Scientists are developing a machine learning tool that uses a smartphone app to screen for tuberculosis (TB), according to a study published in Science Advances. The tool, called TBscreen, aims to accurately discern the signature sounds of TB-related coughs, which differ from coughs associated with other respiratory infections. TB is a highly infectious bacterial disease, and the need for accurate screening tools is increasing as incidence and mortality rates rise. Current gold standards for TB diagnosis, such as sputum culture or GeneXpert molecular tests, are highly accurate but costly, especially in resource-challenged regions.
Researchers from the University of Washington in Seattle and Kenya’s Center for Respiratory Diseases Research in Nairobi collaborated on the study. They found that TBscreen, when fed audio of coughs through a smartphone microphone, identified active TB more accurately than when fed through expensive microphones. The app is still in the investigational phase and not yet ready for distribution. However, given the rising numbers of TB cases globally, its development is timely.
The team is training the machine learning tool to recognize the pattern and frequency of coughs caused by TB. They are also working on training the app to distinguish TB-related coughs from those caused by other respiratory disorders. The tool must discern nuances and factors affecting cough patterns, some of which are inaudible to the human ear, to accurately screen for TB.
Lead author Manuja Sharma, an engineer at the University of Washington, explained that the mechanism of cough production varies based on factors such as mucus properties, respiratory muscle strength, mechanosensitivity, and chemosensitivity of the airways. The study design was constructed to minimize background noise and environmental variability between the control and TB disease groups, ensuring that the machine learning model trains on differences in cough features rather than ambient noise.
Although the app is still in its early stages, it holds promise as a low-cost and high-tech screening tool for TB. With the increasing incidence and mortality rates of TB, particularly in resource-limited areas, the development of an accurate and accessible screening tool is crucial. The researchers hope that TBscreen will provide a solution for these regions, where healthcare infrastructure and diagnostic tools are inadequate.
In conclusion, the development of TBscreen, a machine learning tool integrated into a smartphone app, has the potential to accurately screen for tuberculosis by analyzing cough patterns and frequencies. The app’s ability to distinguish TB-related coughs from those caused by other respiratory disorders may provide a cost-effective and accessible solution for screening high-risk populations in resource-limited regions. Further research and development are necessary to ensure its accuracy and effectiveness before its distribution worldwide.
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it