Interexaminer reproducibility of ordinal and interval-scaled plaque indices

Although many plaque scoring indices have been proposed for clinical or research purposes, only few of them are commonly used in the dental literature. Owing to the ordinal nature of these established indices, only nonparametric tests can be performed, while statistical analysts usually prefer parametric methodologies. Furthermore, these indices do not discriminate between the proximal and the broad surfaces of the teeth.

The aim of this study was to assess the interexaminer reproducibility for one established ordinal plaque index and two recently proposed interval-scaled indices.

Six subjects received a professional tooth cleaning and stopped all oral hygiene measures during 48-62 h. Accumulated plaque was disclosed and separately scored on all teeth except third molars by two examiners using the visual plaque index (VPI), the axial plaque extension index (APEI) and the proximal plaque extension index (PPEI). In total, 168 teeth (48 molars, 48 premolars, 24 canines and 48 incisors) were scored.

Statistically significant differences were found between the means of buccal VPI (p<0.05, Wilcoxon test), mesiobuccal, mesiolingual and distolingual APEI, and distobuccal, mesiolingual and distolingual PPEI (p<0.05, t-test). Spearman correlation values between the duplicate VPI measurements were 0.84 for buccal and 0.64 for lingual VPI, while Pearson correlation values varied between 0.51 and 0.83 for APEI and between 0.48 and 0.74 for PPEI. For VPI, 71% of the buccal (kappa=0.60) and 74% of the lingual (kappa=0.53) scores were identical. For APEI and PPEI scores, a linear regression was found with slopes ranging between 0.50 and 0.80. Mean measurement errors were 7.7% for buccal and 0.4% for lingual VPI, and ranged between 0.9% and 46.1% for APEI and PPEI.

The interexaminer reproducibility of all three indices showed fair to good agreement. Buccal VPI and buccal APEI in particular showed good agreement.