Amsterdam : a generic patient reported outcome measure about the severity and impact of ear complaints . A cross-sectional study on the development of this questionnaire

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/coa.12950 This article is protected by copyright. All rights reserved. MS ELS BRUINEWOUD (Orcid ID : 0000-0001-6361-6126)

(eg SSQ and THI) 2,3 or disease-specific (eg COMQ-12 and MDPOSI). 4,5 However, patients visiting an ENT surgeon often report multiple ear complaints simultaneously. The use of multiple questionnaires per patient seems impractical in the outpatient clinic. Therefore, a PROM that addresses all relevant types of ear complaints is desirable in the outpatient clinic. This paper describes the development of the 'Otology Questionnaire Amsterdam' (OQUA), a generic PROM addressing the severity of ear complaints and their impact on daily life. The OQUA can be used in the outpatient ENT clinic.

| Ethical considerations
The study protocol was assessed by the Medical Ethics Review Committee of VU University Medical Center, Erasmus Medical Center, Isala and Amstelland hospital which led to the decision that the Medical Research Involving Human Subjects Act did not apply to this study. Written informed consent was obtained prior to study participation. All questionnaires were completed anonymously.

| Participants
Patients visiting an ENT surgeon with an ear complaint were recruited in two secondary (Isala and Amstelland hospital) and two tertiary (VU University Medical Center and Erasmus Medical Center) clinics, all located in the Netherlands. Patients were eligible to participate if they were 16 years and over, had an ear complaint, were not known to have a learning disability or cognitive impairment and had a good written understanding of the Dutch language, which was the dominant used language in the research setting.

| Study design
This cross-sectional study consisted of two phases: formulation and selection of items based on qualitative analyses (phase 1) and selecting the definite set of items based on quantitative psychometric analyses (phase 2). The development of the OQUA was based on the guidelines provided in 'Measurement in medicine'. 6  persons) of the target population to test the acceptability and relevance of all items of the questionnaire. 6 Therefore, the OQUA-v1 was pilot-tested (N = 32) to identify which ear complaints were rated relevant to patients and to select descriptions of ear complaints that best described the complaints as experienced by patients. Each item contained different descriptions of a specific ear complaint, which were all formulated based on patients' descriptions that were collected during patient interviews. For each item, the description that was selected most frequently was identified. Items were formulated based on these most selected descriptions. This resulted in OQUA-v2.

Keypoints
• A generic otologic questionnaire seems the most desirable PROM (compared to disease-or complaint-specific) as many patients visiting an ENT surgeon experience multiple ear complaints simultaneously and the use of several questionnaires next to each other seems unpractical.
• Eight relevant types of ear complaints were identified: earache, pressure in the ear, hearing loss, tinnitus, otorrhoea, itch, dizziness and loss of taste.  it is recommended to perform another pilot study in a new sample of the target population (again 15-30 persons). 6 Hence, OQUA-v3 was pilot-tested (N = 39) to assess the comprehensibility of the reformulated items using think-aloud interviews.
Items that were perceived unclear or ambiguous were reformulated in OQUA-v4.

| Phase 2: definite selection of items
Field-testing of OQUA-v4 was conducted (N = 352) to perform quantitative analysis. The number of completed questionnaires required to perform proper field-testing equals seven times the number of items on the questionnaire. 10 Quantitative analysis consisted of the following steps: A. Missing values (missings) and distribution of the scores. First, the percentage of missings for each item was determined. Items with <3% missings were considered acceptable and items with >15% missings were considered unacceptable. 6 Second, each item's score distribution was examined. The items with VAS scales were checked for a normal distribution using a Q-Q plot, histograms with a normal distribution curve and the significance in the Kolmogorov-Smirnov test. For items presenting normally distributed scores, the mean and SD were calculated to examine the distribution of the scores. For items presenting non-normally distributed scores, the median and interquartile ranges were calculated. 6 The distribution of the scores of items with a Likert scale was examined by determining how frequently each response option was selected.
B. Inter-item correlation. An inter-item correlation matrix was created. Items with low correlations (<0.2) with any of the other items were removed. Items correlating highly (>0.9) with any of the other items were checked for being identical. 6 C. Explorative factor analysis (EFA). EFA was conducted to gain insight into the factor structure of the OQUA and to examine possibilities for item reduction. EFA serves to identify a set of latent variables (ie factors) underlying the items. A condition to perform EFA is that responses are normally distributed. Generalised least squares were used to correct for non-normally distributed data. 11 The number of relevant factors was determined based on factors with an eigenvalue >1 and the 'elbow' in the scree plot. Orthogonal rotation (ie Varimax) was used to rotate the component matrix. 6 After orthogonal rotation, items were grouped into factors based on their factor loading. A minimum loading of 0.5 was taken as a threshold. 12 Cronbach's Alpha was used to examine the coherency within a factor (ie items within a factor should represent a common latent variable). A value between 0.70 and 0.90 was considered acceptable. 6 3 | RESULTS  Impact-scale. EFA was conducted with the 12 items of the impact-scale. EFA generated two factors with an eigenvalue >1.

