Keeping a check on the heartbeat of India via cloud5 min read . Updated: 28 Jun 2020, 08:28 PM IST
- AI medtech startup Tricog is helping drive remote diagnosis of heart attacks during the coronavirus pandemic
- After tertiary and primary care centres, Tricog founders are planning to take their diagnosis technology to people’s homes
The number of heart attack cases in hospitals dropped by over 50% after the novel coronavirus was declared a pandemic, according to a worldwide survey by the European Society of Cardiology (ESC). That’s not good news though.
The fear of going to the hospital is behind the drop, which suggests many lives are being lost without timely help. “This is the strongest evidence yet of the collateral damage caused by the pandemic," said ESC president Barbara Casadei while announcing the findings earlier this month.
One place better prepared to deal with this situation happens to be Goa. More than a year ago, the state rolled out a project, STEMI (ST-Elevation Myocardial Infarction), with the help of Bengaluru-based startup Tricog. It has deployed 12-lead ECG machines with Tricog communicator devices at peripheral government centres.
The device sends ECG data to the cloud in real time, where artificial intelligence (AI) algorithms help cardiologists give a diagnosis within 10 minutes. The diagnosis goes to a specialist team at the Goa Medical College, which advises field medical staff on what to do.
For example, a woman, 70, came to the Mapusa district hospital complaining of chest pain last week. The Tricog system at the hospital could diagnose the heart attack within minutes, and she was given a clot-dissolving thrombolysis injection. She was then shifted to Goa Medical College for an angioplasty. STEMI is integrated with a specially equipped ambulance service for heart patients to be shifted to treatment centres. The system saves time and cost apart from widening access to tertiary care.
“Despite covid-19 and lockdowns, the number of heart attack cases did not drop in Goa as much as it did in other places around the world," says Charit Bhograj, a cardiologist in Bengaluru, who is the founder of Tricog.
Bhograj started Tricog in 2015 along with three others, Zainul Charbiwala, Udayan Dasgupta and Abhinav Gujjar, who had the digital tech nous to set up a cloud-based ECG diagnosis system. “As a cardiologist, I could save one life at a time. But the Tricog team has an impact on 1,000 lives every day," says Bhograj.
The startup partnered with GE to deploy ECG machines that could provide a 12-channel recording for the Tricog communicator to transmit to the cloud. Now an upgraded version works with ECG models from other manufacturers too. Tricog charges a monthly rental if it provides the hardware. The rest of the pricing is based on usage of its cloud-based diagnosis.
The quality of data is vital for the ECG interpretation over the cloud to be as good as any you would get at a large hospital. Over 200 heart conditions can show up in an ECG. Hence, the need for a clinical grade machine with 12 leads that can detect any of these conditions. Compare that with the single lead ECG app of the Apple Watch, which recently got FDA clearance for use during the pandemic. Such an app has limited functionality and cannot be a substitute for a standard ECG to diagnose a heart attack. ECG apps in other wearables and phones are also limited to a single lead because otherwise you would need multiple electrodes attached to your body. The ECG app of an Apple Watch can pick up about 17 conditions and most of the critical ones are not among those, says Charbiwala. “Only one out of those 17 conditions, called ventricular tachycardia, is something you associate with a cardiac arrest."
It’s useful for somebody with one of the 17 conditions to use the app, but to diagnose a wider range of heart attacks requires a richer dataset. Tricog’s AI algorithms can detect 40-50 conditions with 95-98% accuracy, says Charbiwala. The final interpretation of the ECG comes from a team of cardiologists; the AI engine just speeds up the process with its inputs.
“From a medico-legal standpoint, we still have all ECGs over-read by human doctors, but their job is so much easier because the AI is doing part of it. And it gets better over time," says Charbiwala.
ECG analysis has subtleties, which preclude just looking at a reading and deciding whether it’s a heart attack. Different heart conditions, gender, age and so on come into it. The startup’s challenge lies in ensuring accuracy, speed and consistency in processing thousands of ECGs a day and scaling that.
The challenges ahead
The Tricog system has been deployed in over 3,000 centres across India as well as around 1,000 centres in 12 countries of Southeast Asia and Africa. That’s still a drop in the ocean, especially in India, where it takes over six hours to reach a hospital in more than half the cases of heart attacks, as shown by the Indian Council of Medical Research data.
Yet, doctors are not the most avid adopters of new technology. It takes time for sales teams to persuade governments and hospitals to try a new system. This is even more challenging during the pandemic. At the same time, the demand for remote diagnostic tools has shot up because patients are reluctant to come to hospital.
So Tricog has started a new free service where an image of an ECG can be uploaded for interpretation. It’s a stop-gap arrangement because diagnosis based on the image of an ECG printout has limitations compared to the full-fledged Tricog system. Details are lost in the process of digitizing the image.
“We were conflicted over releasing this service because an approximate answer is sometimes worse than no answer," says Charbiwala. “But we realized that we had to provide some sort of accessibility in times like these. It’s a way for users to experience what a cloud-based service could do for them. Maybe post-covid they will come in and say, ‘Now, show us your big machine’."
Tricog has eschewed jugaad to build a system to provide a high-end ECG diagnosis over the cloud. But covid-19 creates new imperatives. And the startup raised a $10.5 million series B round just before the pandemic, giving it some bandwidth to pursue them.
Diagnostic centres have been reaching out to the startup to extend its ECG analysis to patients’ homes. But it’s more complex than it sounds, because treating a heart attack is time-sensitive.
The idea is that a call to an emergency number like 108 should bring an ambulance with a paramedic and a Tricog ECG machine. “We’ve been talking to doctors in Kerala to deliver a service like this where an ECG can be transmitted quickly from a patient’s doorstep," says Charbiwala. “Kerala has been ahead in several healthcare aspects, but other states are also discussing these possibilities. Large hospitals might be able to do that as well. But right now we’re seeing more interest from diagnostic chains."
Remote diagnosis of heart attacks is bound to happen more, with the pandemic being a big driver, believes feels Charbiwala. But it’s going to require a full-fledged ECG reading, which an Apple Watch, iPhone accessory or Samsung app is still a long way from achieving.
Malavika Velayanikal is a consulting editor with Mint. She tweets @vmalu.