Interview with David Hale, Project Manager of Pill Identification System Pillbox

Posted by: Amy Hooker, Director of Online Marketing Apr 29, 2010 0 Comments

The patient arrives unconscious in the emergency room as physicians search for clues to the cause for her condition. A quick look in her purse reveals a single light orange oval tablet with ML-P17 imprinted on one side—is it related to her state or unimportant?

Thanks to a new application created by Gov 2.0 superstar David Hale and his team at the National Library of Medicine (NLM), the ER doc can access Pillbox and within moments determine that the pill in question, Doxazosin, is used to treat hypertension and cholesterol. Mystery solved, he can move forward in her treatment.

Earlier this week, we sat down with David and talked to him about how Pillbox came about, how he and his team overcame challenges in its development, and what’s next for him at the NLM.

RCSM: So where did the idea for Pillbox come from? How did you get involved?
DH: Several years ago, my division at the NLM took part in discussions with other organizations involved with medication safety, including the National Capital Poison Control Center and the Department of Veterans Affairs. What came out of those conversations was the common need for a system that would make it easier to identify solid dosage medications. The NLM’s Director, Dr. Donald Lindberg, initiated the creation of a pill image database. At that time, I was exploring connections between histology images and toxicology data, so this project was a good fit.

We knew we wanted to take the concept of pill images and build an online identification resource—but things really started coming together when we moved outside the government and learned how pharmaceutical data is integrated into the lives of individuals and communities. I met professor Michael Wesch from Kansas State University, who introduced me to the concept of digital ethnography and impressed upon me the importance of spending time with your users, learning how they utilize technology to ascribe meaning in their lives. We went from focusing on “features and functionality” to “meaning and value.”

I took that advice to heart and spent time with clinicians, patients, pharmacists, nurses, emergency medical staff, and others learning about the problems they were facing and how they were going about creating solutions. At that time, I was doing a lot of Native American health information outreach, and so I’d take time to go to an ER at a reservation hospital or a local fire station and talk with the staff about what they were doing, what problems they faced when it came to identifying drugs, and what types of tools would be helpful for them.

We also started talking to software developers who focused on health IT and mobile health, to the pharmaceutical industry, and to participants at health and government unconferences, such as Healthcamp. Our question was always the same—will you help us build this?

As a result of these conversations, what we found was that while the need for pill images was substantial, there was a greater problem facing citizens: The existing FDA Structured Product Label (drug labels) dataset wasn’t easily accessible and wasn’t in a form that could be used by those who needed the information. So while we originally set out to create a database of pill images, an engaged community taught us that we need to make federal pharmaceutical data more accessible and a tool they can use to create. We initiated conversations with members of the FDA’s SPL team to share this challenge with them.

It sounds like an engaged community has been key to Pillbox’s early success.
Yes. It’s critical to work with communities, prior to the release of data, in specifying its structure and developing systems that enable access in a form that creates value for those communities. This concept of engaging communities and promoting participation in the processes leading up to the release of data is critical to promoting utilization of that data, once released.

Pillbox as a concept is one that is certainly community-built and community-owned. Individuals from all sides of this complex issue have been involved in the building process, and I think that involvement explains the overwhelming response the project has received even though it’s still in beta.

What kind of responses are you seeing?
Well, probably the strongest response was health IT developers insisting that Pillbox have an API (Application Programming Interface) that would allow programmers to create applications that use Pillbox’s data. This is similar to the way Twitter’s API has enabled countless Twitter applications to be built. And by “insisting,” I mean that a couple of developers literally cornered me at every conference I went to, trying to convince me of the value of an API. When I finally sat down, listened to them, and they convinced me of the innovation that an API for Pillbox would promote, I was sold. Then I broke the news to them that I didn’t know how to create an API. But they did—they actually helped to spec and build the API that’s in beta right now.

The applications that are being developed even while the API and Pillbox are in beta are incredible and far exceed what we would have been able to do on our own. For example, one student at George Washington University spent two weeks with the API and created “Pillbox by Voice,” an application that lets you call a phone number and verbally share information about the pill in question—for example, “round” “white” “one score“—at which point it identifies the pill and reads you the ingredients, dosage, and other information. Incredible! There is also a group of developers creating a Facebook game that includes pill identification as part of the game, and others developing medication identification and reminder apps for the iPhone—and remember, this is all happening while we’re still in beta! I’m excited to see what will happen when we begin to move out of beta in the next 60 to 90 days.

I’m certain that a project the size and scope of Pillbox didn’t come together without its share of challenges. What kind of roadblocks did you encounter and how did you overcome them?
Honestly, Pillbox has been somewhat stealth since the very beginning because we’ve gone about the project in a way that didn’t necessarily follow the traditional work flow. As I mentioned earlier, after reaching out to the community and beginning that conversation, we realized that the FDA’s SPL data was where we needed to work. So to overcome that challenge, we knew we needed to once again get people with a variety of types of expertise involved. As you may imagine, the SPL is an incredibly complex regulatory dataset. We assembled a team of nine experts, including regulatory specialists, pharmacists, computer scientists, and database engineers to work through the FDA data and restructure it. A key component of the foundation of Pillbox is this NLM/FDA partnership.

Another challenge that we were able to overcome was to ensure that the pharmaceutical industry was on board with what we were doing. To do this, we needed to address their concerns about issues like anti-counterfeiting measures, intellectual property, and how their products were being represented. Again, this was achieved through conversation and asking the question, “Will you help us build this?” We also spent quite a bit of time talking to innovator and generic firms through collaborations with the FDA and non-government organizations such as HL7 (Health Level 7 International) and the US Pharmacopeia.

We’re connecting government agencies, industry, nonprofits, clinicians, and patients—creating a conversation. Pillbox will continue to evolve as these conversations and communities evolve.

Would you call yourself a Gov 2.0 evangelist?
Yes, I am frequently called an evangelist. [Laughs] I love the place that I’m at, the intersection of Web 2.0, Gov 2.0, and Health 2.0. I’ve taken a lot of my cues from the start-up community. Friends and colleagues in the San Francisco Bay Area and Silicon Valley joke that I’m running a startup within the government. Ironically, I’m going to be expanding on that idea during a presentation at O’Reilly’s Gov 2.0 Expo next month.

What’s next for you at the NLM?
I need to stay with my “startup” a little longer; however, I’m exploring the creation of a system that will enable communities take the lead in selecting government datasets that have potential to solve problems they are facing, work with government agencies and developers to restructure that data in a way which adds value and enables the creation of new applications, while keeping ownership within those communities.

Sounds great! Best of luck with Pillbox and whatever other datasets you decide to bring your community-driven approach to, David. Thanks for your time today!
Thank you.

Amy Hooker

Director of Online Marketing

• Most Recent
• Topics

View all previous posts