Liu Siqi, Oh Heesoo, Chambers David William, Xu Tianmin, Baumrind Sheldon
First clinical division, Peking University School and Hospital of Stomatology, Beijing, China.
Department of Orthodontics, University of the Pacific, Arthur A. Dugoni School of Dentistry, San Francisco, CA, USA.
Eur J Orthod. 2018 Apr 6;40(2):157-163. doi: 10.1093/ejo/cjx043.
Determine optimal weightings of Peer Assessment Rating (PAR) index and Discrepancy Index (DI) for malocclusion severity assessment in Chinese orthodontic patients.
Sixty-nine Chinese orthodontists assessed a full set of pre-treatment records from a stratified random sample of 120 subjects gathered from six university orthodontic centres. Using professional judgment as the outcome variable, multiple regression analyses were performed to derive customized weighting systems for the PAR index and DI, for all subjects and each Angle classification subgroup.
Professional judgment was consistent, with an Intraclass Correlation Coefficient (ICC) of 0.995. The PAR index or DI can be reliably measured, with ICC = 0.959 and 0.990, respectively. The predictive accuracy of PAR index was greatly improved by the Chinese weighting process (from r = 0.431 to r = 0.788) with almost equal distribution in each Angle classification subgroup. The Chinese-weighted DI showed a higher predictive accuracy, at P = 0.01, compared with the PAR index (r = 0.851 versus r = 0.788). A better performance was found in the Class II group (r = 0.890) when compared to Class I (r = 0.736) and III (r = 0.785) groups.
The Chinese-weighted PAR index and DI were capable of predicting 62 per cent and 73 per cent of total variance in the professional judgment of malocclusion severity in Chinese patients. Differential prediction across Angle classifications merits attention since different weighting formulas were found.
确定同伴评估评分（PAR）指数和差异指数（DI）在中国正畸患者错颌畸形严重程度评估中的最佳权重。
69名中国正畸医生对从6个大学正畸中心收集的120名受试者的分层随机样本的全套治疗前记录进行了评估。以专业判断作为结果变量，进行多元回归分析，为所有受试者以及每个安氏分类亚组得出PAR指数和DI的定制权重系统。
专业判断具有一致性，组内相关系数（ICC）为0.995。PAR指数或DI能够可靠测量，ICC分别为0.959和0.990。通过中国权重过程，PAR指数的预测准确性得到了极大提高（从r = 0.431提高到r = 0.788），在每个安氏分类亚组中分布几乎相等。中国加权DI显示出更高的预测准确性，在P = 0.01时，与PAR指数相比（r = 0.851对r = 0.788）。与安氏I类（r = 0.736）和III类（r = 0.785）组相比，安氏II类组表现更好（r = 0.890）。
中国加权PAR指数和DI能够预测中国患者错颌畸形严重程度专业判断中总方差的62%和73%。由于发现了不同的加权公式，不同安氏分类之间的差异预测值得关注。
