Andhra Medical College, one of India’s oldest medical schools, has started trying out a new AI platform that assesses patients’ lung health, including those who have contracted COVID-19.
WHAT IT’S ABOUT
The technology being trialled was developed by Telangana-based Salcit Technologies. Called Swaasa, the AI platform uses machine learning to perform an audiometric analysis of cough sounds, along with temperature, oxygen saturation, symptoms.
Swaasa is being tested in the primary healthcare setting at the Rural Health Training Centre in Simhachalam. Around 2,000 patients are targeted to partake in the study, which will run over a six-month period.
WHY IT MATTERS
In an interview with Times of India, Dr P.V. Sudhakar, principal of AMC, said existing methods of diagnosis such as x-ray, CT scan and other pulmonary tests require a laboratory setup, which can be “expensive and time-consuming”. “Moreover, they are not available in all rural and tribal areas,” he added.
Dr Devi Madhavi, head of AMC’s Department of Community Medicine, also said in the news report that the Swaasa will be “highly useful in identifying the appropriate next intervention”, as the device helps in screening and identifying whether a lung condition is attributable to airways or lung parenchyma or pleura.
According to its website, Salcit aims to replace spirometry, the most common, simple lung diagnostic test, with its AI tool.
THE LARGER TREND
Like Salcit, researchers at the Massachusetts Institute of Technology have developed a low-cost solution for diagnosing COVID-19 that can be deployed in areas where comprehensive diagnostic testing is unavailable. Last year, they introduced an AI tool that also analyses coughs to determine whether or not a patient is positive for COVID-19.
The researchers collected over 70,000 audio recordings of people’s coughs through a website and used these data to develop, train and validate a model that checks specific acoustic biomarkers related to muscular degradation, vocal cord changes, sentiment or mood changes and changes in the lungs or respiratory tract.
After testing, the tool was found to have 97.1% accuracy, 98.5% sensitivity and 94.2% specificity in detecting COVID-19 positive cases.