Research

Comparing a disease-specific and a generic health-related quality of life instrument in subjects with asthma from the general population

Milo A Puhan¹ , Jean-Michel Gaspoz² , Pierre-Olivier Bridevaux³ , Christian Schindler⁴ , Ursula Ackermann-Liebrich⁴ , Thierry Rochat³ and Margaret W Gerbase³

¹  Department of Internal Medicine, Horten Centre for Patient-oriented Research, University Hospital of Zurich; CH-8091 Zurich, Switzerland

²  Department of Health and Community Medicine, Division of Community and Primary Care Medicine, University Hospitals of Geneva, CH-1211 Geneva 14, Switzerland

³  Department of Internal Medicine, Division of Pulmonary Medicine, University Hospitals of Geneva, CH-1211 Geneva 14, Switzerland

⁴  Institute of Social and Preventive Medicine, University of Basel, CH-4051 Basel, Switzerland

author email corresponding author email

Health and Quality of Life Outcomes 2008, 6:15doi:10.1186/1477-7525-6-15

Published:	15 February 2008

Abstract

Background

Few epidemiologic studies have assessed health-related quality of life (HRQL) of asthma patients from a general population and it is unclear which instrument is best suitable for this purpose. We investigated the validity of the Asthma Quality of Life Questionnaire (AQLQ) and the SF-36 completed by individuals with asthma from the population-based SAPALDIA (Swiss study on air pollution and lung diseases in adults) cohort.

Methods

The study included 258 participants with a physician-diagnosed asthma who had completed the AQLQ and SF-36. We assessed floor and ceiling effects, internal consistency reliability and cross-sectional validity with a priori hypotheses that correlations between the specific HRQL domains (e.g. "symptoms" or "physical functioning") and the corresponding external validation measures (respiratory symptoms, need for doctor visits, limitation in activities due to asthma and lung function) would capture similar aspects and be correlated moderately (≥ 0.3) to strongly (≥ 0.5), whereas non-corresponding domains be correlated weakly with each other (<0.3).

Results

The AQLQ showed pronounced ceiling effects with all median domain scores above 6 (scores varied from 1–7). For the SF-36, ceiling effects were present in 5 out of 8 domains. Cronbach's alpha was >0.7 for all AQLQ and SF-36 domains. Correlations between the AQLQ domains "respiratory symptoms", "activity limitation" and "environmental exposure", and the validation measures ranged from 0.29–0.57. Correlations between the "emotional function" domain and the validation measures were also in this range (0.31–0.55) and not as low as we hypothesized. For the SF-36, correlations between "physical functioning" and "role physical", and the validation measures ranged from 0.25–0.56, whereas "role emotional" and "mental health" correlated with these measures from 0.01–0.23.

Conclusion

The AQLQ and the SF-36 showed fairly good internal consistency. Both instruments are limited by ceiling effects, but they appear less pronounced in the SF-36, which also shows a better discrimination between different aspects of HRQL. The SF-36 may therefore be a more valid measure of HRQL than the AQLQ when applied to individuals with asthma from the general population.