The present study aims at describing the assessment of the reliability and accuracy of a survey of dental caries performed in the city of São Paulo in 2002, as part of the SB2000 Project of the Ministry of Health. The appraisal of indexes used methodology internationally standardized by the World Health Organization. During the training of examiners participating in the survey, an exercise of calibration was conducted for the gauging of inter-observer agreement, which involved nine dental surgeons and two groups of 20 schoolchildren. The coordinator of the survey was assigned as the gold Standard for the validation of the oral examination records in the calibration process. Furthermore, during the gathering of data, 58 schoolchildren were reexamined by the same professionals, for the assessment of intra-observer agreement. We calculated high synthetic indicators of reliability for the survey: respectively 99.83% and 96.45% for the percentage of overall intra- and inter-observer agreement, and 0.996 and 0.943 for the corresponding values of kappa statistics. The validity of the study was also evaluated positively, with high global indicators of sensitivity (0.972) and specificity (0.978). The evaluation of reliability and accuracy for each observer produced high and satisfactory indicators. The description of the procedures chosen for the assessment of reliability and accuracy, as well as the presentation of results, can provide methodological indications for subsequent studies of dental caries distribution.
Dental caries; Data collection; Inter-observer variation; Intra-observer variation
