors associated with oral bone loss and periodontitis progression over a 2-year period among post-menopausal osteopenic women with moderate to advanced periodontitis. A better understanding of the oral disease processes could lead to more effective clinical follow-up procedures and more effective therapies in this type of population.