You may not think there's a clear path from studying ocean and climate science, to becoming a high school math teacher, to ultimately ending up in the diabetes device world... but don't tell Dan Goldner in Minnesota. That's actually how he made his way to One Drop, where he's now the main data science guru at the growing startup.
For Goldner, who doesn't live with diabetes himself but whose dad lives with type 2, each career step has been a link in the chain leading him to glucose meter and data platform company One Drop, where there's a lot happening these days -- from new industry collaborations to predictive automated decision support based on "a billion data points tied to patient-reported outcomes." As VP of Data Science Operations since September 2017, Goldner is leading much of that.
"I’ve had two big focuses in my career -- teaching and education, as well as understanding things through analytics," Goldner says. "If you look at my job description at any point along the way, it’s always been either one of those or a blend of both. That really is a consistent theme over the years, and it's what led me to where I am now.”
The One Drop Experience
As a refresher, One Drop is the exciting startup founded in early 2015 in New York by serial entrepreneur Jeff Dachis, co-founder and former CEO of global digital solutions company RazorFish.
One Drop's mission was to remake the glucose meter into something “cool and badass,” while creating a simple, affordable subscription service for diabetes supplies, plus a mobile management platform that would make data more meaningful for users AND allow shared learnings from all the collected data on blood sugar trends in real life by utilizing Big Data analytics.
The company has made great strides -- indeed delivering a simple affordable subscription service for diabetes services tied together by mobile app and platform with a diabetes health coaching service. To date, One Drop has:
- 29 full-time employees
- The One Drop | Chrome product now available in 30 countries
- Roughly 1M Users in 190+ countries
- Mobile app available in 10 languages
- 1,250,000,000 longitudinal biometric health data points on PWDs (as of August 2018)
- 12 peer-reviewed studies that One Drop is proud to say show a "significant drop" in A1C levels
As head of Data Science Operations, Goldner's basic job is to study all that glucose data, identify patterns and trends, and envision ways that this information can be used to improve life with diabetes. While One Drop is based in New York City, Goldner works from his home state of Minnesota where he moved back to a few years ago.
"I'm very excited to be part of this creative team, able to help come up with new ideas and ways to be helpful for people with diabetes to better manage and just enjoy life," he says. "It's been a fun environment for me and I'm grateful to be here."
Lately, One Drop has made headlines with notable partnerships, including a deal with Companion Medical on the new "smart" InPen in which a One Drop Chrome Bluetooth meter is shipped out with each new InPen product at no additional cost to the user (!). Perhaps the most exciting new development at One Drop is a soon-to-be-launched functionality that uses machine learning to predict where glucose levels are headed in the next few hours (!)
An Algorithm to Predict Glucose Levels
At the American Diabetes Association Scientific Sessions in June 2018, One Drop debuted this new feature, called predictive Automated Decision Support, that can accurately predict future blood glucose values -- with 91% of those predictions falling within +/- 50 mg/dL of actual meter readings and 75% within +/- 27 mg/dL. That's pretty darn accurate!
The company's materials explain: "Blood glucose predictions come from One Drop's machine learning models, which are powered by over 1.1 billion data points collected by more than 860,000 One Drop mobile app users worldwide. Importantly, One Drop's models do not require getting to know an individual over time. Unlike other predictive tools, One Drop's models provide accurate predictions for one person based on the aggregate data of all people with similar health profiles. Within minutes of entering a single data point into the One Drop app, a user can receive their first prediction."
They also note that overall accuracy improves as more data is fed into the system (i.e. Machine Learning), and the accuracy for each individual improves as he/she enters more personal health data into the app.
Their initial launch focuses on T2 PWDs who are not using insulin, as a baseline with fewer BG data points to crunch. The sophisticated analytics will not only predict upcoming trends, but also provide "insights and recommendations" to the users. They definitely plan later support for PWDs on insulin, once the system proves itself and gains traction, we're told.
Goldner is at the helm of this data-driven effort, and most of his attention is currently on prepping for the T2 launch by the end of September, he tells us.
Diabetes and the Human Mind
Goldner says he approaches his role with a "spectrum of engagement" in mind for diabetes management. That is, people vary in how they handle diabetes and the technology and data tools available – from those who check BG and CGM data constantly on smartwatches or apps, to those who are savvy enough to build their own systems, all the way to many PWDs who rarely check with fingersticks and aren’t as interested in keeping track of their D-management -- and often aren’t even sure what the numbers they see really mean.
That’s where Automated Decision Support comes in and can help. “Anything that we can do to help augment the human brain...” he says.
For those who are less engaged, Goldner sees the new predictive function as a way to possibly entice someone to be more curious about what’s going on. "Maybe if they see a forecast about a number, they’ll check another time to affect the next few hours. Eventually, their BG log won’t be a retrospective scorecard but a proactive monitor that motivates behavior change."
“Taking a metabolistic system that’s opaque and hard to understand and making it more visible in a timely way, I’m hoping will make it more rewarding for people to think about lifestyle and diabetes changes. We’re not trying to make people engage in ways they don’t want to. But when they do, I want the tools to be a source of relief not one of confusion or discouragement. The more that can happen, the easier it’ll be for people to stay in tune in ways that aren’t overwhelming.”
