In Mumbai, the Shantilal Shanghvi Eye Institute has integrated artificial intelligence (AI) into its community outreach to improve detection of eye diseases among residents of nearby slums. Using smartphone-based retinal imaging devices, healthcare workers can capture and analyze images of patients' retinas within seconds to detect early signs of conditions such as glaucoma and diabetic retinopathy.

This AI-driven approach allows less-trained personnel to triage patients effectively, referring those who need specialist ophthalmologist care. "The game-changer in terms of AI is we have been able to task shift the screening process to less-trained people, who can use the AI to triage people who require the ophthalmologist's opinion," said Radhika Krishnan, head of public health at the Shantilal Shanghvi Eye Institute.

The technology is provided by a Bengaluru-based firm whose AI has been trained on billions of medical images. The company developed the platform to address India's severe shortage of radiologists, with research indicating there is only one radiologist for every 100,000 people in the country. According to Remidio, each screening device costs about US$3,700, with an annual software subscription for unlimited smartphone scans priced at US$1,500.

Despite the benefits, medical professionals emphasize that AI remains a tool rather than a replacement. Roopesh Deshmukh, director and consultant radiologist at the clinic, stated, "To be honest, I have to really check each word in every case before I sign them off because AI can be really perfect but at times it can go haywire."

The AI-enabled screening has also helped reduce the time patients spend in hospitals, which previously could take an entire day and disrupt their work. Building trust among healthcare providers in the AI's reliability is crucial for its continued adoption. Ultimately, this technology enables the institute to reach a larger population for regular triage and timely referrals for treatment or further evaluation.

Sources: CNA Latest News

Sources