This YC-Backed Startup Is Digitising Healthcare In ‘Bharat’ With ML For Health Data

India’s digital healthcare future is taking shape slowly but surely. The recently approved health data management policy gives some much-needed clarity to startups and private companies looking at the national digital health mission (NDHM) as the launchpad for healthtech innovation. 

Leveraging the digital health infrastructure in partnership with software think tank iSpirt and various other stakeholders, there have been a plethora of healthtech startups working to digitise healthcare records in urban and rural parts of the country, which is the first big gap to be solved.  

And while startups such as Farmako, Tattvan E-Clinics, WonDrx, Portea Medicals, Lybrate, DocsApp, MedCords among others are solving the healthcare data deficiency, there’s a whole lot more ground to be covered to match the NDHM’s ambition of digital health IDs for all Indians. 

But even as a majority of these startups are optimistic about the recent policy development, healthcare data is no more a differentiating factor, but what a startup does with the data. 

“Collecting health data is not mainstream anymore as everybody is doing it, but what we do on top of this data matters,” claims Aman Bhandula, founder and CEO of Farmako, which has created a machine-learning based platform for healthcare businesses to digitise healthcare through electronic medical records (EMR), as well as assisting doctors in prescriptions and medicine delivery. 

Founded by Bhandula along with fellow IIT Roorkee graduate Kaishu Sahu in 2019 in college, the idea started with simply digitising health records for patients and developing a digital prescription platform for doctors and clinics. But after pitching to Y-Combinator in the same year and being rejected, the team added more capabilities. 

After applying a second time, Farmako did make it to YC’s All-Virtual Summer Batch 2020 by proving its use-case. In just over a year, it brought close to 1500 doctors on board from Tier 2 and 3 cities, who are using the EMR app for writing digital prescriptions, out of which, 400+ doctors are using the premium service. This includes a Farmako tablet, printers and other value-added services to help doctors digitise prescriptions and automatically store patient data onto a centralised health record system via a virtual health card. 

In fact, the Farmako EMR and health card is integrated across the startup’s network of hospitals or clinics. So when patients walk into any nearby hospital, their records can be accessed in a timely manner, without any doubts or concerns about the authenticity. The company also claimed that its solution is inclined with the government’s plans for a centralised health record system.

Beyond this, Roorkee-based Farmako is using proprietary machine learning algorithms on top of the data to categorise health data systematically, to give patients and doctors quick access at the time of need. This categorisation will be shared upon user consent. Currently, the company claims to have digitised health records of about 40K patients. 

“If you have to upload each and every record by clicking pictures, it would be really difficult. What the government is also trying to do essentially is that if a doctor is using an EMR app, then they would connect this with a health ID so that they’ll have a centralised database,” added Bhandula. 

Farmako’s Machine Learning Platform

Besides Farmako, healthtech startups such as Practo, Doxper, WonDRx and others are also trying to solve a much similar problem. Even digital prescriptions are something others in this category are doing, but Bhandula claims Farmako is faster than the competition. 

“We have built an EMR app, where doctors can write the prescription as fast as writing it on a paper, taking less than a minute,” he claimed, adding that OPD are busy as they have to treat close to 60 patients per day. Farmako is said to be 4x faster than the competition, though these claims cannot be independently verified. 

While digital prescriptions are undoubtedly great for structuring data, it’s not a natural action for many doctors and often takes longer than writing prescriptions. “That is the reason many clinical management tools do not have the efficacy needed,” WonDRx cofounder Pankaj Sindhu told us in an earlier interaction. 

Bhandula said Farmako tried experimenting with machine learning for handwriting recognition but found it “practically not possible,” and he also claimed that it does not solve the big problem, which is saving time. 

“Typically, doctors prescribe 30-40 medicines and that can be digitised using machine learning. If they are writing this (the frequently prescribed medicines) also, it doesn’t make sense. So handwriting recognition is a great tech, but it doesn’t solve the bigger problem,” he told Inc42.

Instead, Farmako offers features like AI-powered voice assistant tools to recognise multiple languages and canvas, which is a text recognition tool for doctors to write while providing special instructions to patients and more. But the company stopped short of suggesting medicines as it would have scared the doctors. 

However, Farmako said that medicines options are available on the app. For instance, once the doctor writes the diagnosis and treatments, suggestions are shown automatically, otherwise, doctors will have to search for it manually and select the medicines. 

The Unit Economics Of Digital Healthcare

In October this year, the company raised a total of $1.5 Mn in seed funding from various investors, including $ 150K from YC, besides HOF Capital and others. With this funding in place, it is looking to expand to five cities over the next six months, including Moradabad, Bareilly and Meerut among others. 

At present, a large majority of the patients using Farmako are being brought on board by doctors, while some are organically using the health card to store health records. Category wise, paediatricians, gynecologists and orthopedics departments have seen the highest adoption, where patients are using the app to store their medical records in one place. 

In terms of customer acquisition cost for the doctors, Bhandula said that it hovers around the $200 (INR 15K) mark for premium customers. This would include the overall setup cost, including tablet, printers and other services. In return, it charges $50 per month from doctors who are using the app. Adding new doctors, hospitals and clinics is a matter of just 20 minutes, claimed the founder, who is targeting 1000+ doctors in the next ten months.  

Of course, the big question is revenue and this is where Farmako needs to continually add services to upsell doctors, hospitals and clinics. Bhandula says it’s early days to share revenue numbers, but he said the growth is 40% week-on-week. In the coming months, it is integrating medical stores, insurance and health plans onto the platform to diversify its revenue sources. It’s a proven strategy, as the company claims to be earning steady revenue with 300 orders per day after it started aggregating medicine delivery. 

There’s another challenge, and that is part of every domain that is being influenced by artificial intelligence or machine learning. From an AI standpoint, many policy experts suggested that these tech solutions might bring in efficiency in the process, but it becomes crucial for the government to come up with an ethical framework. The question of, ‘if something goes wrong, who should be liable?’ 

Farmako’s founder clarified that the government should come up with highly regulated standards for AI in healthcare in terms of the use of the data and the extent of recommendations that it can give a doctor or patient. “At this point in time, we are just providing a tool for the doctor to use and not replacing them, i.e suggesting medicines, showing relevant health records, etc. which is very underlined.” 

Note: We at Inc42 take our ethics very seriously. More information about it can be found here.