PremiumScans and X-ray images from 850 hospitals and clinics across the country go to 5C Network (5CN), a digital platform of radiologists providing diagnostics as a service. The platform algorithmically assigns each case to the most suitable radiologist available, from among the 280 radiologists on the network currently.
The notified radiologist gets a patient’s clinical history and field doctor’s observations along with the scans. AI (artificial intelligence) tools also help in making accurate diagnosis in the shortest time possible. Teleradiology has been around for some time, but a network of hundreds of radiologists takes that to another level. It’s not like a hospital connecting to a remote radiologist who may be snowed under a bunch of scans and unavailable immediately. 5CN can provide on-demand access to a specialist straightaway, just like Uber would assign a driver for a ride.
“When I see a bunch of scans coming in and a bunch of radiologists who are available, it looks like a linear programming problem (of matching diagnostic requests and specialists optimally)," says 5CN co-founder Kalyan Sivasailam, whose expertise is in computer science. “We have an algorithm called Max Flow that optimizes for both the right radiologists and the turnaround time."
This network play is what provides the kind of scale and efficiency that a small teleradiology service cannot match. So, instead of sending its scans online to a particular radiologist or diagnostic centre, a clinic is putting the scans on a network where multiple radiologists can be tapped for immediate action.
Addressing accessibility gaps
This tackles the problem of access to a specialist service that is required in most areas of healthcare. There are around 12,000 radiologists in the country, most of them concentrated in the metros. That creates a huge shortage in the availability of specialists to interpret the 750,000 scans performed each day, which works out to 60 cases a day on average for each radiologist.
This becomes more acute in smaller towns and rural areas where radiologists are scarce. The National Capital Region has over 2,000 registered radiologists while the whole of Uttar Pradesh has less than 200. Delhi, Gujarat, Maharashtra, Karnataka and Tamil Nadu account for nearly 80% of the radiologists in the country.
So 5CN addresses short supply as well as geographical mismatch. It gives access to radiologists wherever they may be based. At the same time, the tech platform helps improve the productivity of radiologists by the speed and efficiency with which it can assign scans and also prioritize them according to the nature of the emergency.
Technology has advanced to help the four-year-old startup focus on building a digital health platform. “Today, compressions are so good that large images can come quickly to our servers. We can build systems that allow radiologists to get notifications wherever they are, and workflows that ensure cases are solved within a particular time, where we know the exact status of a case at any point. We could not have done all these things 10 years back when technologies did not exist to run a business like 5CN efficiently," points out Sivasailam.
The platform approach brought together three aspects of radiology that were being digitized—storage and retrieval of images which used to be siloed, teleradiology, and the application of AI to diagnosis. “Our platform is basically three parts: we are a system of record, we offer on-demand access to radiologists, and these specialists are aided by multiple AI models that help them to be faster and more accurate," adds Sivasailam.
Radiology AI companies can plug into the 5CN platform not only to scale up their business but also to test and validate their models on massive data. One of these is Mumbai startup Qure, whose FDA-approved AI product qER helps emergency room staff take action faster on CT scans of the brain. Instead of only being deployed on the systems of hospitals that are clients of Qure, its products have now become available to a much larger network of radiologists on the 5CN platform.
Dig deep approach
Ease of communication on the network’s user interface also facilitates consultation between doctors and radiologists. Take the case of a woman who fell off a bike, as described by a doctor on the 5CN blog. This was a standard trauma case where the assigned radiologist reported a “non-haemorrhagic contusion with haemorrhagic focus", which in ordinary language means an injury to the brain with a little bleeding. But the radiologist had a “Spidey sense" that there was “something off" in this case. He consulted a doctor on the platform who asked to be connected to the patient to elicit more details. It turned out that she felt weak and had a seizure right before the fall. Then a second set of contrast-enhanced CT scans revealed there were tapeworms in the brain.
“The fact that I had seen cases like this from this geography—where pork is commonly eaten—allowed me to zero in on the diagnosis. As expected, she said that she was a regular pork eater," says Dr Kushagra Garg in the blog post. “There are some cases that make me realize how incredible my profession can be."
5CN’s product-first approach is what differentiates it from earlier teleradiology companies. This is also a much more scalable model. “Earlier, teleradiology companies were built small and stayed small. We have achieved some level of escape velocity. The challenge for 5CN now is to understand how we can scale faster and go international," says Sivasailam. “I believe any hospital or diagnostic centre that’s not using 5CN is at a disadvantage. It’s my duty to make sure they have it, because I know the kind of impact we can have in helping generate better patient outcomes."
The startup got early backing from Unitus Ventures and Axilor Ventures, which runs an accelerator programme in Bengaluru. Its go-to-market push also came from being associated with GenWorks, a medical equipment maker backed by GE Healthcare. It’s a win-win proposition because GE can sell a CT scan machine to a hospital or diagnostic centre in a tier-2 town with the assurance that access to radiologists will not be a problem. In this way, 5CN gets introduced wherever GE machines related to radiology are sold.
Radha K., partner at Unitus, explains why such a partnership is vital for a health tech startup. “Healthcare is a complicated area to address in India because of the incentives of different stakeholders. For example, there is the whole reference network (where a doctor refers a patient to a diagnostician or vice versa) which creates nexuses that operate outside the tech ecosystem. So it’s hard to solve a healthcare problem with only a tech-led approach. You also need some kind of an offline delivery mechanism."
5CN’s partnership with GenWorks thus proved as important for its scaling as its platform approach. “They picked us when we were at an early stage because they saw we were solving a pain-point for their customers," says Sivasailam.
Malavika Velayanikal is a consulting editor with Mint. She tweets @vmalu
