An intelligent stethoscope for the diagnosis of lung diseases
Respiratory diseases are one of the leading causes of death and disability in the world. According to the World Health Organization (WHO) more than one million children under five die annually of acute lower respiratory infections such as pneumonia. It represents 16% of overall deaths for children under five years old, 80% of these deaths are within the first two years of their life.
Low-and-Middle Income Countries (LMICs) are the most affected by pneumonia with a high prevalence in sub-Saharan Africa and south-east Asia. In high-income countries, asthma and Chronic Obstructive Pulmonary Diseases (COPD) affect 400 million people, killing more than 3 million each year, and increasing constantly due to air pollution.
Pneumonia kills about one children every 30 seconds worldwide.
World Health Organization (WHO) 2015
Identification of respiratory diseases at an early stage is essential for delivering the appropriate treatment, but diagnosis often relies on direct auscultation with a stethoscope and needs medical expertise. Thus, auscultation cannot be currently referred to health care workers, pharmacists, nurses, or even patients.
Therefore, we imagined Pneumoscope, an innovative concept of an unsupervised digital stethoscope powered by Artificial Intelligence. The solution is a device made of a microphone which records lung and heart sounds coupled with AI-based algorithms that analyze sound features and associate them with a disease. The results are then displayed on a user-friendly app, helping the user to grade breathing difficulties and optimize treatment delivery without any medical knowledge required. The project is the result of a close collaboration between the Hôpitaux Universitaire de Genève (HUG), Terre des Hommes foundation and EssentialTech Center.
– Proof of Concept (POC)
– UNIGE I-Teams 2019, 1st Prize (Jury & Public)
– ECAL Innovation by design 2019, 1st Prize (Jury & Public)
– Clinical Studies HUG
– Prix de l’Innovation HUG 2018, 1st Prize