Comparison of peer assessment rating (PAR) index scores of plaster and computer-based digital models

Abstract

The peer assessment rating (PAR) index is a valid and reliable tool for measuring malocclusion on plaster models, but it has not been shown to be valid and reliable when used to score computer-based digital models. The purpose of this study was to determine whether the PAR index is a valid and reliable measure on digital models. The study sample consisted of 48 pairs of plaster and digital pretreatment models. One examiner, calibrated in the PAR index, scored the digital and plaster models. The overall PAR scores were examined for reliability and validity by using analysis of variance and the intraclass correlation coefficient (ICC). Reliability of the components of the PAR score was compared with values originally presented by Richmond et al (1992). No significant differences were found between overall PAR scores of plaster and digital models (P = .82), and scores were highly correlated (ICC = 0.95; lower confidence boundary (LCB) = 0.92; upper confidence boundary (UCB) = 0.97). Intraexaminer reliability was excellent for both plaster models (ICC = 0.98; LCB = 0.97; UCB = 0.99) and digital models (ICC = 0.96; LCB = 0.94; UCB = 0.98). Reliability of all components of the PAR score generated on digital models except for buccal occlusion was similar to those of Richmond et al. PAR scores derived from digital models are valid and reliable measures of occlusion.