Propriedades psicométricas da versão brasileira do questionário Rapid Estimate of Adult Literacy in Dentistry (BREALD-30) aplicado em adolescentes

Abstract

The objective of the study was to evaluate the psychometric properties of the Rapid Estimate of Adult Literacy in Dentistry (BREALD-30) instrument in Brazil for the ages of 12 and 15 to 19 years. It was a cross-sectional comprised by 375 adolescents for each age group from public and private schools in a city in Northeastern Brazil. The parent/caregiver responded to a sociodemographic questionnaire. Two previously calibrated researchers applied the BREALD-30 and performed the clinical exams (dental caries). Kappa coefficients for the BREALD-30 word-by-word agreement between the gold standard and researchers were 0.889 and 0.884. Agreement between both researchers was 0.870 and intra-examiners were 0.898 and 0.871. Intraclass correlation coefficient (ICC) measured the BREALD-30 total score for inter-examiners (0.987; CI 95%: 0.970-0.995 and 0.874; CI 95%: 0.860-0.895) and intra- examiners (0.973; CI 95%: 0.921-0.991 and 0.994; CI95%: 0.982-0.998) agreement. For the clinical exam, dental caries was diagnosed according to the Nyvad’s classification criteria. Kappa values between the gold standard and both examiners were 0.894 and 0.936, while inter-examiner agreement was 0.890 and 0.880. Reliability was measured through internal consistency and test-retest reliability. BREALD-30 scores were classified in high, moderate and low, according to tertiles. The convergent validity was assessed through the correlation between BREALD-30 and the Functional Literacy Indicator (INAF) scores using the Spearman Correlation Coefficient (p <0.05). Divergent validity was evaluated by comparing BREALD-30 scores between sociodemographic variables, using the Chi-square and Chi- Square test with linear trend (p <0.05). For predictive validity, it was tested the association between oral health literacy (OHL) level, through BREALD-30 scores, with the presence of active and/or inactive dental cavitation through a multiple logistic regression model. It was observed that for the age of 12 years the BREALD-30 presented a mean score of 18.5 (SD = ±5.46), while for the 15-19 years-old group the mean score was 20.8 ((SD = ±4.7). For the age of 12 years and for the age group of 15 to 19 years the instrument presented good internal consistency, with Cronbach's alpha of 0.871 and 0.834, respectively. The instrument showed good test-retest reliability (CCI = 0.898 and 0.974; Kappa = 0.804 and 0.808), and item-total correlation of all items was satisfactory. The instrument had convergent validity with INAF for 12 years (rs = 0.558, p <0.001) and for 15 to 19 years (rs = 0.652, p <0.001). Participants with higher OHL levels were from private school (p <0.001), belonging to Economic Classes A and B1 (p <0.001), and had parents with higher education (p <0.001), indicating the divergent validity of the instrument. Regarding predictive validity, adolescents with low OHL were more likely to have cavitated dental caries, for both the 12 year-old (OR = 2.37, 95% CI: 1.48-3.80; p< 0,001) and for the 15 to 19 years group (OR = 1.96, 95% CI: 1.24-3.11; p=0,004). It is concluded that the BREALD-30 instrument showed satisfactory psychometric properties for the age groups analyzed, and can be applied as a fast, simple and reliable measure of oral health literacy.