In other words, the better and easier the tools, the more likely they are to shape the course of people's diabetes management.
We love that as an obvious Data Brain and Tech Nerd, Goldner says emphatically: “What I try to do is to make everything easier... Diabetes is hard, and it doesn’t have to be.”
He says he has a bunch of ideas about how this ADS is going to help people, but like everything he's done in his life, it's a bit of an experiment and he's looking forward to seeing what happens post-launch.
"The most exciting thing for me is going to be seeing how this plays out in the field. I want to see what’s really happening and if I’m right, and how we can make it even better going forward."
In some ways, he sees this as the apex of his unconventional path to One Drop -- landing in a place where he impact so many lives.
From Ocean Science to "Data Gaps" Expert
Growing up in the Minneapolis area of Minnesota, Goldner says he's always loved the water. He has fond memories as a kid standing in rivers and streams for hours, fishing and later scuba diving and having the same passion for the ocean. He also loved math, and all of that came together in merging his passions for math and the ocean when he reached college -- at Harvard, nonetheless. Goldner eventually earned a Ph.D. in Ocean Physics at MIT, basically becoming an authority on everything from water currents and climate change, to waves and how water moves around this planet linked to the solar system.
“The oceans are very big and ships are very small, so even though there are terabytes of data about the ocean, there’s not nearly enough to measure everything that’s going on in the ocean," he muses. "It’s even much more difficult than measuring what’s going on in the atmosphere. And so, what I studied was ways to get the most knowledge you can out of the data you have, even if you need more and there are errors or gaps that aren’t measured."
Now in his current role studying D-Data every day, the parallels between ocean science and diabetes data aren't lost on Goldner.
“It’s a combination of looking at the data statistically and what patterns emerge from that, but also using what you know," he says. "In the ocean context, it’s using these measurements in the data but also filling in the blanks with what you know on the physics of the water. And likewise in diabetes, you have pure machine learning models looking at data, but also what we know about how the pancreas functions and insulin works and how all of the factors of life come into play to affect the data. We can combine what we know about diabetes with the data measurements we have, to get the very best picture of what’s going on."
He took a brief detour after his undergraduate years to teach math at a private high school, embracing that long-time love for mathematics. But then he returned to MIT/Woods Hole Oceanographic Institute to complete his Ph.D. And after that, he went in what some might consider a totally different direction – consulting on business development through data analysis.
In his independent consulting role, Goldner said he had a hand in various industries and Fortune 50 companies, operations at manufacturing factories, to working on the FAA’s management of the aviation system, and NASA’s project on building a new spacecraft. He also did work in the Pharma industry, aimed at analyzing the cost-effectiveness of marketing pills via various commercials, to working with payers to get on formularies, and maximizing profits in different ways.
To him, it was the same type of job he'd had when studying the ocean -- looking at a lot of different data sets, recognizing the gaps, and being able to fill those in to guide decision-making, operations and business development.
Goldner says he loved it and learned a lot, but after a decade or so he missed working with kids in the classroom. That sparked the next chapter in his career path.
Consulting to Classrooms, and Back Again
He’d been hearing the national narrative about how inner-city public schools and how they were supposedly failing, but his analytical mind needed to see first-hand data to prove it. So he went back to teaching, spending a year getting his public school credential before taking a high school math teacher position at “turnaround school” in Boston.
“We managed to bring that school to the point where it was the first school in Massachusetts to get out of turnaround status and get back on its feet,” he says. “That was a very exciting chapter and I learned a lot. Remember, it’s a blend of what the numbers say (about school success or failure) and what you know about how a system works.”
From there, Goldner went back into general business consulting and that’s how he hooked up with One Drop in the diabetes space. As it turns out, he and One Drop founder Jeff Dachis had mutual acquaintances, as they both grew up in the Minneapolis area and were in the Boston area at the time when their paths crossed. In early 2017, Dachis just happened to be looking to find a data analytics expert to expand the diabetes company.
And the rest is (One Drop) history.
New Learnings from a Type 2 Dad
So is this where Goldner would have imagined he'd end up? Pretty surely not... but it's had some major advantages.
He's gotten to fulfill his dream of being both a scientist and professor; he loves his daily work at One Drop and says he still occasionally teaches analytics at the Carlson School of Management at University of Minnesota.
And perhaps even better, his current work helps his own father, who lives with type 2 diabetes. They now regularly talk diabetes and Goldner says he has so much more appreciation for what his dad (and mom) live through each day. His dad has of course become a One Drop user himself and is quite happy with the product and subscription service.
"I’m extrapolating what I see from him and realizing how much willpower and brain-power people with diabetes bring to that situation – and how we can help at One Drop," he says. "I’ve been learning a lot about the realities of living with diabetes, and how people can figure out what they need to do. It's pretty amazing. That's the best part of all this, seeing what's behind the data and being able to get that back to people to best help them."
Nicely put, in the surprisingly down-to-earth words of a big data science brain!