Esther Dyson Q&A: On digital health and getting to the end game

Don’t be afraid to try. Work under the assumption that it will work and if it doesn’t work, figure out why. In a startup you are doing things people have never done before.” - Esther Dyson

Q & A with Travis Good and Esther Dyson

Travis Good, MD: Esther, you’ve said in the past that as an angel investor you like to focus your investments on disease prevention versus technology around disease treatment. Can you tell us a bit more about that philosophy?

Esther Dyson: In general, you get a better return on keeping things working and investing in capacity than you do in trying to repair things that are broken. To some extent, people sometimes take that for granted. They do general maintenance to their cars and they conduct home maintenance, but often forget when it comes to their own bodies. People generally don’t go to the doctor until they are sick and by that time, it is extremely expensive to try and fix the problem. It’s expensive not just for you, but also for the healthcare system. It’s also expensive for your employer and you are probably less productive there. For individual investors and for society in general, it makes more sense to keep us healthy. Obviously, a big market exists for fixing people who are broken, but given that I have limited resources, I’d rather try to put the people who are trying to “fix what is broken” out of business by keeping everybody healthy.

Let’s start with the simple proposition that health is an asset. It might be our most important asset. And like any asset, if we cultivate it we increase the value it returns. We’re not talking about fixes when health is broken; those are important but they happen too late. — Way To WellVille

Travis: I would agree that obviously prevention is the ideal path so that makes sense from an investment standpoint, as well. I want to shift gears to something you founded in 2013. You rebranded the Health Initiative Coordinating Council (HICCup).

Esther: In 2013, I founded Way to Wellville, which is a nonprofit and it does exactly what we are talking about. The investment pays off but the business model doesn’t necessarily pay back to the investor, so that’s why Way to Wellville is a nonprofit. We are trying to show the value of investing in health. Ultimately, governments, communities, and employers will realize there’s a huge ROI. In 2014, we selected five communities from among 42 that applied. Now, the Wellville team of five are working with actual communities to help make them healthy. By the end of ten years, we hope to have enough success and visibility that we actually persuade other people to follow our path.

These are all real places with real people. They are not social experiments; they are not new towns, but just places that have decided they would like a better future for their residents. [Wellville is] helping them build that for themselves.

Travis: My understanding is that the goal with these communities is to see which one could most improve five measures of health and economic vitality over a five year period. That was 2014; we are almost in 2018 now. Could you update us on how that’s going?

Esther: As I mentioned, we pivoted. It did start out as a contest, but it turned out the communities were not interested in beating one another, they really wanted to collaborate and learn from one another. It also became clear that ten years is a more realistic timeframe. Like any startup, we’ve morphed. It turns out that getting communities and community leader groups coordinated and organized has a long gestation period. It’s very much like a startup; it takes a few years to build your team, try out an NVP, then discover that you need to change everything. We are now much more focused on children and parents and prevention of diabetes. If you can help the parents be better parents, or you change the children’s experience positively, that’s what’s going to pay off best in the long run.

Discussions now center on the opioid crisis; they keep talking about the drugs, but they should also be talking about why people are so vulnerable to these drugs. What makes them unable to resist? Not everybody who takes an opioid suddenly gets addicted, yet a number of people do and that’s partly because their childhood was so insecure. It’s very hard if you’ve grown up not being able to trust the people around you. There’s a lot of data on so-called adverse childhood experience that show direct correlation between that and addiction, unemployment, crime and all kinds of social ills. So we are trying to go upstream and address it from the start.

Travis: Are you working with the communities to define those measures and strategies, or are they coming to you?

