Innovations in healthcare are very much needed by society to keep healthcare affordable, but many entrepreneurs are reluctant to start a health startup due to the extra complications in this domain. But is it really more difficult to start a business in the medical domain? In this interview we ask this question to Marieke Schoutsen, who has made the jump into the medical domain and is now CEO of the health startup Earlydoc.
What is Earlydoc exactly?
Via Earlydoc we want to provide everyone with online reliable medical analysis. Most online medical information is generic and scary. Earlydoc is different: it gives you the information that your doctor would give you. Ultimately, this should lead to less visits to the doctor. We provide a website where you can fill in your complaints and the system then tells you whether you should go to the doctor. Eventually we want to provide people with a very intelligent online medical file.
When did you start Earlydoc?
I founded Earlydoc one and a half years ago, in March 2012. I had just got my masters in Medical Informatics. During my masters I worked on my thesis at the Leiden University Medical Center, which focused on medical decision support, and I got this idea for a freely accessible medical advice service. My idea then was that such system could also function as a ‘flagship project’ like the Nuon Solar car that many student teams have been working on. After graduation I decided to try and build such a system.
Did it take more time than expected to get where you are now?
Yes, both the technical side and the content side have been challenging. We have been working on it now for 1.5 years. And I guess I expected and wanted to move even faster than we have. Building a first system was not so difficult, but you need to fill the system with the right content that has been validated by doctors. Also finding investors in the Netherlands and finding the right team members takes up a lot of time.
How did you start with the company?
I spent the first three months on building a prototype system myself, in PHP. I am not a strong programmer but I know a lot about front-end creation and user experience design. When the first prototype was available, I got 15,000 users without much marketing effort. That’s when I started looking for investors. This also took a few months. When the funding was secured, I could start finding a technical co-founder and other team members.
What did you do to find the right technical co-founder?
I interviewed at least 20 people before finding Travis Dunn. I was not looking for just any developer that understood the product, but needed someone that is able to be a CTO. It was a frustrating process, because with most candidates I had to explain the technical issues to them, instead of the other way around. I eventually found Travis via Github. He has a very good understanding of what it means to work with medical data and there was an immediate ‘click’.
Many people in IT believe that dealing with medical data is hard due to additional regulations. Is this true?
It depends. Our system takes most of the medical data directly from patients who type it in themselves. This type of medical data is not more regulated than other private data. Basically you need to do three things:
- You need to report to the CBP that you are storing personal data (this is something many startups forget)
- You need to make sure you have a very good level of security.
So in our case it is not true that medical data is harder to handle than any other data.
So when do you need extra measures as a medical startup?
If you receive information from medical professionals (general practitioners or hospitals), then you do need to meet extra demands. One way to meet these demands is to get CE-certification for the software. We started this week with this process, together with a lawyer. We could not find any firm specialized in CE certifications, so we have to find out ourselves what the process entails.
What are the additional costs, such as legal costs, that such certification entails?
Until now we have not had any large extra expenses. Earlydoc receives much support from advisors who appreciate what we are doing and are willing to help us. So the additional costs for working in a medical domain are limited, for now.
You offer your system for free on the Internet. Is this sustainable in the long term, and if so, how will you monetize?
We want to keep the system available for everyone without charging money for it. For one, this is important to many of the medical professionals that are involved in validating the system. We have been very careful not to become too commercially focused, because many stakeholders in the Dutch medical world are wary of overly commercial initiatives. We are exploring a few business models. One of them involves medical insurance companies. We save them money by preventing costly doctor visits, and if we can quantify this, they are the most logical customers of Earlydoc. Secondly, we provide an appointment and information feature to general practitioners, making their practice more efficient and reachable at all hours. Another possible model would be to provide Earlydoc to employers as a way to manage absence. Earlydoc can aggregate data that we collect to provide general statistics to employers. We do not share any individual information of patients using the system, but rather: general statistics about types of absence, absence predictions and trends.
For your system to be useful, you need to input medical knowledge. Do you create this knowledge yourself, or do you use existing medical handbooks?
Our starting points are the existing protocols and literature that is already used by GP’s and GP assistants. However when you want to create decision models from this, you often have to fill in the gaps. We do this by asking doctors whether they would want to see the patient in such a situation. One of our largest efforts is the validation of our system.
So how do you validate your system?
All or our content is checked by doctors. We talk to them about our logic and show them output with different scenarios and they give us feedback. We also validate by placing the system in doctors’ waiting rooms and have patients use it. The patients take their answers and advice with them to the doctor in the consult and then the doctor can check or augment the outcomes of the system against his actual judgment of this patient’s condition. This way we not only gain trust that the system’s advice matches that of actual doctors, but this way doctors also gain confidence in the outcomes of the system.
Your company is located at ‘Boven de Balie’, a collective office space for startups and entrepreneurs. Why and how did you start this initiative?
When I started with Earlydoc, my Boven de balie business partner had a similar space here in Amsterdam, that was always fully booked with many people still applying. It made sense to start looking for another suitable space. We are not in the very center of Amsterdam, above a restaurant/cultural center.
Is it subsidized?
No, we charge a fixed amount per desk per month. The difference with normal office space is that you can rent per desk, that you sit in a shared space with likeminded people and that we organize extra events to connect people to each other. The rent we charge is significant so we have to provide extra value, for instance by organizing events.
You started your company in The Netherlands. Was this a conscious choice?
When I started Earlydoc I was not thinking as big as I do now. I just wanted to build this first system. Along the way, and also influenced by talking to investors, I have started to think bigger. In hindsight The Netherlands is not the easiest market for companies like Earlydoc. The US health market is much bigger, less fragmented. The potential customers or partners in Holland in the healthcare sector move slower, because there is less of a commercial incentive for them. Also in the Netherlands there are few investors. I also found that investors in for instance the UK are more likely to help you create a strategy that fits going international, and see more potential in early stage companies.
What other markets are you now looking at?
In principle our technology can be used in any market, even outside health care and we might license our technology for other domains. We cannot enter all markets at once, so we are focusing on medical markets that are either large (i.e. US) or similar to The Netherlands (i.e. Germany). However every new market requires you to develop new content and validate it again. There are different medical protocols and cultural differences in every market.
Did you or do you participate as a company in incubators, accelerators and business competitions?
With Earlydoc we try to focus on our business and we don’t participate in many events. The only event we have participated in was the Accenture Innovation Awards, because 5 people independently recommended me to look at it. We ended up in the finals and gained some publicity.
We have not applied to any Accelerator program because we already have a network of advisors that help us. We do not need an accelerator program for that. I also try to limit my networking time. It is valuable only up to a certain point.
Looking back at the past 1.5 years, what would you do differently?
In the past year we let ourselves be influenced by outsider suggestions and have add many extra features to our system, such as making appointments. These extras are nice but they are not important to the core of our product, which is decision support. Looking back I would not have done these nice extras but would have focused exclusively on making progress with our core product.
About the author: Sieuwert van Otterloo (twitter: @entreprenl) is a management consultant with a focus on IT-enabled business. He gives software related advice via the Software Improvement Group, startup and innovation advice via Inbys and manages his own venture fund Otterloo Ventures. He writes for Startupjuncture and Frankwatching.