Stratifying asthma severity in children using cough sound analytic technology.

We use the collected data and cough recordings obtained in the Breathe Easy Study (ACTRN12618001521213). This dataset consists of a training set with coupled cough recordings, clinical data and single adjudicator diagnostic evaluation and a testing set with cough recordings, clinical data and a consensus panel adjudicator diagnostic evaluation. Phase 1 of the trial is to use the training set to develop the mathematical models to predict asthma severity using well known commonly used clinical scales (eg Pulmonary Score). Phase 2 of the trial is to use the testing set to perform a diagnostic accuracy study. The Breathe Easy protocol requires patients with defined respiratory conditions (including Asthma, COPD, Bronchiolitis, Croup, Interstitial Lung Disease, Pneumonia, Upper and Lower Respiratory infections, Bronchiectasis, Emphysema), confirmed by expert review, to provide one series of (5-10) coughs, recorded by a smartphone for analysis and technical development of an algorithm to produce a digital diagnostic test. This procedure takes 30 seconds when the patient can provide a voluntary cough and up to 5 minutes if spontaneous. Children (who are able to cough on demand or are doing so spontaneously) and adults will be approached by a trained research nurse. The recordings are collected at two Perth hospitals in the emergency department, inpatient wards and ambulatory care settings.