The Road Toward Well-Being — Notes From South Summit Panel
Here is the transcript slightly edited. The video is at the end of this post.
- The rise of symptom checking apps, and their benefits
- How new devices can predict, prevent and more
- Pros and cons of sharing health data
- The challenge or the 7,000 rare diseases that affect over half the population
- Why people like bots more than doctors
- How doctors benefit too
Ben (HAX): You are an entrepreneur with a health tech company and we are investors in about 30 med-tech and wellness device startups. It’s interesting we’re following the pharma guy because in a way we’re both putting some pressure on pharma. But let’s start with introductions. Daniel?
Daniel (Ada Health): Ada is what we call the world’s health guide. We give you an app that you can have a conversation with, like with your family doctor.
You open the app and tell about the problem you’re having. It will ask you a series of questions then it will give you an idea of what kind of condition is causing your symptoms and what steps you should take to address it. So basically we’re replacing the good old ‘googling your symptoms’.
Ben (HAX): Is it a kind of chatbot for medical diagnostic?
Daniel (Ada Health): Yes. We’ve been around for 7 years and spent the first 5 to develop a decision-support system for doctors. So the patient version that we launched about 1.5 years ago is only the tip of the iceberg of the technology that we built. The patient version is being used a lot more and became the number one medical app in 140 countries and every 2 seconds someone is entering a new request.
Ben (HAX): Amazing. Quickly about HAX: it’s the hardware branch of SOSV, a global early stage fund with $400 million under management. HAX has offices in Shenzhen — to build things — and Silicon Valley. Over the past 6 years we’ve done 200 hardware investments. We invest at the prototype stage with a first check of $100k and help iterate fast using our acceleration program in Shenzhen, then follow-on with up to $1.5 million in later rounds.
We started with consumer devices but about 3 years ago we started to look more closely at health tech. At that time Fitbit was hot and the question was ‘Is everything done now that we know how many steps we walk?’. Obviously there was more room.
Since then we’re invested in companies focused on diagnostics — including 4 blood-testing companies (testing pathogens, blood type, and more) or ‘closing the loop’ to give you instant feedback or directly heal you.
For instance, one of our startups called Flow Neuroscience uses brain stimulation to treat depression.
It’s all devices, but always with some A.I. or machine learning. Ada is pure software doing machine learning before it was cool, isn’t it?
Daniel (Ada Health): That’s true. 7 years ago, and until a year ago my friends didn’t understand what I was working on — the first few years I wasn’t either (laugh)— but there is clearly a shift in the industry and in the wider population about all these tools to manage your own health.
The basic need to find out what’s wrong with you has always been there — 9 out of 10 people google themselves before visiting a doctor, or even when they come back from such visit.
With doctors, more than 80% of diagnosis is based purely on patient history. Once it’s in the hands of patients — where this personalized health aspect comes in —the input will make it even more accurate. In sleep too there is so much you can do. We spent a third of our life sleeping.
Getting the data
Ben (HAX): The challenge is really getting the data. Let’s ask: who had a health check this year? (few raise hands) Everyone else: your data is more than a year old.
This summer I actually got a health check in the Philippines. There were two clinics right in a shopping mall. You just chose your tests on the menu: cholesterol, fertility, ultrasound, STDs ... For $200 and within two hours I had baseline data I never had before. Having data when you’re healthy and getting regular checks allow early detection of conditions.
One friend from Silicon Valley recently lost a lot of weight. He discovered he was gluten intolerant and pre-diabetic. He stopped bread, pasta, rice, cakes and potatoes. He also switched to a 16/8 diet. Now he’s not even pre-diabetic.
One thing we discovered though our investments is that you can have all the devices and data you want, the biggest contributor to health problems is behavior. And it’s the hardest thing to change. Do you have ways to tackle this with your software? Fitbit and others are trying to do that in some small ways with notifications or reminders. How about you?
Daniel (Ada Health): People use Ada when they already feel sick and have symptoms. As long as you’re fine, you don’t worry unless you’re one of the ‘worried well’.
Users tell us that finally figuring our their condition leads to change in behavior. There are over 7,000 rare diseases. A lot of people suffer because it takes 5 to 7 years to even get a diagnosis.
We have over 100,000 reviews in the app store (current rating is 4.8) and one said ‘I just wanted to test your app because last year I finally got a diagnosis for a condition. I asked 20 doctors and it took 10 years to find out. With Ada it took 5 minutes.’
We’re now trying to help users manage their conditions so it doesn’t worsen.
Ben (HAX): One new buzzword that emerged is P4 medicine: preventive, predictive, personalized and participatory. ‘In the future, donating your data will be more useful than donating an organ’. How about sharing data?
In the future, donating your data will be more useful than donating an organ.
Daniel (Ada Health): Users decide what to share and with whom. We’ll ask for permission to contact them if their condition matches clinical trials, so they can help advance medical research. Companies then choose if they would like to connect with users. So it’s a double opt-in.
Learning from data
Ben (HAX): Are there things you figured out from the aggregation of anonymized data?
Daniel (Ada Health): We have 5 million users, and expect 100 million by 2020. There is a lot of data coming in all the time and we can see certain trends. For instance there was an outbreak of Lassa fever in a country in Africa. We didn’t have it modeled in our database yet as a disease. After we modeled it we could clearly see a spike.
