Adult asthma scores - development and validation of multivariable scores to identify asthma in surveys

Publication date: Available online 18 July 2018

Source: The Journal of Allergy and Clinical Immunology: In Practice

Author(s): Ana Sá-Sousa, Ana Margarida Pereira, Rute Almeida, Luís Araújo, Mariana Couto, Tiago Jacinto, Alberto Freitas, Jean Bousquet, João A. Fonseca

Abstract

Background

One of the questions in epidemiology is the identification of adult asthma in studies.

Objective

We aim to 1) develop and validate multivariable scores for adult asthma identification in epidemiological studies; 2) explore cut-offs to rule-in/-out asthma, compared to asthma diagnosed by a physician after clinical examination and diagnostic tests, blinded to the self-administered questions.

Methods

We analyzed data (n=711 adults) from a nationwide population-based study. The predictors were self-administered questions identified in a literature review (A2-score) and from the GA2LEN questionnaire (GA2LEN-score). Scores were developed using exploratory factor analysis. Internal consistency, discriminative power and diagnostic accuracy were assessed.

Results

The A2-score comprises 8 questions (including "Did a physician confirm you had asthma?") and GA2LEN-score, 6 questions (including "Have you ever had asthma?"). Both had high Cronbach's alpha (0.89 and 0.85) and good AUC (90.4% and 89.0%). The scoring is the sum of positive answers. Asthma is present (rule-in) for scores≥4 (Specificity:99.2%; PPVs:93.3%, 91.7%; Accuracy:89.4%, 87.4%, respectively for A2-score and GA2LEN-score). Asthma is excluded (rule-out) for A2-scores of 0-1 and GA2LEN-score of 0 (Sensitivity:93.1%; NPV:98.2%, 98.0%; Accuracy:89.4%, 82.8%).

Conclusions

These practical scores can be used to rule-in/-out asthma in epidemiological studies and clinical screening/triage settings. They may help physicians in primary care or other specialties to screen asthmatics using a simple score with a high level of discrimination and to identify the best candidates to be referred for a diagnostic workup. Moreover, their use may contribute to reducing the inconsistencies of operational definitions of asthma across studies and surveys.