Department of Pediatric Dentistry, Universidade Luterana do Brasil, Canoas, Brazil.
Dent Traumatol. 2010 Apr;26(2):143-8. doi: 10.1111/j.1600-9657.2009.00858.x. Epub 2010 Jan 5.
This cross-sectional study aimed to investigate the factors associated with dental trauma in preschool children using Poisson regression analysis with robust variance.
The study population comprised 888 children aged 3- to 5-year-old attending public nurseries in Canoas, southern Brazil. Questionnaires assessing information related to the independent variables (age, gender, race, mother's educational level and family income) were completed by the parents. Clinical examinations were carried out by five trained examiners in order to assess traumatic dental injuries (TDI) according to Andreasen's classification. One of the five examiners was calibrated to assess orthodontic characteristics (open bite and overjet). Multivariable Poisson regression analysis with robust variance was used to determine the factors associated with dental trauma as well as the strengths of association. Traditional logistic regression was also performed in order to compare the estimates obtained by both methods of statistical analysis.
36.4% (323/888) of the children suffered dental trauma and there was no difference in prevalence rates from 3 to 5 years of age. Poisson regression analysis showed that the probability of the outcome was almost 30% higher for children whose mothers had more than 8 years of education (Prevalence Ratio = 1.28; 95% CI = 1.03-1.60) and 63% higher for children with an overjet greater than 2 mm (Prevalence Ratio = 1.63; 95% CI = 1.31-2.03). Odds ratios clearly overestimated the size of the effect when compared with prevalence ratios.
These findings indicate the need for preventive orientation regarding TDI, in order to educate parents and caregivers about supervising infants, particularly those with increased overjet and whose mothers have a higher level of education. Poisson regression with robust variance represents a better alternative than logistic regression to estimate the risk of dental trauma in preschool children.
本横断面研究采用稳健方差的泊松回归分析，旨在调查学龄前儿童牙外伤的相关因素。
研究人群为巴西南部卡诺亚市公立托儿所 888 名 3-5 岁儿童。父母填写了调查问卷，评估与自变量（年龄、性别、种族、母亲教育水平和家庭收入）相关的信息。由五名经过培训的检查者进行临床检查，以根据 Andreasen 分类评估创伤性牙外伤（TDI）。其中一名检查者负责评估正畸特征（开颌和前牙覆盖）。采用稳健方差的多变量泊松回归分析来确定与牙外伤相关的因素及其关联强度。还进行了传统的逻辑回归，以便比较两种统计分析方法的估计值。
36.4%（323/888）的儿童发生了牙外伤，3 至 5 岁之间的患病率没有差异。泊松回归分析显示，母亲受教育年限超过 8 年的儿童发生该结局的可能性增加近 30%（患病率比=1.28；95%CI=1.03-1.60），前牙覆盖大于 2mm 的儿童增加 63%（患病率比=1.63；95%CI=1.31-2.03）。与患病率比相比，优势比明显高估了效应的大小。
这些发现表明需要针对 TDI 进行预防指导，以便向父母和照顾者传授有关监督婴儿的知识，特别是那些前牙覆盖较大且母亲受教育程度较高的婴儿。稳健方差的泊松回归比逻辑回归更能估计学龄前儿童牙外伤的风险。
