The Potential of a Smartphone App for TB Diagnosis Through Cough Sounds
Identifying the unique features of cough sounds associated with different diseases has long eluded scientists. However, a breakthrough may be on the horizon with the development of a new machine learning tool that aims to differentiate the signature sounds of tuberculosis (TB).
The Significance of Cough in Respiratory Infections
Cough is a key symptom of respiratory infections, and the variations in cough patterns and frequencies across different diseases have prompted the quest for a smartphone app capable of accurately detecting TB-related coughs.
Efforts to create a cost-effective yet efficient TB screening tool have been ongoing for years, especially in regions with limited healthcare infrastructure and diagnostic resources.
With the resurgence of TB incidence and mortality rates after a period of decline, the need for precise screening tools has become more pressing. While current diagnostic methods like sputum culture and GeneXpert tests are highly accurate, their affordability in TB-endemic areas remains a challenge.
Developing a TB Screening App
An international research team is exploring the hypothesis that the distinct cough patterns and frequencies associated with TB can serve as reliable indicators for the disease’s detection through smartphone technology.
The investigational app, known as TBscreen, is still in the experimental phase and not yet available for public use. Nevertheless, its potential relevance has grown significantly amid the global TB crisis.
Recent findings published in Science Advances by collaborators from the University of Washington and the Center for Respiratory Diseases Research in Kenya shed light on the research progress. The team comprises experts in engineering, computer science, medicine, and infectious diseases.
Enhancing TB Detection Through Sound Analysis
Through TBscreen, researchers observed that the app, in conjunction with a smartphone microphone, exhibited superior accuracy in identifying active TB cases compared to traditional high-cost microphones.
Manuja Sharma, an engineer at the University of Washington, explained, “By studying the characteristics of cough sounds as a diagnostic tool for TB, we recruited individuals with TB-related and non-TB-related coughs in Nairobi, Kenya.”
The machine learning system is undergoing training to recognize the specific patterns and frequencies present in TB-related cough sounds, while also learning to distinguish them from coughs associated with other respiratory conditions.
As the development of the TBscreen app progresses, the potential for a groundbreaking, cost-effective TB screening tool becomes increasingly promising, offering hope for improved diagnostic capabilities in TB-endemic regions.