中国正畸专家对错畸形严重程度的主观判断一致性研究

Consistency of the subjective evaluation of malocclusion severity by the Chinese orthodontic experts

目的:分析中国正畸专家主观判断错畸形严重程度的一致性。方法:从参与中国正畸疗效评价标准研究的6所院校所提供的具有完整病例资料的完成正畸治疗的2 383例病例中以院校来源和安氏分类作为分层因素随机抽取120例病例,由全国69位正畸专家,根据病例治疗前的模型、头颅侧位片、面像、曲面断层片和病历基本资料主观判断每例错畸形的严重程度,选取轻度、中等偏轻、中等、中等偏重和重度5个严重程度级别之一作为每例的判断结果,并对主观判断结果进行专家自身可靠性和专家之间评价一致性的分析。结果:采用加权Kappa检验分析专家自身可靠性,8.33%的专家自身一致性达到极好水平(Kappa≥0.81);78.33%的专家自身一致性达到好及以上水平(Kappa≥0.61);96.67%的专家自身一致性达到中等及以上水平(Kappa≥0.41)。采用组内相关系数(intra-class correlation coefficient,ICC)检验分析专家之间评价一致性(r=0.989,P<0.01),专家之间评价一致性非常好。结论:中国正畸专家依据病例治疗前的模型、头颅侧位片、面像、曲面断层片和病例基本信息对错畸形严重程度进行主观判断的专家自身可靠性好,且专家之间评价一致性高,这为建立错畸形严重程度的客观分级系统奠定了良好的基础。

Objective： To assess the consistency of the subjective evaluation of malocclusion severity by the Chinese orthodontic experts.Methods： Sixty-nine Chinese orthodontic experts subjectively evaluated the malocclusion severity for 120 cases which were selected randomly from 6 University orthodontic clinics by checking each case＇s pretreatment records including study cast,lateral head film,panoramic radiograph,facial photographs and patient chart.Each orthodontist was asked to independently rate the severity of every case into five grades： mild,mildly moderate,moderate,severely moderate and severe.Rating data was finally gathered to evaluate the intra-judge＇s reliability and the inter-judges＇ consistency.Results： Weighted Kappa test revealed that 8.33% orthodontists showed excellent intra-judge＇s reliability（Kappa≥0.81）,78.33% orthodontists showed good intra-judge＇s reliability（Kappa≥0.61） and 96.67% specialists displayed general intra-judge＇s reliability（Kappa≥0.41）.And intra-class correlation coefficient demonstrated a high level of inter-judges＇ consistency（r=0.989,P0.01）.Conclusion： Good intra-judge＇s reliability and inter-judges＇ consistency can be demonstrated in the subjective evaluation of malocclusion severity by the Chinese orthodontic experts,which could the basis for establi-shing the objective grading system of malocclusion severity.