Validação para língua portuguesa do Brasil e desenvolvimento da versão reduzida do Hong Kong Oral Health Literacy Assessment Task for Pediatric Dentistry ( HKOHLAT-P)

• Ano de defesa: 2019
• Autor(a) principal: Ramon Targino Firmino
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; Programa de Pós-Graduação em Odontologia; UFMG
• Palavras-chave em Português: Alfabetização em saúde; Criança; Estudos de validação; Saúde bucal; Avaliação de programas e instrumentos de pesquisa; Inquéritos e questionários; Tradução; Cárie dentária
• Link de acesso: http://hdl.handle.net/1843/31156

The aim of the study was to cross culturally adapt and validate for the Brazilian Portuguese language the Hong Kong Oral Health Literacy Assessment Task for Paediatric Dentistry (HKOHLAT-P) questionnaire and to develop a reduced version of the instrument. The cross-cultural adaptation comprised translations, unification of translations, back-translation, analysis by an expert panel, pre-test, discussion by the expert panel and production of the final version in Portuguese. Two hundred parents/caregivers and preschoolers aged 3 to 5 years old from Campina Grande, Paraíba, participated in the study. One of the child’s parent/caregiver answered to the Brazilian version of the HKOHLAT-P (BHKOHLAT-P), a sociodemographic and reading habits questionnaire, the Early Childhood Oral Health Impact Scale (BECOHIS) and the Brazilian Rapid Estimate of Adult Literacy in Dentistry (BREALD-30) instrument. BHKOHLAT-P was answered a second time by all participants, after a 15 days-interval, to assess test-retest reliability. Children were examined for the diagnosis of dental caries (ICDAS-II) by a calibrated examiner (K = 0.83-0.99). The reliability of the instrument was measured by internal consistency and test-retest reliability. Exploratory (EFA) and confirmatory factor analysis (CFA) evaluated the dimensionality of the instrument. Convergent, divergent, concurrent and predictive validity were evaluated. Poisson regression models tested the effect of family income and education on the association between BHKOHLAT-P and exploratory variables (p<0.05). Item response theory was employed to reduce the number of items of the instrument, using item characteristic curves, discrimination and difficulty parameters. The reduced version had its dimensionality assessed by FCA; its reliability and validity tested. The instrument showed excellent internal consistency (Cronbach’s alpha = 0.93) and testretest reliability (ICC = 0.95). EFA and CFA showed the instrument to be unidimensional. There was a positive correlation between BHKOHLAT-P and BREALD-30 scores (r = 0.71; p<0.001) (convergent validity), the number of years of schooling (r = 0.60; p<0.001) and hours spent reading (r = 0.34; p<0.001) (concurrent validity). There was a negative correlation between BHKOHLAT-P and B-ECOHIS scores (r = - 0.23; p=0.001) and with number of cavitated teeth (r = - 0.15; p=0.026) (predictive validity). At the multivariate model, BHKOHLAT-P scores were not associated with B-ECOHIS scores, number of cavitated teeth or with the presence of caries (divergent validity). The reduced version comprised 30 items (BHKOHLAT-P30). CFA showed the reduced instrument to be unidimensional. BHKOHLAT-P-30 had excellent internal consistency (Cronbach’s alpha = 0.91) and test-retest reliability (CCI = 0.95). BHKOHLAT-P-30 scores were correlated with BREALD-30 scores (r = 0.71; p<0.001), number of years of schooling (r = 0.60; p<0.001) and hours spent reading (r = 0.33; p<0.001). There was a negative correlation between BHKOHLAT-P-30 and BECOHIS scores (r = -0.21; p=0.002) and with number of cavitated teeth (r = -0.18; p=0.010). In conclusion, the long and reduced versions showed good reliability and validity. The reduced version demonstrated psychometric properties similar to those of the long version, with a substantial reduction in the number of items.