Ben (HAX): So you could associate the reported symptoms with the disease.
Daniel (Ada Health): Yes. There are a lot more things we can learn. For instance I didn’t expect such a high percentage of mental health cases. It’s almost 1 in 3 with a mental health component, and there is a lot of variance between countries. The more data we collect the more we can refine our model. Just by using the app we can help others. You can download it already in the app store in Spanish, English, German, Portuguese and French. We’re working on Mandarin, Hindi and Arabic.
From Treatment to Prevention
Ben (HAX): In the previous talk, the spending on healthcare was going higher and higher but seems to be maxing out. Are we shifting budgets from symptom-based reactive medicine to prevention?
The most motivated are generally not pharma companies and clinics but insurers and governments, who are very motivated to keep you healthy. Have you had talks with insurers, for instance?
Health care spending is maxing out. We will shift from treatment to prevention.
Daniel (Ada Health): Yes, of course. A lot. We are working with ‘payers’: governments, insurers and sometimes corporates — like in the U.S — and with health systems. We’re getting approached by those players from all over the world. In Germany we’re working with some of the largest health insurers, as well as in other countries, and with the NHS. Clearly they’re interested in:
- Having more engagement with their covered population,
- They see us as first layer of primary care, like the GP in most countries.
We can then help direct the patient to the right next step in the care journey. One big problem in many countries is that people go straight to the emergency room because they often don’t know better.
Another aspect is to be able to catch things early. Right now we’re preventing the $1,000 problem to become the $100,000 problem. The next step is to prevent it from ever becoming a $1,000 problem, so we’d love to work with some of your device startups for instance.
We’re preventing the $1,000 problem to become the $100,000 problem. The next step is to prevent it from ever becoming a $1,000 problem
The Challenge of Adoption
Ben (HAX): You mentioned you were surprised people liked to talk to your chatbot more than to a real human.
Daniel (Ada Health): Yes. Our background is different from many companies in our space. They started as Skype-like tele-health companies, which didn’t scale very well because you still need one doctor for one patient. So they built symptom-checkers and called themselves A.I. companies. Some of them also claim to be better than doctors — don’t believe it.
At the beginning we also thought we should offer the opportunity to speak to a human doctor at the end of the automated assessment. So we offered that in the UK, but the conversion rate was extremely low. When we surveyed user satisfaction, it was a lot higher with the automated Ada service.
We also saw that for certain conditions like mental health or sexual health, or in the Arab world people are more comfortable talking initially to a doctor. Coming back to the reviews, people wrote ‘I know it’s just a bot, but it already reduces my anxiety to be able to have a conversation like that.’
Ben (HAX): That’s interesting. In some professions, people are worried software is going to eliminate jobs, but in the health space there are not enough practitioners to take care of all the problems. So anything that allows a doctor to treat not 10 but 100 or 1,000 patients or see them at later stages of diagnosis allows them to have more impact.
Daniel (Ada Health): There is a global shortage of 7 million health workers. If you look at India or some parts of Africa, more than a billion people never see a doctor in their life. Now if you only need a smartphone that’s an amazing improvement.
More than a billion people never see a doctor in their life. Now if you only need a smartphone that’s an amazing improvement.
Ben (HAX): One topic: behavior change / human augmentation ?
Daniel (Ada Health): Our idea was to augment the intelligence of doctors, not replace them but empower the patient, and make the journey more efficient. Doctors can now start with a more detailed briefing to make a better ultimate diagnosis.
Ben (HAX): Back to the company Flow Neuroscience that makes a brain stimulation device to treat depression. It could potentially stimulate any area of your brain and help you memorize things better or else, leading to ‘human augmentation’, or being ‘super well’.
### Q&A with the Audience ###
Q: Describing symptoms for rare diseases is not always easy. How do you deal with that?
Daniel (Ada Health): We have about 10% of the 7,000 rare diseases. We can sometimes help describe symptoms better. Many of our 130 staff in Berlin, London, Munich and NYC are doctors and keep adding to our knowledge database.
Q: Is there a future where those solutions will be paid for by public money?
Ben (HAX): A few of our companies start by working with doctors or hospitals to help with diagnosis or improve their processes, and save time or money. Some get into the reimbursement system in which case it could be covered by public money. It’s on a case-by-case basis and it’s a long road: you need the clinical trials, the approval, and all the paperwork to get into the reimbursement system.
Daniel (Ada Health): For us it’s happening now.
- We’re in active negotiations with several governments like the NHS in the UK, and with governments in Asia and Africa.
- We’re already working with some NGOs to make it accessible to those who need it most in parts of Africa, with several foundations like the Gates Foundation.
- We also just won the ‘Frontlines of Health’ at the MIT Solve Challenge, and the “A.I. for the Betterment of Humanity’ prize by the McGovern Foundation.
It’s quite obvious: we want to empower patients to have better outcomes, but clearly there are also quite huge cost savings potential. This is why private and public health insurers are so interested in working with us. We are quite convinced we can probably save them 10% of their total costs thought the various benefits that we offer.
Ben (HAX): I think we’re out of time. Thanks everyone for your patience and your questions.
Daniel (Ada Health): Thank you.