Using AI to detect mental health disorders likely to be counterproductive

AI-based vocal analytics may allow patients to detect the disease at home by just speaking into a smartphone application, eliminating the need to be physically present at treatment facilities for diagnosis.

NEW DELHI: Advances in artificial intelligence (AI) technology have made it possible to detect mental disorders such as anxiety and depression from cues in a person's voice, but doctors warn that adopting such tools in haste may prove to be counterproductive.

Despite the staggering number of patients being reported to suffer from disorders like depression, India struggles with the stigma of mental health diseases -- deterring patients from getting the right help in time.

Read Also: Hypoglycaemia, Mental Health Side Effects: Label side-effect warnings of Fluoroquinolone, DCGI tells Drugmakers

AI-based vocal analytics may allow patients to detect the disease at home by just speaking into a smartphone application -- eliminating the need to be physically present at treatment facilities for diagnosis.

"This technology will be helpful to people who are bound by the stigma of going to a doctor for therapy or diagnosis," Dr Rajendra Singh, a psychiatrist based in Bhopal, told PTI.

Last year, CompanionMx, an application developed by the behavioural analysis firm Cognito Corp, launched an AI-based mobile mental health monitoring system, clinically validated by the Harvard Medical School in the US.

The app uses objective data instead of self-reports to track the mood of the users by analysing their voice for acoustics and behavioural biomarkers, in order to predict the core symptoms of mood and anxiety disorders.

Read Also: Digital health services are transforming the health sector in India: Nadda

Similarly, Sonde Health a US-based digital medicine company uses vocal analytics, licensed by the Massachusetts Institute of Technology (MIT)'s Lincoln Laboratories in the US, to monitor and diagnose depression, dementia, Parkinson's and a host of other diseases.

"Sonde has collected voice and health data from over 10,000 volunteers in India as a part of the ongoing validation of its platform. We are currently developing partnerships with healthcare and commercial entities in India," Jim Harper, the co-founder of Sonde Health, told PTI.

However, health practitioners in India are yet to warm up to adopting such unconventional means of diagnostics.

"I think the prospect of a non-invasive technique of diagnostics and the idea of AI-based analytics in pathology seems interesting, though I am apprehensive of how it will pan out in India," said Arunima Patel, CEO of Mumbai-based IGenetics Diagnostics.

Read Also: Healthians to expand in South India, invests Rs 50 crore

"Even if the technology is US Food and Drug Administration (FDA) approved, people in the industry are still pretty conservative in terms of adopting such unconventional types of diagnostics methods," Patel said.

"My concern is that sometimes people with mental health disorders tend to hide or even pretend otherwise so the system of vocal analytics may be subject to manipulation. I want to know if it is completely foolproof," said Dr Satyam from Ram Manohar Lohia (RML) hospital here.

Dr Shweta Sharma, a senior clinical psychologist from Columbia Asia Hospital here felt that diagnosis through speech analysis is not enough to ascertain mental health as it takes an experienced psychologist to observe multiple biomarkers to reach a diagnosis.

With the stigma surrounding mental health, Sharma expressed concern if people consult vocal analytics based apps for diagnosis, they would take treatment into their own hands and may not even bother to seek proper therapy.

"It would be like how people use Google for diagnosis based on a few vague symptoms. I think it is going to create more chaos than the order in this regard," Sharma.

She also noted that every individual is different and that the manifestations of the symptoms of mental disorders may vary from person to person.

"Without going through the entire personal, interpersonal case history and comprehensively evaluating the extrinsic and intrinsic forces that push a person to a psychological disorder, it is hard to determine the disease without interacting with the patient, especially in the case of mental disorders," said Singh said.

The idea of a technology that allows people to diagnose themselves by speaking into their smartphones at home may be revolutionary, but until clinical trials offer conclusive results, getting the doctors to help in case of mental health issues may be the best way forward.