On September 14th, I had the opportunity to attend the OASIS (Open-Source Alliance for Social Innovation & Sustainability) Summit, thanks to the generous sponsorship from the Zerodha team. The main purpose of attending the summit was to explore how open-source technologies could be leveraged to provide Healthcare products and services at a lower cost, especially when we have a deep understanding of the problem statement and the price point that our target audience can afford. Additionally, I was excited about the chance to meet Nithin (whom I've long aspired to work with) and Kailash Nadh (the CTO of Zerodha and founder of FOSS United - Free and Open Source Software). Their philosophy of building Zerodha's tech has inspired us to believe that technology can make quality healthcare affordable for the bottom 90% of Indians. There's a humorous story about my first encounter with Kailash, which I will write in future essays.
See the smile on my face (Fan moment, lol), but that doesn't mean I don't disagree with his many POVs
As I left the summit around 5:00 PM, I was overcome with emotion, seeing the possibilities of making quality healthcare affordable for the bottom 90% of Indians through technology. I did not doubt that achieving the price point, we were aiming for, was possible, given the available technologies; it was just a matter of building upon them.
Returning to Bangalore after three years, aside from attending the event, I have been fully engrossed in work, preparing for our next meeting on the 21st, once again thanks to Zerodha and the Rainmatter team. :)
While taking notes, for the above meeting, on how the technology we're developing at Jile Health can benefit companies serving the top 5% of Indians in the health and fitness space, I was astonished to discover that the yearly subscription costs for individual packages averaged INR 50,000! This is the cost of online coaching and training. I pondered the rationale behind these high prices and concluded that it could be due to either a supply constraint (a shortage of such coaches) or a demand surplus (more people seeking personal coaches).
I have yet to verify these figures with a coach, but when I did the math, I realized that, on average, these coaches charge INR 1,000 per hour (INR 4,000 per month per client, managing up to 50 clients per month, resulting in total earnings of INR 200,000, with a monthly workload of 200 hours). In this scenario, the cost can only be reduced by increasing the number of clients or the number of hours worked. In any case, the chances of the price dropping below INR 800 per hour are slim. This means that fitness, nutrition, and diet coaches will only be affordable for the top 1 to 2% of India's population. If we want to provide personal coaching to a billion people, economics will never support it.
Having been on the ground for the past 13 months, interacting with patients and smaller healthcare providers daily, I created a document titled "All My Thoughts on the Future of LLM and AGI in Healthcare" shortly after the launch of ChatGPT. Most of these thoughts are based on real-world issues faced by patients, providers, and payers, with a 15-year timeline in mind, backed by in-depth research. Let's take the example mentioned earlier.
Given the current per-hour cost of these coaches, only 1 to 2% of Indians can afford their services. Yet, 70 to 80% of healthcare expenses in India are related to chronic diseases that can be largely managed through lifestyle changes. However, the current coaching model is affordable for only a small fraction of the population. To make healthier choices accessible to the next 40% of India's population, we need to ensure the availability of these coaches.
The good news is that unlike healthcare professionals (doctors) or healthcare facilities (hospitals, clinics), these coaches are not directly involved in core medical aspects. They all follow a fixed and standardized protocol, which can be taught in a few weeks or months. This means that by combining the protocol followed by these coaches with evidence-based data points, fine-tuned using Llama-2, and supported by Health Vitals extracted from health records using an open-source ABDM-enabled HMIS called Bahamni, we can convert this information into Indian languages using the open-source Indian Large Language Model Bhashni. We can then transform it into AI coaches (using Bhashni's text-to-audio capabilities), which can be made available via smartphones (WhatsApp bot) at almost no cost to the rest of the Indian population, that will deliver the same outcomes to the users. And Instead of paying for individual subscription packages, users can pay a fee to access all subscriptions for a year, similar to how Netflix offers its entire content library on a monthly basis.
One misconception I had early on about health records was that they were solely for medical purposes. While it's true that health records are used in medical contexts, their real value extends to non-medical use cases:
Risk Evaluation Profile for Health Insurance Premium
Reducing Health Insurance Premium cost
Fraud prevention
Incentivizing individuals to be healthy
Personalization of Healthcare products and services etc
You may already be aware of this, but let's revisit it anyway. In its initial iteration, Netflix used to rent out CDs of movies and TV series. Users could rent and pay for individual CDs for a few weeks or months. Eventually, Netflix introduced subscriptions, allowing users to rent an unlimited number of CDs in a month for a fixed price. And eventually making all these content available via smartphones through streaming. Can you imagine paying for individual movies and TV series on Netflix, or how many times you rented them?
There's a saying that sometimes users don't know what they want. If Netflix had insisted on offering only streaming services, it might have faced significant user backlash. Therefore, they initially offered streaming as a complimentary service free alongside their physical CD monthly subscription. Over time, users transitioned to streaming, which has now become an integral part of our entertainment experience, and individual renting seems antiquated.
The key distinction between Netflix and India's healthtech industry is that unlike Netflix, which invested billions in producing original content, we don't need to spend billions; we can leverage what is readily available (and nearly everything is available). We can build and distribute it in a way that aligns with user preferences, rather than what an individual organization thinks users want.
I sometimes wonder why consumers from the top two urban quintiles took longer to adopt UPIs or why sellers from the top two urban quintiles are not yet enthusiastic about listing on ONDC (especially considering the potential for acquiring users at zero customer acquisition cost), or why I don't see more people in cities having ABHA (Ayushman Bharat Health Account). I don't have an answer to that. However, If history repeats itself, as we all believe, and drawing from past digital revolutions, I predict that the future of healthcare and healthtech in India will resemble the evolution of Netflix. The "Netflix of Healthcare" appears inevitable, and individual health records are the billion-dollar streaming content that will make it accessible to a billion Indians. This could be the most significant technological impact on 1.4 billion Indians…
I am in Bangalore till Sunday 01:00 AM and if you think we should meet I would love to meet. Thanks for reading, I shall see you all next week
PS: When I wrote the first doc of Jile Health, the mission was - Making Quality Healthcare affordable for the bottom 90%, 1.2 billion Bhartiya. And in the past few months - I have written on multiple occasions - the dream of affordable quality Healthcare for everyone in the world is possible. Because I am not seeing any reason for not happening…
27 years old Suman posing with 23 years old Kailash Nadh (he looks 23, no?) who could be Carbon neutral at the individual level carries a handkerchief, and water bottle, and created a new philosophy on designing a tech stack