Did you know? The optic nerve has atleast 1.7 million nerve fibres. The human eye shares several vascular and neural similarities to the brain – making it one of the most effective natural biomarkers for illnesses such as Alzheimer’s Disease, diabetes and hypertension. Over the past few years, a lot of research has been dedicated to understanding how AI can be utilised to scan retinal images to detect illnesses – this technology is non-invasive and doesn’t take more than 10 minutes, allowing medical staff to detect illnesses faster and provide a course of treatment in a timely manner. In 2018, Google’s DeepMind and UK’s Moorfields Eye Hospital conducted a joint research project where the AI tool was able to accurately interpret eye scans and detect the presence of an anomaly or illness.

Today, an Indian startup Radical Health is using a fundus camera or a retinal camera to scan a human eye, and a proprietary AI algorithm can provide an alert on a possible ailment in a matter of a few minutes. This particular technology is especially proving to be handy during the ongoing COVID19 health crisis in Maharashtra – by screening patients using the fundus camera for comorbidities like diabetes and hypertension, Radical Health is able to segregate high and moderate risk patients immediately, and ensure they are given the right course of treatment.

Independent research has pointed that from COVID19 mortalities world over, nearly 48% have succumbed to the illness on account of having comorbidities, most commonly diabetes and hypertension. Many of these patients weren’t even aware they had an underlying condition until they tested COVID-positive. The presence of comorbidities leads to a compromised immune system and could make them more susceptible to death or a significant deterioration in health, compared to COVID-positive patients without comorbidities.

Maharashtra is one of the states to be worst affected by COVID19. The state alone has 615,000 confirmed cases, with 20,687 deaths (most recent numbers) and 438,000 recoveries. Radical Health’s easy-to-use solution is now being used by the Brihanmumbai Municipal Corporation’s (BMC) network of hospitals to screen patients, in order to limit the fatality rate and help patients receive timely care. Rito Maitra, cofounder of Radical Health says, "We have one of the largest databases of fundus images. We carry out nearly 15,000 AI-based diagnostic cases every week. For COVID, our testing numbers vary between 1,000 to 1,500 per week and are currently focusing on bigger BMC wards in Mumbai." While Radical Health’s engineers build and train the algorithms, a team of doctors and retina specialists provide strategic & research-based counsel to the engineering team.

(A patient in a BMC ward in Mumbai getting tested for comorbidities with a fundus camera)

Why Govt of Maharashtra Sees AI As A Game Changer

Last year, Radical Health got an opportunity to do a three-month pilot project with the Government of Maharashtra. That’s when Maitra met Janak Shah, who was working as an advisor to Chief Minister’s War Room – a specialized unit set up by former CM Devendra Fadnavis to expedite projects to improve rural and urban infrastructure in various areas including healthcare. Shah, who is now associated with Wadhwani AI in Maharashtra, also helped set up the BMC COVID19 War Room, under the leadership of MC Praveen Pardeshi and AMC Ashwini Bhide, to lead relief efforts when the infection began to spread rapidly in March in the state. Shah says, “The three month pilot with Radical Health was extended to 11 months, owing to the ease of deployment and efficacy of the solution presented by Rito and his team. Pre COVID, we have screened 2.5 lakh patients in Maharashtra.”

When the war room was set up, Shah reached out to Maitra asking him to help deploy his technology for screening patients. “The good news in India is that 80% of COVID patients are asymptomatic, which means home isolation and self care should help them recover. But 20% are in the moderate to high risk category, who need that timely intervention to ensure they don’t succumb to the disease.”

What Shah and Maitra realized while driving the pilot in rural Maharashtra was that most people live with comorbidities. Some conscientious individuals do routine health checkup and take preventive steps, but this is a very small percentage of the population. “During our rural outreach, we found that 92% of the patients screened didn’t even know they had a preexisting condition. Regular health checkups in rural India is a real challenge, and its even harder to scale tech-based solutions due to spotty Internet and some resistance to modern methods of screening.”

They realized that this portion of the population would be the most at risk during COVID19.

It has been 6 weeks, and Radical Health has been screening patients in BMC wards. Per the latest stats shared by Maitra, a total of 5881 patients have been screened by the AI. 4135 have been identified as ‘low risk’, 1567 as ‘medium risk’ and 179 as ‘high risk’. Of the lot, 728 had diabetes and 1210 had hypertension and 102 had other ailments such as asthma, jaundice, TB, kidney stones, glaucoma and age-related macular degeneration. Among the 728 with diabetes, the AI was able to detect 489 new cases and from the lot of 1210 patients with hypertension, the machine could detect 824 new cases. Patient awareness levels for both diseases was less than 34%.

“The healthcare infrastructure in our country cannot support all patients. In fact, this is the case in most countries. At such times, this kind of technology can immediately segregate the high risk patients from low risk ones, and divert them to a course of treatment appropriate to their condition. This will greatly help medical staff manage their workload as well,” explains Maitra.

Shah adds that the current crisis requires a priority-based approach to manage the caseload competently, and technology can play a crucial role. While the fatality rate is 2.15% - among the lowest in the world – it’s a significant number nonetheless. “The focus is on increased testing, bringing down the fatality rate, ensuring there are sufficient beds, medicines, medical staff and ways to manage existing ailments. Technology will be a huge help in managing all these variables.”

Eventually, the plan is to screen all of Mumbai, especially the slum populations. Radical Health is in talks with 2-3 more Indian states to execute similar pilot projects. Maitra says the experience of working with a state government was an eye-opener. “The reception and access we received has been very encouraging. We were able to improvise our offering in just five days for the Maharashtra government, and one of the reasons this process was so quick was due to the efficiency of the government team. This kind of collaboration, acceptance and faith shown by government in startups is very inspiring, and I hope this encourages more entrepreneurs to innovate.”

Creativeart: Freepik