Esther: We put out a call for applications in 2014. Forty-two communities applied; we picked five. Our role is like a personal trainer, or perhaps a startup coach. It’s still the community that has to do the work; it’s their people; it’s ultimately their achievement. We might eventually make ourselves unnecessary, but in the meantime, we can give them advice, introduce them to what the other communities are doing, introduce them to vendors and to potential funding sources. Way to Wellness was funded by me and I’d certainly love others to join, but the community projects are funded directly by funders, whether it is Medicaid, or philanthropy or a government grant; it’s between those funders and the communities. It’s theirs, not ours and we are not looking to be visible or recognized. We are looking for them to take on the responsibility, do the work, make a difference, train their own people and have something to be really proud of.

Travis: It’s exciting to hear when people are working with communities, especially some small communities. What happens after that initial 10-year period?

Esther: The goal is to build something sustainable. Instead of, “Oh we have a class for 60 overweight women.” We want to hear, “We have now trained 30 dieticians and nutritional counselors, exercise coaches and they are community members, and they now have jobs and they are keeping those people healthy.” This is the problem: Just as people get addicted to drugs, ironically, communities and nonprofits often get addicted to pilots. The pilots are great. We often hear, “This pilot’s going to be really wonderful and we’ll have a trial with 30 people.” The pilot works then we often hear “We’ve run out of funding and the funders want to do something more innovative.” We are not trying to do something innovative; we are trying to do something that works and then scale it.

When Walmart opens a store, they are not running a pilot to see if the store works. If the store somehow isn’t working, they find a new manager, they do more training, but they don’t say, “Oh, the concept of this really doesn’t work.” We are trying to get that kind of accountability. These programs work, if effective people are brought in, screened and trained. People need transportation, as well. There are lots of local challenges that you need to overcome. There’s no magic here; there’s simply implementation, persistence, and being in it for the long-term, versus “we’ll get some results and then see what happens.”

Travis: As you were talking through that in terms of building something that is sustainable, real, and can scale, I imagine that’s the same thing you would say to some of the companies that you invest in and how they should view pilots.

Esther: Yes, very much. Don’t be afraid to try. Work under the assumption that it will work and if it doesn’t work, figure out why. In a startup you are doing things people have never done before. People have done diabetes prevention programs before. They have trained communities better. What we are not doing is coming to these communities and doing exotic research and turning people into specimens. We are helping them do things that typically don’t scale. Instead of doing something that is often different, you are trying to do something new that hasn’t been tried before and learn from failing. In our communities, you learn from failing. Maybe we need to train people better or maybe we didn’t market this program effectively enough or we need to provide some type of incentive or we need to provide child care so that people can attend the course, because hey, they don’t have childcare — what a surprise! It’s those kinds of things.

It’s very easy to guarantee someone privacy, but you also need to make their data usable. Many companies are trying to do that, but they don’t understand that they need to train users and understand the workflow in hospitals. Many companies think they have a solution, but they don’t necessarily know how to actually get it installed and working in the actual healthcare environment.” -Esther Dyson

Travis: I want to shift gears a little and talk about one of the company that you are invested in, I read something about a merger between Digi.me and Personal. In terms of joining forces, they are looking to address the challenges around enhancing privacy with healthcare data. But at the same time, enabling people to share and potentially combine and analyze data. As you think about your investment there, can you share your thoughts on the importance that must be made by healthcare IT companies on privacy and security?

Esther: Privacy and security are really important but to some extent, there are regular privacy breaches, which is an issue to fix. Privacy and HIPAA often obstruct good healthcare. What you need for good healthcare is to know more about the person, not just their result on a particular test. Where do they live? What is their income level? In case of opioids, have they applied for drugs at any other place? Do they have multiple sources of supply? It’s not a complicated thing, it’s having both security, privacy and effective data sharing. That’s what is challenging. It’s very easy to guarantee someone privacy, but you also need to make their data usable. Many companies are trying to do that, but they don’t understand that they need to train users and understand the workflow in hospitals. Many companies think they have a solution, but they don’t necessarily know how to actually get it installed and working in the actual healthcare environment.

Travis: How do you advise companies to help them navigate that path towards implementing the solution and into the workflow of the health system?

