It’s 2:30 a.m. on a Saturday in Baltimore, and the room I’ve just walked into is a mess. There are soldering guns, wires, wood planks, air compressors, drills, heat guns, and two students (one barefoot) doing hardware hacking. Needless to say, I’m intrigued.
The two students are Mitch Gaines and Andrew Negal. Somewhere else, their colleagues Muskaan Khosla and Ruchee Shrestha are getting some rest, but Mitch and Andrew are continuing their work on a serious challenge — building a solution for Active Bed Sore Prevention that tracks the pressure points of a patient laying on a bed, and would inflate or deflate specific pockets in the bed to prevent bed sores from forming. It’s an important problem to solve (patients die from bed sores), and a fine thing to tackle in the early hours of the weekend.
I found myself in that room as one of the 10 mentors Elsevier sent along to the 2017 Johns Hopkins MedHacks. I’m a software and solutions architect myself, and my role involves designing software and cloud-based solutions for Elsevier’s Clinical Solutions. Other mentors were product managers, pharmacists, clinicians, software engineers, big data experts and Amazon architects.
The event itself was organized around three tracks: Medication Adherence, Access to Care, and Patient Safety and Quality. The 650 students who gathered were divided into 98 teams. As the teams self organized around their concepts, they were able to pitch their concepts to mentors to gather feedback and feasibility.
At events like these, you always get a strong sense of how sharp and innovative tech students are. The projects being presented were wildly diverse. The Simple Talk project was a speech-to-text engine to simplify medical jargon. Another was Hot Take, a geospatial heatmap that highlights quality of care in a specific area based on open data from the National Institutes of Health (NIH). As we worked with the different teams to refine their concepts, we challenged their assumptions, and asked them to really push the envelope of what could be done. That’s how we work at Elsevier, and it was remarkable to see how these hackers took up the challenge.
In fact they took it further. It dawned on us that the students weren’t constraining their concepts because of barriers that we consider practical constraints. Constraints that we consider every day such as adoptability and regulatory bodies simply weren’t a consideration. This allowed the students to shoot for the stars and truly be innovative. After realizing this, we had to adjust the way we as mentors challenged their thinking; we left constraints off the table to see where the concepts went.
Amazon Web Services donated AWS accounts for each of the teams, which allowed them to make progress quickly. We encouraged the teams to leverage Amazon services such as Alexa, DynamoDB, SNS, SES and Lambda as much as possible. The PillPal team ended up using Alexa connected to Lambda functions, DynamoDB for storage, and SNS for push notifications. For a team that had never really worked with AWS before, that was pretty amazing.
Day two got really interesting. The teams had formed and refined their concepts and were actually building things and needing guidance and help. This is where the software engineering mentors got to field the tough questions, help define architecture, circumvent firewalls rules (by using non-standard ports for services), and help to define MVPs for the concepts. As it got later, we just walked around to the different groups asking if they needed help and rolling up our sleeves to help – that’s how I found myself looking at a DIY bed sore project well past midnight.
The system Mitch, Andrew, Muskaan and Ruchee built used an Arduino microcontroller connected to pressure sensors on the mattress, and the Arduino controlled solenoids that would adjust to inflate or deflate the pockets of the mattress. Bed sores are a huge problem today, and this elegant solution was built for $74. And while this is just a prototype, it appears to be a great way to tackle targeted pressure against the body. It went on to win first place.
My personal favorite, though, was PillPop, a phone app where you take photos of your prescription labels, and the app add reminders to your Google calendar. While it sounds simple, this app was built using React Native, and used Tesseract (optical character recognition software) and natural language processing to read “take three times daily” or “take in the morning” so it knew when to add alerts to your calendar.
What’s inspiring is seeing students tackle real-world problems and seek solutions in the same way that our own tech teams do. They didn’t consider the barriers we think about every day. This allowed them to innovate and to prove a solution helped a problem, rather than trying to understand the marketability or compliance of a solution.
It’s fascinating to see students from different disciplines comes together. I personally met students focused on biomedical engineering, mathematics, software engineering, cardiology, medicine, and many other focuses. When teams with that wide of a practice area come together, they learn from each other, and they challenge each other.
That’s the kind of culture that drives innovation.