Hacking Healthcare:
Modernizing Healthcare IT
A chat with Piotr Orzechowski, CEO at Infermedica
Healthcare innovation is at an all time high. 2014 had record breaking investments in digital health, and 2015 is keeping pace. According to Rock Health’s mid-year report, funding in this space is growing faster than VC funding overall for the second year in a row.
Despite all of the noise in the healthtech world, we’re at a crossroads. Software is eating the world, as Marc Andreessen noted years ago, but healthcare is still lagging behind for a variety of reasons — expensive, legacy systems, a difficult regulatory environment, or the simple fact that people don’t always do what’s best for them.
At BaseHealth, we believe that the developers and engineers behind today’s most promising technology are the ones who will forge the road ahead, invent new solutions to legacy problems, improve the healthcare experience, and ultimately create a healthier future.
That’s why we’ve started this series — to recognize and celebrate the people who are making healthcare better via behind-the-scenes innovation. We are giving them the spotlight they deserve.
Our second interview is with Piotr Orzechowski, the CEO of Infermedica, a startup building a medical diagnostic engine to help developers build smart, AI-based applications that help doctors and patients make better decisions.
Piotr is passionate about artificial intelligence, machine learning and clinical decision support. He is an entrepreneur with a strong technical background. He spends his free time playing guitar. You can follow him on Twitter at @_orzech.
What are you working on right now and how will it impact healthcare and/or the developer world? What are you most proud of in your work?
For a couple of years now I’ve been developing a medical diagnostic engine at Infermedica. Given a patient’s data (such as symptoms or risk factors), our engine computes probabilities of likely conditions and suggests relevant questions to ask. While there’s still a long way ahead of us, you can think of Infermedica as an API for medical diagnosis. Our goal is to make this technology available to all healthcare IT developers out there so they can power their products with Artificial Intelligence.
What is the most promising sign in healthcare today? What gets you out of the bed in the morning?
Being an engineer at heart I’m really excited about the digitalization of medical data, especially when it comes to electronic health records.
Right now most of patient data is kept in an unstructured format (and frequently in paper form) which makes it difficult to extract meaningful information. Fortunately, that’s changing and it’s a very promising sign. Imagine you had all of your patients’ data structured and well-codified so you could write sophisticated queries just like you do for any other database. Apart from being extremely valuable for business and management purposes, having an ability to search thousands (or millions) of patients records by symptoms, history of conditions and their correlations is a great opportunity for a clinical decision support, clinical trial enrollment and medical research in general.
Besides regulation, what else is holding healthcare IT from truly modernizing?
I will limit my answer to the field I’m familiar with, which is the clinical decision support. I strongly believe that a wide adoption of computer-assisted diagnostic support or treatment planning will significantly improve patient outcomes and would be of a huge economic gain.
Even though technologically we are almost there — think of leading players like IBM Watson, GIDEON or Isabel Healthcare — there are still some obstacles. The usual answer here is that doctors’ mistrust slows down the clinical decision support from spreading.
In my opinion the reason is different — none of the existing tools are fully compatible with the way clinicians work.
These tools just don’t fit clinicians’ workflows as they are too time consuming to use or don’t provide enough really relevant information. We should do our best to create solutions that are efficient and effective in time-limited environments. The question is whether we’ll see the widespread adoption of healthcare AI in two, three or ten years.
How will APIs transform the healthcare industry?
APIs will play a significant role in transformation of healthcare technologies because they allow startups the space to move and innovate much faster than large-scale players and organizations.
There is already a number of great APIs such as ones offered by Eligible, Validic, BetterDoctor, PokitDok, Medline or BaseHealth. What we do at Infermedica is also API-driven. An important fact here is that there are more and more API-friendly EHR providers, like AthenaHealth, Dr. Chrono, or Practice Fusion. This creates opportunities for many healthcare engineers.
What advice do you have for young developers getting into the healthcare world?
Healthcare is one of the most challenging domains to start your project. Whenever you have an idea, always consult with someone who has extensive knowledge and experience in this field — usually a physician, nurse, healthcare manager or a patient who went through what you’re trying to solve.
Also it’ll be much more difficult to apply classic customer development methods since you will not have access to the real users for a long time. Find your medical co-founder as soon as possible. I wish I knew that before! Never rely on the assumption that doctors will simply buy your product (no matter how good it is).
If you could do it all over again, what would you have done differently in your career and/or research to date, so as to have a bigger impact today?
I think I’ve made a ton of mistakes along the way and a lot of things could have been done better. We pivoted our strategy probably 7 times before Infermedica reached a model that has gained traction. Yet I wouldn’t trade my mistakes for the right choices. Though I do wish I had a medical co-founder early on — that would have saved us a lot of time!
Want to be featured next? Leave a comment about the working you’re doing in the comments below.