Esther: I’m generally pretty skeptical. I was just at the doctor earlier this week and had blood tests. I got the tests back and they were in a PDF. What I’d really like to see is my FitBit data and other personal data, to determine how the blood tests changed over time. What I had was a long list of PDFs from various dates that were completely useless. Part of the problem is that most of these institutions are using large legacy mainframe systems that are incredibly cumbersome. You really start to think: “Throw this all out and start fresh.” That’s what a number of vendors are doing but of course, it is very hard to replace the huge installs of these systems.

I’m on the board of a company called Wellpass, and invested in quite a few others, including 4D Healthware, Healthcelerate, HealthLoop, Vital Score. They all sell into healthcare systems, and need to understand how they work. I don’t think a twenty-two year old who just graduated from Stanford can know how to fix the healthcare system. You need to have someone that has actually worked in a hospital or an insurance company on your team. That’s not just so you have contacts, but more importantly so that you understand the context, how the system works, how the data actually flows. I went to a continuing medical education conference once; they had this table where they were handing out medical advice for doctors. There was a stack of stapled handouts — twenty-five sheets each, double-sided, so fifty pages per handout. Each page was a spreadsheet about some particular thing the doctor should do with their patient during the patient interview — which of course lasts max fifteen minutes. As a doctor, they are probably spending eight or nine minutes typing and and five minutes actually talking. Now here’s this sheet of fifty things they are supposed to be doing; some of them they’re already doing, and some are only for teenagers and or for women. But you, as a vendor, you have to understand that you are either replacing one of those fifty things, or adding to them. Do you really understand that context? Can you fit into it or make it better with your extra page? Of course, some things are for the administrator, or the nurse discharging patients. But whatever your tool does, how does it make those people’s lives better and easier?

That’s the challenge. Healthcare is a very cumbersome and change-resistant marketplace. The companies that have been the most successful are the ones that are eroding things around the edges. I love this marketplace because it is interesting and important but, if you want to be in this marketplace, you need to understand the challenges.

Travis: As we think of how to augment, expand, extend, or in some way work with the EHR and EHR data — taking that data and transform it into something a bit more meaningful than a CCD or some big PDF — I’m curious what you are finding there.

Esther: One company out there called PicnicHealth will take your medical records and automatically parse them as much as they can and then have humans do the rest. They will have people on the phone calling to get your medical documents from all these different vendors or different providers and they will make them all into one piece that is intelligible. That is really an exciting proposition. As I was just talking about my blood draws, I realized I should just call them up and see how they are doing and use the service.

The challenge, of course, is that there are so many different varieties and layouts. There is an accepted medical vocabulary but there are always little exceptions here and there. A blood value is not something you want to get wrong. For the patient, but also for the doctor, you want to make sure that if you are comparing data from multiple sources, you want to make sure the data is absolutely consistent across all these different sources. It’s a challenge and there is both a real danger to people and real liability to companies if they do it wrong. So, it’s going to be a slow process. The more standardized things get, the better. Whether you are on a Mac or a PC, we can amazingly have this conversation, across different kinds of computers and it is working, more or less.

Travis: It is interesting to see how different companies are approaching that challenge, both from a consumer standpoint and from FHIR and direct-to-consumer enablement authorization of data.

Esther: At one end, somebody like Apple and HealthKit is really fostering that. It’s possible; consider how it happened in banking, with Intuit. They had this tool Quicken that would let individual bank customers interact with their bank portals. Intuit started by reverse-engineering the interfaces of a couple of big banks - probably Citibank and BankAmerica (I’m not sure) - so users of those banks could download their own data and manage it for themselves, even consolidating data from multiple accounts. At first, the banks were very resistant but after the first two or three were done, and the customers started engaging more with their banks, the other banks noticed. They started saying, “Wow, they are not losing their customers; in fact, those banks are getting even more buy-in from their customers. We want this, too.” They went to Intuit and said, “Please do an API [application program interface] for us.” Suddenly the dynamics had changed. Intuit came back and said, “If you want to hook up your banks to us, then you have to do the work, but sure, we’d be glad to let you address our customer base.” With luck, that’s what’s going to happen in healthcare. There are going to be a lot of vendors now that are interested in working with HealthKit and other intermediaries. You are beginning to get this widespread recognition that, of course, individuals should be able to see their Fitbit data matched against their RunKeeper data and someday matched against blood draws and all the other stuff.

