RESUMO
OBJECTIVE: Translations of validated questionnaires help to compare different countries/cultures populations and establish protocols for global health. OSA-18 is a validated disease-specific questionnaire for pediatric Obstructive Sleep Apnea (OSA). Our aim was to validate OSA-18 in Greek and correlate it with polysomnography results and OSA severity. STUDY DESIGN: Prospective instrument validation study. SUBJECTS AND METHODS: OSA-18 was translated in Greek and back into English. Children undergoing polysomnography due to snoring were recruited prospectively. OSA-18 was completed by parents during the initial clinic visit (test), in the evening prior to the sleep study (retest), and 3 months postoperatively for subjects who underwent adenotonsillectomy. Internal consistency and test-retest reliability were evaluated. Validity was assessed by exploring correlations between OSA-18 and AHI, by comparing OSA-18 of OSA and non-OSA groups, and by calculating questionnaire's sensitivity and specificity for detecting OSA. Total scores of non-OSA, mild, moderate and severe OSA subgroups were compared. In OSA children who underwent adenotonsillectomy, preoperative and postoperative total scores were compared to assess responsiveness. RESULTS: Test-retest questionnaires were fully completed for 141 children. OSA-18 in Greek had good internal consistency (Cronbach's alpha 0.951 for test and 0.947 for retest) and test-retest reliability (Pearson's correlation coefficients between test and retest scores: 0.850-0.946; P<0.05). Total and subscale OSA-18 scores and AHI were significantly correlated (Spearman's correlation coefficients: 0.376-0.633; P<0.01), while children with OSA had higher total OSA-18 score than those without OSA [median (interquartile range): 61 (35) vs. 38 (22), respectively; P<0.001)]. Sensitivity was 53.4%, suggesting poor validity compared to polysomnography. All OSA severity subgroups had significant higher score than non-OSA. OSA-18 scores postoperatively were significantly lower compared to preoperatively (22.91±5.49 vs. 67.13±15.27, respectively; P<0.001), indicating good responsiveness. CONCLUSION: Greek OSA-18 is an instrument with satisfactory internal consistency, reliability, and responsiveness, but it is a poor predictor of OSA severity.
