The application's broad, long-term objective is to develop, evaluate and commercialize a device for continuous monitoring of a child's asthma components (symptoms, lung function variability, nighttime awakenings, etc.) over a few weeks. Asthma is the most common pediatric chronic condition in the US. It affects 6 million children, causes 0.8 million emergency department (ED) visits, 0.13 million hospitalizations and $18 billion annually in healthcare costs. To reduce the burden of uncontrolled asthma, NHLBI evidence-based guidelines require periodic assessment of the child's asthma risk/impairment, and corresponding adjustments to treatment, until asthma control has been established. When assessing asthma impairment, pediatricians depend heavily on the self-report, which is unfortunately unreliable. This leads to underestimation of asthma impairment, undertreatment, and long paths to achieving control. To efficiently and effectively establish asthma control, a wearable device could be used to provide pediatricians an accurate, recent history of asthma components. Currently, no practical solutions exist for weeks-long monitoring of symptom frequency and lung function variability, both of which are critical components for assesing asthma impairment. Via the following Specific Aims, this application will use advances in embedded systems and machine learning to develop a small, wearable device for monitoring asthma symptoms that is (1) acceptable to patients; (2) able to robustly detect cough and wheeze symptoms; and (3) able to passively measure lung function variability.
Specific Aim 1 : Train and evaluate an algorithm to detect pediatric asthma symptoms (cough and wheeze) on a low power, small form factor wearable device. Specifications: 90% sensitivity; false alarm rate: 1 cough epsiode/day or 1 wheeze episode/day. Evaluate algorithm against medical expert (physician) scoring using the two best available asthma scoring tools (AS: asthma score; PRAM: Pediatric Respiratory Assessment Measure).
Specific Aim 2 : Design and evaluate algorithm to detect lung function variability on a low power, small form factor wearable device. Specifications: Using respiratory signals from sensor patch, detect variations ? 10% in forced expiratory volume in 1 second to forced vital capacity (FEV1/FVC). Evaluate algorithm against spirometry gold standard.
Uncontrolled pediatric asthma is a serious public health concern. There are proven therapies and management guidelines to achieve asthma control in most patients, but poor patient self-report of asthma daily symptoms and components has made this an elusive goal in practice. This project will develop and evaluate a wearable device that provides physicians with an accurate report of their patient's recent asthma symptoms and pulmonary function variability.