Travis: I want to follow up on something else you said. There’s been a ton of investment in the digital healthcare space over the last five to ten years. You mentioned about having someone with experience on the actual care side. Can you share two or three consistent success threads that help take a startup from a seed investment up through multiple series, from A to B to C?

Esther: Basically it all comes down to the team. You need to know something; you need to work together effectively; but then you need to be willing to learn and change into it; be agile and be humble and all this kind stuff. It’s really the attitude of “I love what I’m doing, I’m learning from what I’m doing and I’m responding to what I’m learning.” Then, it helps if you have a good idea and you are working with interesting problems. There are lots of specifics to health or healthcare but the fundamental things that make startups successful are pretty much the same.

Travis: Curious on team, and investing in team, as you look at presentations and pitch decks, do you have an opinion on an ideal team or an ideal mix of people on a team that you like to see and you like to invest in.

Esther: Very clearly, it doesn’t consist of the three best people in the world, it consists of the best team of three or four, or whatever. A lot of people miss one out of the three or four, but you need someone with medical industry knowledge and connections; you need someone with technical capacity to build what you need; you need someone who understands, whether it is the consumer or the healthcare system that you are selling into; but, you also need someone who understands the user proposition, especially in healthcare where often the payer is not the same as the user. The payer may be the insurance company or the individual, but it may be that the payer is the hospital, but the user is a doctor or a nurse.

It goes back to the old software industry, where there was this wonderful set of companies called value-added resellers (VARs). They would take the product and go find a corporate customer. They would train the IT department and help IT install the software. Then they would train the users — the employees making phone calls or doing accounting or customer support. The VARs understood it was their problem to make the buyers — and the buyers’ employees — like the product. Without VARs, the software vendors’ customers would usually pilot the software and think it was great, but then they would buy it and find out they couldn’t get their employees to use it. The vendor got the money the first few times and started getting bad references for the product.

I think we need a better understanding of the role of VARs (or whatever you want to call them these days) in the healthcare market. Whether it’s a third-party VAR or your own sales support, you need someone who understands how the hospital works, or how the insurance company works to insert their product into the workflow and actually make it useful and make the end-user customer say, “Oh my gosh, this makes my job so much easier. I can do referrals automatically instead of having to handwrite something on a piece of paper and give it to my nurse and she puts it on my desk and it gets lost and I don’t get the referral back and forth with the other doctors.” And, of course, the patient doesn’t get the benefit of going to see the dermatologist.

Travis: Before we conclude the interview, Esther, I wanted to ensure we covered everything on the Way to Wellville so that we didn’t miss anything there.

Esther: Well, I did want to mention the five Wellville communities, just in case in case anybody listening either lives in one of them or is nearby. The Wellville partner in Muskegon County, Michigan, which is the one that I work with most closely, is called 1 in 21. The Wellville partner in Spartanburg, South Carolina, is called Spartanburg’s Way to Wellville. There’s Hope Rising in Lake County, California, which is just north of Silicon Valley. There’s Way to Wellville: Clatsop County, northwest of Portland, Oregon. And there’s the Triple Aim Collaborative in North Hartford, Connecticut. These are all real places with real people. They are not social experiments; they are not new towns, but just places that have decided they would like a better future for their residents. We are helping them build that for themselves. You can find us at wellville.net.

Travis: Wonderful. Thank you for joining us today, Esther.

Esther: Likewise, it was great. Thank you very much.