Multilevel calibration procedure for the oral health national multicenter survey in primary teeth

Abstract

Background: Early childhood caries (ECC) requires systematically collected and standardized data.

Aim: To describe a novel multilevel calibration procedure in primary dentition.

Design: Calibration method involved two calibration levels: the first (L₁ ) involved an interexaminer agreement between three main investigators, the group leaders (GLs) in the following level; the second level (L₂ ) involved three groups of 11 paediatric dentists and interexaminer agreement assessment according to the GLs in each group. The study sample consisted of 650 primary teeth surfaces in eight children (mean age 6.56 ± 2.22 years). Surface-by-surface percent agreement, tooth-by-tooth percent agreement, Cohen's kappa, and Fleiss kappa statistics were used to calculate interexaminer reliability. Statistical analysis was performed with SPSS 27.0.

Results: Surface-by-surface percent agreement regarding ICDASepi-merged revealed almost perfect agreement (>90.00%) on both L₁ and L₂ . Kappa values and ranges showed good agreement at both L₁ (overall κ = .95) and L₂ (overall κ = .98) and almost perfect consistency was detected between GLs at L₁ (>91.30%) and substantial agreement at L₂ (>85.00%). All examiners at L₂ showed almost perfect positive agreement (sensitivity = 96.77%-100%) when detecting the presence of dental plaque.

Conclusion: The calibration procedure appeared feasible prior to organizing multicenter epidemiological oral health survey in large population groups of preschool children, with higher number of examiners.

Keywords: calibration; early childhood caries; epidemiology; reliability.