| Patient characteristics
The scree plot showed an obvious 'elbow' after the first factor.
Therefore, EFA was repeated using one factor. All 12 items 4. Indicate the severity of itching in or on your ear on the line below. 5. I have an itch in or on my ear. 6. It feels as if there is something in my ear.
7. Indicate the severity of your tinnitus (this can be a hum, murmur, beeping noise or buzzing sound) on the line below. 8. I hear a hum, murmur, beeping noise or buzzing sound.
9. Indicate the severity of your hearing loss on the line below.

| Strength and weakness of the study
Following the guidelines of COSMIN, the required steps to develop a questionnaire were taken in the various phases of the study. 10 The study population represented a heterogeneous group by including participants of different ages, experiencing different types of ear complaints and an even distribution of male and female participants. Another positive feature of this study is that there were only few missings (<0.3%). Moreover, the missings were obtained from few respondents (7%) and were distributed over various items. This suggests that missings were not caused by poor formulation of an item.
A limitation of this study is that the origin of the participants was not evenly distributed across clinics. Fifty-eight participants (74%) in phase 1 and 303 participants (86%) in phase 2 originated from a tertiary clinic. This may have led to a selected display of distributions of complaints within the study population. However, every type of ear complaints was sufficiently represented in the study population. Another limitation of this study is the fact that the OQUA contains eight items answered by a VAS. Previous research has shown that not every patient, in particular the elder patient, is able to complete a VAS correctly. 16,17 This was also observed in this study. However, almost no problems occurred when the OQUA was completed digitally. Therefore, it is preferred to complete the OQUA digitally.

| Further research
Before implementation, the reliability and validity of the OQUA need to be investigated. In addition, OQUA's format will be further developed so that scores of different ear complaints can be presented graphically, which will give a quick overview of each patient's complaints. To optimise the comparison of patients' perspectives before and after an intervention, a third construct should be added next to complaints and impact, namely 'benefit and satisfaction'. The OQUA has the potential to provide the scores on all three parts of the questionnaire (complaints, impact and benefit and satisfaction) separately. We aim to examine OQUA's test-retest reliability and responsiveness in future research.

| CONCLUSION
The development of a generic otologic PROM covering all relevant ear complaints is presented. This generic questionnaire can be very useful as many patients with an ear complaint report multiple ear complaints simultaneously. The OQUA addresses eight types of ear complaints and contains items about impact in a 34-item questionnaire. The OQUA is based on thorough analyses including in-depth interviews, think-aloud interviews, pilot studies and field-testing. Further studies will focus on validity, reliability, the optimisation of graphical representation and satisfaction and benefit scores.

ACKNOWLEDG MENTS
We would like to thank MSc L. Rothuizen (ENT department of Erasmus Medical Center, Rotterdam, the Netherlands), dr. E.O. Laurens-Thalen (ENT department of Isala, Zwolle, the Netherlands) and dr.
R.J.A.M. van der Hulst (ENT department of Amstelland hospital, Amstelveen, the Netherlands) for their contribution to patient inclusion in this study. We would also like to thank MSc E. Haitsma for her contribution in the first phase of this study.

CONFLI CT OF INTEREST
None to declare.