Background: There is a need for a validated self-assessment questionnaire for cognitive impairment in subjects reporting subjective tinnitus. The objective was to develop a patient-reported outcome measure.
Methods: This was a prospective, non-interventional, multicultural study. The 30-item "Attention and Performance Self-Assessment Scale" (APSA) was linguistically validated in Germany, Mexico and USA and was analyzed for content and structure. The analysis included descriptive statistics of baseline data, item characteristics, test-retest reliability (intra-class correlation coefficients, ICC), definition of internal consistency (Cronbach' s alpha), and explorative and confirmatory factor analysis to define the structure of the scale. Correlations with various tinnitus scales and subscales from the Hospital Anxiety and Depression Scale (HADS) were done to estimate convergent validity.
Results: The data for 211 subjects aged 30 through 60 years, (mean= 48.5 years, SD= 8.3) with mild to moderate tinnitus (mean Tinnitus Handicap Inventory-12 (THI-12) total score 11.2, SD= 5.3) were analyzed. The majority of subjects had sub-clinical scores for anxiety and depression (HADS below 11 points). Sequential principal factor analyses of the APSA resulted in a subscale which included 20 (APS20) of the original 30 items and two correlated subscales (AP-F1, AP-F2) defined by 9 items each. Both factor solutions were confirmed by confirmatory factor analysis. Test retest reliability of the APS20, AP-F1 and AP-F2 (ICC ≥ 0.87) and internal consistency (Cronbach's alpha ≥ 0.89) are high. APS20 correlated moderately high with HADS (depression: 0.54; anxiety: 0.62) and THI-12 total (0.52). In a few cases, AP-F2 correlated higher than AP-F1 with other scales (e.g. HADS-depression with AP-F1: only 0.46, but AP-F2: 0.59).
Conclusions: APS20, AP-F1, and AP-F2 have good psychometrical properties. The scales will add value to the assessment of cognitive aspects of quality of life and mental health in the population with subjective tinnitus. The subscales AP-F1 and AP-F2 may be helpful for detecting specific cognitive failures and may be sensitive to different interventional effects.