With only a few hours of work under their belts, the newly-formed teams at the Boston Blue Button Innovation Challenge gave their first practice pitches in front of a jumble of mentors and other hackathon participants.
True to their word, the mentors left no part of the pitches unturned, pursuing every possible problem and unearthing every flaw in the tech, business plan, and even the pitch itself.
“Bad specific problem, good specific solution,” said Dr. Daniel Karlin, assistant professor of psychiatry at Tufts Medical Center, to the concussion evaluation team after pitching their mobile eye-tracking app.
“We really want to be wowed,” said Rohan Jotwani, first-year Tufts MD/MBA student, to the concussion team. “You absolutely have to have a demo tomorrow.”
Sugarfly, a team that had pitched a mobile app for diabetics that morning, gave a noteworthy personal pitch that started with a touching personal story.
“I have four family members with diabetes,” said Creed Mangrum, a computer science student at MIT, as he began the team’s pitch. “My sister has had Type I diabetes for twenty years.”
Mangrum explained to the audience that his sister has recently had difficulty sensing when her blood sugar is low.
One time, her blood sugar went low while she was at work, and her coworkers found her unconscious at her desk hours later. Another time, she went low while driving on the highway at 80 miles per hour, and she flipped her car four times.
“Diabetics aren’t always sensitive to situations in which they will go low,” Mangrum said during the pitch, “so we want to help them live better with the condition.”
Their solution? Using machine learning from wearable accelerometers, smartphone glucometers, and Blue Button demographic information to predict hypoglycemic events.
“When the machine learning algorithm thinks you might be low,” Mangrum explained, “an app pops up on your phone and asks you, ‘How are you feeling right now?’”
Answering with high, good, or low and then inputting data from a glucometer allows the algorithm to learn about the changes in your blood sugar levels to alert family members and friends to check in on you, explained one of the Sugarfly team members.
As the pitch ended, the mentors immediately tore into the team on the machine learning component of their hack.
“I love the personal story,” said mentor Dr. Alisa Niksch, pediatric electrophysiologist at Tufts Medical Center and medical device researcher, “but you need to work on machine learning. It can’t just become a buzzword.”
“And don’t discount the Type 2 diabetes market,” she added.
Visit-to-Visit, with their Facebook-like web app that gathers patient information from Blue Button and wearable tech, gave a robust pitch that was well received by the mentors.
In their pitch, they revealed that their app would use existing APIs from wearable tech companies to pull data only on the day of the office visit, to avoid pulling too much data into the software.
“I like this a lot,” said mentor Dr. Brenton Fargnoli, clinical fellow in medicine at Harvard Medical School. “The episodic nature of the push avoids the continuous stream of too much data into the doctor’s office.”
“Doctors aren’t data scientists,” responded team member Michael Tomko. “They’re practitioners. They’re not going to read through your RunKeeper stats. But they will read a graph of how many calories you burned since your last doctor’s visit.”
After watching the last few teams, I retreated upstairs to the hack workspaces to talk to the teams about the work that needed to be done before the next day’s final pitches. I caught up first with the concussion group, regrouping in room 218 to debrief.
“We got some tremendous feedback,” said Ned McCague, “and everyone recognized the value of what we’re trying to do.”
But, the team recognized that the mentors were concerned about a demo and the team’s definition of the concussion problem.
“We need more prototyping,” said Chris Sanders, the pharmaceutical consultant for the team. “We’re definitely going to show screenshots of what the app looks like tomorrow.”
In regards to Dr. Karlin’s “bad specific problem” comment, the team had discussed at length how they would clarify their pitch for tomorrow.
“What the mentors were really getting at was to hone the pitch to really show how the problem of sports concussion will actually play out,” explained Sanders.
“We have to show how the app is used on the sideline,” added Dr. Tanzid Shams, “right in the moment.”
Matt Doiron speculated that the mentor’s concerns might also have to do with the nature of the concussion problem.
“There’s a ton of media around concussions,” explained Doiron, “but the problem is still a little more nebulous. The research is very much in its infancy, and that presents a huge opportunity for us.”
The teams continued their labors late into the night, eagerly preparing for the next day’s final judging and the prospect of winning venture capital and the opportunity to pitch their hacks to biotech CEOs,
Check back as I watch these three teams perform their final pitches as part of the Boston Blue Button Innovation Challenge.
Send this to a friend