114 Chapter 6 percentiles in the corresponding sample from the general population with the same sex and age. To ensure balanced representation across age and sex categories, quota sampling was applied: participants were enrolled until each predefined age-sex stratum was filled. Due to our aim to construct normative scores for the general population, we anticipated a skewed distribution of severity scores for most subscales of the OQUA, with many individuals scoring zero or near zero. Based on the general prevalence of hearing loss and tinnitus a higher score in older aged individuals or age groups was to be expected in these groups. For the anticipated skewed distribution of the item scores linear regression was not considered appropriate for the derivation of norms. Instead, norms were developed based on zero-inflated beta regression.12 Adjustments for age and sex were implemented when appropriate, according to a stepwise model selection based on the Akaike Information Criterion.12 Due to the limited number of participants, additional adjustments for educational level per age category and sexe was not feasible. The modeled median, 75th and 95th percentile of the distribution of scores in the general population was calculated. The 95th percentile was taken as the threshold between a normal and abnormal score of each subscale (each ear complaint). Hence, 5% of the normal population would be expected to score above this threshold. Results Five hundred Dutch adults completed the Otology Questionnaire Amsterdam. Four patients had to be excluded because of mentioning recent visit to a ENT surgeon for an ear complaint or having a relevant ENT diagnosis in their medical history. The remaining 496 participants were eligible for inclusion and analysis. Out of the 496 participants 243 were male, 253 were female. Age range was 18-91 years of age with mean age of 54. All age groups as defined in the methods were evenly distributed with 70-72 participants per age group and a 50/50 male-female ratio. A hundred and seventy-three respondents were retired, 246 were employed, 77 were unemployed (20 incapacitated, 26 stay-at-home parent, 17 student, 14 job seeking). 355 participants were none smokers. Scores on each of the 9 subscales ranged from 0 to 100. In Figure 3, scores for each OQUA subscale are plotted as function of age and sex. For all subscales, scores were heavily skewed to the right with many individuals scoring zero or near zero and had proportions of zero scores between 28% and 58%.
RkJQdWJsaXNoZXIy MjY0ODMw