Calibration of Examiners for Chile National Children's Oral Health Survey

Objectives: To evaluate the inter- and intra-examiner agreement, reliability, and accuracy of examiners during the training and calibration process for caries and malocclusion detection in Chile's national children's oral health survey.
Methods: The training process began with theoretical instruction, incorporating online activities. Clinical pictures were evaluated during in-person sessions that included group discussions for feedback and calibration. An in vitro workshop was conducted to assess inter-examiner agreement for each outcome through practical exercises and evaluations of caries in donated teeth and malocclusions by two experts. For caries detection, intra- and inter-examiner reliability was evaluated first in adults and subsequently in children under 10 years old. Intra-examiner reliability was assessed on the same day by randomizing and re-identifying patient cases. The children’s assessment followed the International Caries Detection and Assessment System (ICDAS) combining initial scores and using frontal lighting without tooth drying to replicate school-based field conditions. For malocclusion detection, the workshop session utilized clinical pictures and stone casts. Analysis was performed using Stata 18®. We used kappa with quadratic weights for caries detection, while agreement proportions were determined for malocclusions.
Results: The training involved five dentists for caries detection and seven dentists for malocclusion. The assessments included 20 patients and 10 stone casts. For caries, agreement with the senior examiner ranged between 0.81 and 0.90, with weighted kappa values between 0.79 and 0.93, intra-examiner reliability ranged from 0.82 to 0.94, and inter-examiner reliability ranged from 0.86 to 0.93. For malocclusions, agreement with the standard examiner ranged between 0.70 and 1.00.
Conclusions: These results indicate that the training and calibration process achieved high levels of agreement and reliability among examiners. These findings emphasize the importance of structured training protocols to ensure data quality in national oral health surveys.