Good very early morning from gate 1 at Lafayette Regional Airport, waiting for my flight home after a great visit to the second annual Cajun Codefest. As I gushed last year, this is a truly special place with a combination of awesomeness that I’ve never experienced anywhere else. If you can gauge enthusiasm by the amount of late night time spent looking at local real estate on the web --- I like Lafayette a lot.
The HealthVault team was present in force: Phil and Saeed did the hard work, staying up late into the night helping teams integrate HealthVault into their solutions. I got the easy job and just had to try to keep people awake during my presentations and help with judging.
This year, the Codefest theme was “Own Your Own Health.” A pretty wide open idea, and I was frankly a bit worried that without a bit more focus and only 27 hours to work, folks would get too scattered to deliver real results --- but while that happened to one or two teams, it certainly wasn’t the norm. Three standouts for me:
The $25,000 winner, “Team Break Fix”
Inspired by a pretty overwhelming personal story of adverse medication events, this team set out to attack the problem of tracking a deceptively simple question, “are you taking your meds?”
I take one pill every day, and even I struggle to remember if I took it after a morning on autopilot. My wife Lara has a complicated regimen and spends a pretty nutty amount of time every week filling a special pill box to try to keep track. The stats are pretty clear on this one, and the opportunity to make things better is pretty overwhelming.
That said, it’s also a problem many are trying to address, including everything you could imagine from super-gadgets like bio-powered pills to old fashioned phone calls for folks that haven’t even filled their scripts. So what new were we going to see here?
Make it easy to track the event of taking a pill by attaching a cheap (often less than a dollar even at low retail volumes on Amazon) “near field communication” sensor on the pill bottle. Bump that up against your phone as you take the pill and forget it. Nice! There are some real issues here with multiple pills, and NFC isn’t nearly ubiquitous yet --- but the idea is solid.
Involve the care circle --- both professional and personal -- via text message when doses are missed. Again, there’s workflow complexity here to avoid classic alert overload, but the solution was surprisingly advanced given they’d only had a day to focus on it!
The “most market-ready” $5,000 winner, “New Old Schoolers 2.0”
There’s a reason that medicine is so often the setting for television shows --- beyond the obvious emotional draw, healthcare almost always comes down to detective work: what’s the trigger for this symptom or condition? To wit: Is Jan really allergic to Tiger, or (spoiler alert) is it just the new shampoo they’re using?
Less dramatic but just as important is the ability to spot trends before they become a big deal. Relatively minor weight gain in a person recovering from one heart attack can easily add enough stress on the muscle to cause another. Small increases in rescue inhaler use can telegraph an upcoming serious attack. Rising anxiety can cause acute psychological incidents. And so on.
The Schoolers took on this problem by creating a mobile solution built to find correlations and trends in these “observations” over time. First, the app helps collect data conveniently and quickly, using a single screen composed of green-to-red sliders that capture “current state” of anything: pain, stress, nausea, you name it. It also records point-in-time events like an asthma attack or unanticipated need for insulin. And while they didn’t implement it in their 27 hours, the solution plan includes tracking other external factors like local pollen count or air quality.
The “all in one” nature of asking the user for data is pretty key. There are some things we can measure automatically, and that’s great. But many of the observations we need, at least with today’s technology, require asking and answering. And that’s a pain to do every day over a bunch of questions. Having everything in one screen with simple, quick sliders --- that remember your answers from yesterday --- go a long way towards easing the burden of keeping up.
Next, the team recognized that identifying patterns in a mass of variables is really tough even when you DO have the data. So they created color-based visualizations to try to help --- e.g., if the pain measure is “reddest” every Friday --- it’s a hint that may help you find a trigger you didn’t realize existed.
And finally, they realized that the uses for this tool can extend beyond individual care into public health and other research realms. In Louisiana near to where we were coding, they burn off sugar cane fields every season. Wouldn’t it be useful to have people across the community report back on coughing, asthma attacks and other air-quality related symptoms during this period? By correlating these with location and time we might be able to mitigate community-wide issues we doing really even know are happening.
