Bhat Meghashyam, Do Loc G, Roberts-Thomson Kaye
School of Dentistry, the University of Adelaide, Adelaide, South Australia, Australia.
Department of Public Health Dentistry, Sharavathi Dental College and Hospital, Shimoga, India.
J Investig Clin Dent. 2018 Aug;9(3):e12326. doi: 10.1111/jicd.12326. Epub 2018 Feb 9.
The aim of the present study was to determine the association between dental visiting and missing teeth using propensity score (PS) adjustment to control for confounding bias, and to compare the estimates with those obtained from traditional regression models.
A population-based study was conducted on adults aged 35-54 years in India. Multistage stratified cluster random sampling was used. Data were collected through interviews and oral examinations. The exposure factor was 'dental visiting', and the outcome was number of missing teeth. Sociodemographic factors, oral hygiene practices, periodontal disease, and caries experience were the covariates. Inverse probability weight (IPW) calculated from the PS for dental visiting from a logistic regression model was used to balance the covariates. The association between dental visiting and missing teeth was estimated from log-binomial regression models with and without using IPW.
Of the 873 participants, 77.7% visited a dentist. The ≥1 missing teeth prevalence was 65.3%. Post-IPW adjustment covariate standardized bias between groups with or without dental visit was lower than the pre-IPW adjustment. Those who visited a dentist had an adjusted prevalence ratio of 2.40 when IPW was used, and 2.03 when IPW was not used.
Dental visiting was strongly associated with missing teeth in this rural population.
本研究旨在利用倾向得分（PS）调整来控制混杂偏倚，确定看牙与缺牙之间的关联，并将估计值与传统回归模型得出的结果进行比较。
对印度35至54岁的成年人开展了一项基于人群的研究。采用多阶段分层整群随机抽样。通过访谈和口腔检查收集数据。暴露因素为“看牙”，结果为缺牙数量。社会人口学因素、口腔卫生习惯、牙周疾病和龋齿经历为协变量。使用从逻辑回归模型计算得出的看牙倾向得分的逆概率权重（IPW）来平衡协变量。通过使用和不使用IPW的对数二项回归模型估计看牙与缺牙之间的关联。
873名参与者中，77.7%看过牙医。≥1颗缺牙的患病率为65.3%。使用IPW调整后，看牙组与未看牙组之间的协变量标准化偏差低于使用IPW调整前。使用IPW时，看过牙医者的调整患病率比为2.40，未使用IPW时为2.03。
在这个农村人群中，看牙与缺牙密切相关。
