• Carole N M Brouwer, Anne G M Schilder, Henk F van Stel, Maroeska M Rovers, Reinier H Veenhoven, Diederick E Grobbee, Elisabeth A M Sanders, and A Rianne Maillé.
• Department of Paediatrics, Spaarne Hospital, Postbus 770, 2130 AT Hoofddorp, The Netherlands.
• Qual Life Res. 2007 Oct 1; 16 (8): 1357-73.

AbstractIn this study the reliability and validity of generic and disease-specific questionnaires has been assessed focusing on responsiveness. This is part of a study on the effects of recurrent acute otitis media (rAOM) on functional health status (FHS) and health-related quality of life (HRQoL) in 383 children with rAOM participating in a randomized clinical trial. The following generic questionnaires were studied: 1. RAND general health rating index, 2. Functional Status Questionnaire (FSQ Generic and FSQ Specific), 3. TNO-AZL Infant Quality of Life (TAIQOL), and the following disease-specific questionnaires: 1. Otitis Media-6 (OM-6), 2. Numerical rating scales (NRS) for child and caregiver (NRS Child and NRS Caregiver), and 3. a new Family Functioning Questionnaire (FFQ). Reliability was good to excellent (Cronbach's alpha range 0.80-0.90, intraclass correlation coefficient range 0.76-0.93). Moderate to strong correlations were found between the questionnaires as well as between questionnaires and relevant clinical indicators (r = 0.29-0.49), demonstrating construct validity. Discriminant validity for children with few versus frequent episodes of acute otitis media per year was good for most questionnaires (P < 0.004) but poor for the otitis media-related subscales of the TAIQOL (P = 0.10-0.97) and both NRS (P = 0.22 and 0.48). Except for the TAIQOL subscales, change scores were significant (P < 0.003) for generic and disease-specific questionnaires. Effect sizes were somewhat higher for disease-specific compared to generic questionnaires (0.55-0.95 versus 0.32-0.60) except for the TAIQOL subscales, which showed very poor sensitivity to change. Anchor-based methods resulted in a somewhat larger range of estimates of MCID than distribution-based methods. Combining distribution-based and anchor-based methods resulted in similar ranges for the minimally clinical important differences for generic and disease-specific questionnaires: 2-15 points on a 0-100 scale. Apart from the generic TAIQOL subscales, both generic and disease-specific questionnaires used in this study showed good psychometric qualities and responsiveness for use in clinical studies on children with rAOM.

18 keywords...

hide…