RESUMO
OBJECTIVE: Automated threshold audiometry (ATA) could increase access to paediatric hearing assessment in low- and middle-income countries, but few studies have evaluated test-retest repeatability of ATA in children. This study aims to analyse test-retest repeatability of ATA and to identify factors that affect the reliability of this method. DESIGN: ATA was performed twice in a cohort of Nicaraguan schoolchildren. During testing, the proportion of responses occurring in the absence of a stimulus was measured by calculating a stimulus response false positive rate (SRFP). Absolute test-retest repeatability was determined between the two trials, as well as the impact of age, gender, ambient noise, head circumference, and SRFP on these results. STUDY SAMPLE: 807 children were randomly selected from 35 schools in northern Nicaragua. RESULTS: Across all frequencies, the absolute value of the difference between measurements was 5.5 ± 7.8 dB. 89.6% of test-retest differences were within 10 dB. Intra-class correlation coefficients between the two measurements showed that lower SRFP was associated with improved repeatability. No effect of age, gender, or ambient noise was found. CONCLUSIONS: ATA produced moderate test-retest repeatability in Nicaraguan schoolchildren. Participant testing behaviours, such as delayed or otherwise inappropriate response patterns, significantly impacts the repeatability of these measurements.
Assuntos
Audiometria , Ruído , Humanos , Criança , Reprodutibilidade dos Testes , Audiometria de Tons Puros/métodos , Limiar Auditivo/fisiologiaRESUMO
OBJECTIVE: Determine whether an electronic tablet-based Wireless Automated Hearing-Test System can perform high-quality audiometry to assess schoolchildren for hearing loss in the field in Nicaragua. STUDY DESIGN: Cross-sectional. SETTING: A school and hospital-based audiology clinic in Jinotega, Nicaragua. SUBJECTS AND METHODS: Second and third graders (nâ=â120) were randomly selected for hearing testing in a school. Air conduction hearing thresholds were obtained bilaterally using a Wireless Automated Hearing-Test System at 1000, 2000, and 4000âHz. Referral criteria were set at more than 25âdBHL at one or more frequencies. A cohort of children was retested with conventional audiometry in a hospital-based sound booth. Factors influencing false-positive examinations, including ambient noise and behavior, were examined. RESULTS: All children with hearing loss were detected using an automated, manual, or two-step (those referred from automated testing were tested manually) protocol in the school (sensitivityâ=â100%). Specificity was 76% for automated testing, 97% for manual testing, and 99% for the two-step protocol. The variability between thresholds obtained with automated testing was greater than manual testing when compared with conventional audiometry. The percentage of participant responses when no stimulus tone was presented during automated testing was higher in children with false-positive examinations. CONCLUSION: A Wireless Automated Hearing-Test System identified all children with hearing loss in a challenging field setting. A two-step protocol (those referred from automated testing are tested manually) reduced false-positive examinations and unnecessary referrals. Children who respond frequently when no tone is presented are more likely to have false-positive automated examinations and should be tested manually.