Everyone converges on a central theme, but we all bring our unique perspective
Earlier this month, I had the opportunity to participate in an event aiming to innovate and disrupt some of healthcare’s most intractable problems—the Healthcare Grand H@ckfest. A hackathon hosted by MIT H@cking Medicine and sponsored by big names—such as Kauffman Foundation, Joslin Diabetes Center, Global Genes, and the Massachusetts General Hospital. The event brought together over 400 people to meet, form ideas, and build prototypes—all within 48 hours. Participants traveled anywhere from 1 mile to halfway around the globe, bringing their experience as designers, developers, entrepreneurs, HCPs, researchers, and more to address the issues affecting healthcare today.
And They’re Off
Outside the Massachusetts Institute of Technology’s Media Lab, a placid, partially frozen Charles River lay before a sedate Boston skyline; a peaceful Saturday morning, in contrast to the energized excitement building in the room. Though we had all brought different perspectives and ideas, we all shared the common desire to improve the state of healthcare. After keynotes and a networking mixer Friday evening, Saturday morning was a feverish rush of meetings and brainstorming. The floor was open for anyone to express a problem or pain point, and it was up to everyone else to try to come up with a solution. Teams formed organically as individuals bounced from group to group, refining potential solutions until each team had solidified around 1 idea by early afternoon. We now had 24 hours to show that our solution could work.
Healthcare Challenges That Hackathons Can Address
The pitches were separated into 5 tracks, focusing on diabetes, rare disease, telehealth, global health, and hospital IT. Despite each track presenting separately and earning separate prizes, I was able to watch some of the pitches and final presentations for both the Hospital IT and Global Health tracks (I was in the rare disease track). Echoing the collaborative essence of the event, it seemed to me that a lot of the problems or solutions from 1 track could equally apply to 1 or more of the other tracks; particularly when it came to improving the quality of doctor-patient interactions or managing care.
Some of the greatest hurdles leapt were about improving the quality of information provided from tools to physicians. These included high-tech solutions involving wearable technologies (i.e., Google Glass) allowing an HCP to sit face-to-face with a patient rather than huddle over a computer screen as he/she reviews charts or inputs information. Or, relatively low-tech involving SMS messaging to relate the status of remote equipment or a patient’s therapy adherence to a central monitoring center, which could be easily reviewed by a physician.
Tackling Rare Disease Diagnosis
My team, Facial Metrics sought to tackle the problem of rare disease diagnosis, hoping to reduce both the average time to correct diagnosis (currently over 7 years) and the number of misdiagnoses. With over 7,000 rare diseases manifesting any number of symptoms—rare or common—it’s difficult for any doctor to accurately recognize even a handful of these disorders, let alone know which specialist to refer a patient to. But, the majority of these thousands of disorders are caused by genetic abnormalities that can manifest as abnormal phenotypic expressions during embryonic and childhood development. Using a facial recognition app developed by our mentor, Sharon Moalem, MD, PhD, our team was able to measure some of these abnormalities and correctly identify heightened risk for specific disease states from photos of people with those disorders.
It all works from a simple photo of a face. From this, a map of the face is extracted, identifying key landmarks (e.g., corners of the eyes or mouth, pupils, tip of the nose, etc.) of the individual’s face. From these points, relative and absolute distances can be measured and compared to what is considered a normal variance for the respective measurement. A value outside the normal range indicates risk for an associated genetic condition. While not a true diagnostic tool, it can aid a physician in narrowing down possibilities to just a handful and informing him/her about what tests to perform next.
What the Team Accomplished
A hackathon is not about building a finished product; it’s about proving a concept. No, my team’s app will not be available in the app store next week, but we did show that it is possible to programmatically correlate genetic aberrations with phenotypic abnormalities from nothing more than a photo. Yet, there remains a lot of work to be done before such a tool is in the hands of physicians or curious patients: expanding the concept to identify more genetic conditions, improving accuracy of measurements, and simplifying the user experience.
So, what’s next? Win or lose, the great thing about hackathons is that no idea is really better than any other. Everyone converges on a central theme, but we all bring our unique perspective and end up working on related but different problems. Any one of the multitude of teams could continue developing their idea and bring it to market; or just abandon it and use the experience to inform the next hackathon endeavor. As for my teammates, I think we’ve all expressed interest in seeing where this idea can lead. It has the potential to positively impact the lives of millions of people who might otherwise wade through years of uncertainty and misdiagnoses.
If you’re curious to learn more about our efforts, I recommend checking out the hackathon from the perspective of 2 of my teammates in their blog posts:
• Building a Successful Healthcare Hack: Rare Diseases and the Grand Hackfest
• Facial metrics: Journey from phenotypes to genotypes
Fail Fast, Fail Often
Regardless of which, if any, teams finish their projects, the crowd-based, collaborative nature of hackathons seems like the future of innovation in healthcare. Innovation comes at the price of repeated failure, and hackathons are the embodiment of the “fail fast, fail often” credo. A group of people with unique experiences, skill sets, and perspectives are free to create imaginative, innovative, and disruptive technologies that can enhance, complement, or outright replace the best we have now. No matter how many ideas fail to start, the odds are in favor of some great solutions coming out of the fray.
- 03 April 2014 at 4:04pm
- How Technology Will Improve the Future for Rare Disease | SirenSong
[…] and designers to take on this challenge and they made great strides in just one weekend. The facial recognition software team won top ...