The student $2,500 winner, Geaux KNES
On its face this was just a really well-executed iPhone-based exercise companion that helps suggest exercise programs given your personal situation, and then uses GPS to track walks and runs. But beyond the fact that these two (yes, just two) students built a metric TON of code and great design in the 27 hours --- their app has some really powerful new ideas.
First, they didn’t track only distance and speed --- but build their system around exercise intensity. By combining speed with grade (elevation gain/loss), they computed how hard the user is working, which is a key factor in understanding the benefit you’re getting from a given session. I’ve seen simple elevation tracking in quite a few apps, but not this use of intensity as a combined metric.
Next, they took those intensity numbers and translated them into language that match the guidelines we all know about: twenty minutes of vigorous exercise three times a week or whatever. But what is “vigorous” to me may not jibe with what those guidelines mean --- it’s a pretty subjective word. The Geaux KNES app does this translation for me, so I can keep myself in “moderate” or “vigorous” range in real time. Nice!
HealthVault as an enabler for “suites of innovation”
In parallel with the coding, we got to hear and interact with super-smart folks like Farzad Mostashari and Brian Sivak. This was my first time meeting Brian in person, and it was a lot of fun --- he’s a great guy. I will forever regret that I was too pooped out after the long days to visit Festival International with him and the gang this year … but at least Saeed made it last night, and I guess that just gives me one more reason to come back next time.
Farzad made a point that was incredibly relevant to HealthVault. There is a huge amount of innovation going on, but if you want to see that innovation really make a difference at scale and in the real world … you need to make it easy to adopt. Docs are being bombarded by hundreds of niche “apps” that may be awesome, but can’t possibly all be integrated one-by-one into clinical environments and workflows. We need to find a way for these apps to fit into a “suite of innovation” that reduces the overhead of adoption.
Um, this is why HealthVault exists!
We work all day, every day, to help make it easy for apps to integrate with the healthcare system. We do the hard work of moving data between enterprises and their patients or members bi-directionally:
OUT of clinical systems, pharmacies, labs, Blue Button, HIEs, monitoring devices, etc. etc. INTO a platform where your apps can use it. So your app can be auto-populated with data from the providers without ever even knowing who those providers are.
INTO clinical systems OUT of your app where providers and care managers can act on it. For example, the Schooler’s app generates a ton of data and tries to help the individual find the patterns. Often a provider is going to way better at identify the patterns. One option to get that data to the provider would be a new “app” for the doctor to look at, but this is a non-starter --- docs have enough screens to log into and look at --- you have to be integrated into their workflows. HealthVault will help you get that data back into an EMR where it can be most valuable.
A very telling moment for me during judging was when Brian asked the Geaux KNES team --- who had implemented a slick two-way sync of exercise data with HealthVault --- if they’d considered connecting to devices like the Fitbit. The poor kid started explaining how they could do that because Fitbit has an API, etc. etc. --- and I couldn’t stop myself from shouting out, YOU ALREADY CONNECTED TO FITBIT AND 200 OTHER DEVICES --- because Fitbit is already part of the HealthVault ecosystem. Any user with Fitbit data will AUTOMATICALLY see it in the Geaux KNES app, and they didn’t even realize it!
Sheesh. This was funny in one sense, but really just a bummer, and I wish I could figure out how to make people “get” what it means to connect to a true platform like HealthVault versus a million one-off APIs. Ah well, we’ll just keep swimming and delivering the message again and again until it sticks.
The other thing we learned about HealthVault --- was that our current Java SDK has fallen behind the curve of development and needs work. Not sure how we’ll address this … whether we should keep on the strategy we have there or just reset with a newer-style REST API that’s easier to consume without an SDK. I got some strong opinions on this during the sessions. ;)
This post is already way too long, so I’ll refrain from waxing lyrical about the amazing Crawfish boil, the incredible music of Terry Huval and his sons, breakfast with boudin and cracklings, and all the other great stuff that makes me wait all year to come back to Lafayette. Suffice to say that it was an amazing time, I was energized by the enthusiasm and great work of the coders, loved catching up with new and old friends, and am humbled by the hospitality Lafayette’s leaders show to a Northwest interloper like myself. I carry my “Honorary Cajun” title with a great deal of pride!
Thanks again to everyone in Lafayette, and see you next year!