How our hackathon project inspired Google to fight the opioid crisis

How our hackathon project inspired Google to fight the opioid crisisTaylor CorbettBlockedUnblockFollowFollowingMar 7For those who have never participated in a hackathon, it’s a unique experience.

There’s something about the combination of sleep deprivation, junk food, tight deadlines, and a competitive (but collaborative) spirit that gives birth to innovative ideas.

Yet, for all their popularity, a common critique of hackathons is that they have little real impact.

Once sugar highs fade, Red Bull wears off, and participants return to their daily routines, the ideas and prototypes they worked so hard to develop usually fall by the wayside.

This is a shame.

While many of the projects developed at hackathons are meager prototypes, the ideas behind those prototypes can have real merit.

With additional work, some supportive partners, and a bit of luck, these ideas can grow to have real impact.

I know because it happened to my team.

In December 2017, my team and I (three data scientists and one software engineer) attended the U.


Department of Health and Human Services (HHS) Opioid Code-a-thon.

We, along with 50 other teams, had 24 hours to comb through over 70 provided datasets to help HHS, health practitioners, and the public find new ways to address the opioid crisis.

Not coming from a public health background, it was (and still is) difficult to wrap our minds around the many interdependent actors, behaviors, and incentives that feed the opioid crisis.

In such a complex system “quick wins,” particularly those that could be addressed in 24 hours at a code-a-thon, can be elusive; but that’s not to say they aren’t there if you look hard enough.

In the weeks leading up to the event, we read as much information about the opioid crisis as we could get our hands on.

From academic journals to Reddit, we invested significant time attempting to understand the type and scale of challenges posed by the crisis.

It was through this research that we came across a startling fact: 71 percent of those who abuse prescription painkillers obtain their drugs (gifted, bought, or stolen) from family or friends.

CDC — Policy Impact — Prescription Painkiller Overdoses — 2011This fact forced us to reassess our assumptions about the most common pathways to addiction.

Not only was there danger in overzealous physicians or drug dealers, but in the millions of old and unused medications that are sitting in medicine cabinets around the country.

In isolation, these medications can seem harmless, but in the aggregate we know this excess medication is ultimately contributing to thousands of deaths every year.

This realization forced us to keep coming back to the same question: What can we do to safely remove the opioids (and other medications) that are already in circulation?Safe drug disposal is not a new concept.

For years the DEA’s Diversion Control Division has provided licenses to select pharmacies around the country that allow them to accept old or unused medication.

In addition, the DEA runs annual National Take Back Days during which the public can return medication to additional sites operated by local law enforcement.

While these programs created the infrastructure necessary for safe drug disposal, finding how, when, and where you could dispose of medication safely remained a difficult task.

Information about licensed disposal locations, hours, and limitations was dispersed across multiple federal, state, and local government websites as well as the websites of private companies like CVS and Walgreens.

On top of this, getting the public to engage in preventative behaviors, such as returning unused medications, is notoriously difficult.

It’s not that we’re bad people, we’re just busy, tired, and overwhelmed humans with jobs, bills, and families.

While many of us aspire to do the “right” thing, my team realized there wasn’t a resource that enabled the public to easily find and utilize the existing take-back locations around them.

To truly motivate the public to return their unused and excess medication, we realized we needed to build a tool that accomplished two things: 1) built a sense of urgency and 2) made it as easy as possible for the public to find take-back locations near them.

Armed with these facts, a lot of data, and copious amounts of coffee we started the code-a-thon.

24 hours later we emerged with our prototype: TakeBack AmericaTakeBack America Prototype — April 2018In short, TakeBack was designed to eliminate barriers to safe drug disposal by making it easier for members of the public to find licensed disposal locations near them.

TakeBack utilizes county level opioid risk and prescribing data (provided by HHS) with drug take-back location data we had scraped from various websites.

Within TakeBack, this data is merged and displayed at the county level using a combined heat and point map built in a dynamic and easy to use interface.

Ultimately, we believed that presenting this information in a localized, concise, and compelling way would motivate the public to return their old and excess medications; others agreed.

After demoing our tool during two lightning rounds, TakeBack was selected as one of the three winning solutions by a panel of judges from HHS, Stanford Medicine X, Socrata, and others.

Team Visionist — HHS Opioid Code-a-thon Awards Ceremony (Photo Credit: HHS)Following the code-a-thon we continued to give TakeBack some TLC.

Initial updates were small; adding a few new disposal locations or tweaking our map, but soon enough others started rallying around us.

First, our leadership at Visionist encouraged us to improve Takeback’s UI, covering pizza deliveries as we worked late one evening.

Soon many of our amazing colleagues came to our aid, some staying after hours to assist with automating our data flows, fine-tuning our user experience, or perfecting our user interface.

As TakeBack’s underlying data and features continued to grow, visits to the site started to increase.

While we felt comfortable with TakeBack’s technical bits, we strongly believed that future enhancements to TakeBack should be guided by its users.

For this, we first reached out to the code-a-thon’s sponsors HHS.

Having hosted the competition, HHS’ Chief Data Officer Mona Siddiqui and her team met with us to brainstorm how we could better promote our tool.

These discussions led us to present TakeBack at Health Dataplooza 2018 and the Roundtable on Data Sharing Policies, Data-Driven Solutions, and the Opioid Crisis.

At these events, we had the opportunity to interact directly with patients, physicians, and other health practitioners.

We found these conversations to be incredibly helpful and led us to make enhancements to TakeBack that we never would have thought of on our own.

Then in September, out of the blue, we were contacted by Google.

It turns out Google (an original sponsor of the Code-a-thon) had seen our demo and felt that TakeBack had a lot of potential.

In fact, Google was so inspired they told us they wanted to replicate TakeBack’s insights within Google’s Maps and Search products.

Starting in October, we held repeated calls with Google’s legal and data science teams to review our data, methods, as well as TakeBack’s underlying code.

Last week, Google, HHS, and Visionist finally announced the release of this new tool to the broader public.

Needless to say, we are honored that both Google and HHS saw TakeBack’s unique value and we appreciated getting to work with both groups over this past year to scale TakeBack to a national audience.

We believe these new integrations with Google Search and Maps will be an invaluable resource to others and could even help save lives.

As a data scientist, there is no prouder moment than seeing your tool be made available to benefit so many.

Yet, this is just one step in the right direction.

Through this experience, we’ve learned just how complex and multifaceted the opioid crisis is.

Truly solving this crisis will require continued contributions from those both in and outside of the healthcare system.

At Visionist, my team continues to look for opportunities to apply our analytic skills to the opioid crisis and beyond.

I hope this post will encourage some of you to join us in this effort.

Whether it be a late night idea, a class assignment, or a hackathon project of your own, both we and others want to hear what ideas and prototypes you are bringing to bear against this crisis.

Our experience demonstrates that such projects can be diamonds in the rough.

If you believe in your project, continue to give it your time and attention.

Share it with others and seek their honest feedback.

You never know who it is going to inspire or what it could ultimately become.

Thank you again to everyone who made this possible including my incredible team at Visionist that helped make TakeBack into what it is today (Dillon Mulroy, Kelsey Campbell, Will Sutton, Brian Lehman, Matt Loff, Eric Reidelbach, Trish Willard, Enrique Caballero, Greg Friedman, and Ruth Goulding).

Also, thank you to Google’s Legal Team and HHS Office of the Chief Technology Officer.

.. More details

Leave a Reply