BMC to soon screen patients using voice samples, AI
MUMBAI: After conducting antigen and antibody testing, the Brihanmumbai Municipal Corporation (BMC) is set to diagnose citizens for Covid-19 on the basis of their sound waves.
By next week, the civic body will commence the non-invasive voice analysis of suspected and confirmed Covid-19 patients at the NESCO jumbo Covid-19 centre in Goregaon. The results of the analysis can indicate the presence of the virus in 30 seconds, and those testing positive, will undergo a reverse transcription polymerase chain reaction (RT-PCR) swab test to confirm if they are infected.
Using artificial intelligence (AI), Dr RN Cooper Municipal General Hospital, which is run by the civic body, will conduct the study on 2,000 patients at the centre. “It will take two-three months to conclude the study,” said Suresh Kakani, additional commissioner, BMC.
RT-PCR is considered the gold standard test for diagnosing Sarscov-2 – the virus that causes Covid-19 – but the result through this testing method takes more than eight hours. In contrast, a rapid antigen test can give results within 30 minutes, but this method can give a higher rate of false negatives according to the Indian Council of Medical Research (ICMR).
Dr Neelam Andrade, dean of Nair Dental Hospital, who is in-charge of the NESCO facility, said that though the concept of diagnosing patients through their sound waves is new, the United States of America and Israel are using it. When the symptoms of Covid-19 manifest, a patient starts developing breathing problems, which affects the amount of air exhaled, thereby interacting with inflamed muscles on its journey to generate voices or speeches. These interactions impact voice modulations— measurable qualities that form the basis of their biomarkers. Depending on its variations, an individual can be detected with Covid-19, said Dr Andrade.
“There is a voice app which can be installed on phones. The suspected patient will be asked to count a few numbers before the device. The voice sample will automatically get synced with the main server of the provider. Then, through artificial intelligence, the result will be procured within 30 seconds,” she said.
The app will analyse voices of suspected, positive and negative patients. Depending on their vocal biomarkers (VB), they will be diagnosed. For instance, if a person’s VB is below the standard point (0.5), the person will be considered as negative. But if anyone records VB above it, the person will be considered a suspected patient and undergo RT-PCR test for a confirmed result.
Cooper hospital, which has been given the responsibility of conducting the study, is yet to get the final nod from its ethical committee. “We are hoping to get it in a day or two,” said Kakani.