Interview: Rachel Lunsford on creating the Blue Button program and “Midwestern nice”
In this episode of Rewiring Government, Josh talks to Rachel Lunsford, the project manager of the Blue Button Initiative at the Veteran Affairs Agency. They discuss the program’s development, managers who say “yes,” user feedback, and using “Midwestern nice” to get things done.
Use the player above to listen, or subscribe on iTunes and Google Play! You can also add our RSS feed to your favorite podcast app. If you like this episode, rate and review us on iTunes, and tell your friends.
A lightly edited transcript is below.
One reason I started this podcast is that I wanted to capture the stories of how some of the technology initiatives in government were put into place, what it really took to implement them. Looking back over the last seven or eight years, I think one of the most disarmingly simple yet powerful projects anywhere in government is the VA’s Blue Button Program. So I’m thrilled to have Rachel on the program. Rachel, welcome to the show.
Thank you. Thank you for having me, and thank you for those kind words about Blue Button. As the first product manager, it’s my baby, and it’s always nice when somebody says your baby is pretty.
Haha, that’s right. Let’s start with the basics, for those who don’t know. How did you come to the VA when you did? What was your tenure like? And what is the Blue Button Program, exactly?
Sure. I was brought in [to the VA] around 2008 as a civil servant employee. I worked in the Office of Information and Technology, which was in the middle of centralizing their efforts. Previously, all of the IT staff answered at a local level to the hospital administrators, and maybe at a regional level to their regional directors. They began their centralization efforts over the early 2000s: there was kind of a shift in who people reported to, and how work was getting done. I would say [a lot of questions] were in a state of flux: How did you get requirements through? How did you get projects through? There were definitely processes in place, but it was a state of flux.
Part of my initial job was to come in and help analyze. Did projects have the right resources allocated to them? Did they have the right budgets allocated to them? That sort of stuff. There was a team of three or four of us at the headquarters’ level, that were responsible for some of that analysis.
That was literally just before President Obama got elected. It was that summer before. What I learned in being in the federal government at the time is that transition period is always interesting, because you have to have continuity, right? That’s kind of why the bureaucracy exists: [to] have continuity between the administrations. You also get this distinct sense that you’re holding your breath for something to happen, especially with the intensity of the 2008 election.
What President Obama brought is just this idea of, “We’re changing things. Things aren’t going to be the same old, same old anymore. We’re definitely going to try and do the right thing by citizens.” I distinctly remember, now looking back at it, this electricity in the air: “We know something is coming. We’re not sure what’s coming yet, but we know it’s coming. We’re going to have to think about how to improve what we’re doing [to achieve] what we always talked about, delivering world class service to veterans at the VA.” That’s kind of the beginning of all of it.
Then, in early 2010, two years into the Obama administration, there was a group of folks who got together and said, “What can we do to make healthcare better and easier for people?” More specifically, [how can we work to improve] some of the services that government offers? What can Social Security do? What can the VA do? What can all these services do that fit with the idea of open, transparent, collaborative, participatory government?
I wasn’t there, but I always love this story of this meeting at the Markle Foundation in New York. Being kind of the government dork that I am, I would say the kind of innovative leaders of the time were all there. Todd Park, Aneesh Chopra, and Peter Levin, a bunch of folks like that who were all there saying, “We could come up with a billion ideas.” The great folklore from this meeting is, no one really knows who said it, [but] it seems like everyone thought at the same time: “Why don’t we put a button on the VA’s website? Why don’t we put a button on the Centers for Medicare and Medicaid Services website, and the Social Security website, that people can literally push to download a simple version of their health and benefits information?”
I love the notion that everybody was thinking it and, at least from what I’ve heard, no one really likes to take credit for the idea. I think it’s one of those things that are just, as you said, so simple and powerful that it almost doesn’t matter who came up with it. Because after that, we were hitting the ground running, trying to figure out what that really meant and what it would look like.
I think this part of the story is genius from a strategic perspective: you have this electricity in the air, and there’s a sense that technologists are coming into the administration from the digital world for the first time. They could sort of do anything, right? And there wasn’t really a concept of what that would mean in practice.
The idea of releasing to people what is essentially a text file, letting them access their own information and make the connections to get it from their doctor to their preschool to their pharmacy, and to keep it that simple, took a tremendous amount of strategic foresight. Everyone knew that changing government was not going to be a straightforward process, and to start with such a high-level shift is, I think, one of the great strategic moves that’s happened in the space.
What did it look like in practice to go from that meeting at the Markle Foundation to you, where you had to take the ball and run with it?
We had lots of challenges, but the VA is used to having challenges. I think for the forseeable future, there will always be challenges associated with providing care for veterans. Especially since a different type of care is needed, because of the sorts of things that we send our soldiers out into.
We were very lucky that we had, I think, a leadership team that saw those challenges as opportunities. The Assistant Secretary for I&T was Roger Baker. You had my boss, Peter Levin, who was the CTO. The two of them both were like, “Whatever we need to do, let’s bring those challenges to the table, and let’s make sure we figure out how to get rid of those roadblocks.” That mentality went to up to Secretary Shinseki and his leadership team, the deputy secretary, everybody.
You start seeing this [momentum] when you start bringing in that mentality at the top level. At that lower, grassroots level, when someone says, “I have this good idea, and I wonder if this is my chance to say whether we should do something about this?”, they think, “Well, it’d be nice if we could let people download their records, but there’s kind of this bureaucracy in front of us.” There’s somebody who says HIPAA tells us we can’t do it. There’s somebody who says that it’s unsafe to legally let people download this stuff on the internet.
We found the people who were saying “yes” and the people who were looking for the ways to get around those things. I don’t mean to [circumvent] those things legally, but figuring out how to accommodate HIPAA in a safe way, how to make sure we are complying with HIPAA but not letting it slow us down. We’re figuring out how to make sure that the lawyers are happy, without saying we can’t do something.
To that end, we began saying, “Let’s just take a small bite.” Let’s figure out whatever data is already available on My HealtheVet, and just put that in a downloadable file. Then, when we’re starting to look at [file formats], you look at all the standards available in healthcare, you could really get yourself into a mess of trying to pick one.
So we decided we’re going to go back several standards. We’re going to go back to the most basic thing we could use, and use ASCII. It’s not going to be pretty. It’s not going to be glamorous, but you can read it, and so can a computer. That was kind of our two things that we wanted to make sure we knew we could accommodate, that if somebody pulled it out and read it, they could understand it, and a computer could also read it. We found some folks on the front line, who were already thinking about ways of doing this. My job was to help connect my bosses to the folks that were in the field. We asked: What are their problems and concerns, and how do we [address them]? How do we make sure we have the resources, and the money allocated?
In the beginning, there was no real money associated with creating the button, it was just whatever time those of us put into the project. It was our salaries, more or less. It was kind of a project of love in the beginning: we knew it was the right thing, we thought it would be relatively easy, and we thought we had a pretty good proof of concept. By the time it got through the processes, it was probably mid-summer, and we actually had a concept running with this ASCII text file. A handful of people who came together to really make sure that you could read it, and that there were 81 columns and all of that in the file.
I think you just said a couple of really important things that we’ve seen over and over again, for people who are trying to improve the way government works for their constituencies. One is find the people with the right mindset, people who are going to help you do new things, and go from there. The other is do innovative things before they become institutionalized, before they necessarily have a budget, and find ways to make those things happen.
For government projects that involve releasing data, I’m concerned that it’s often seen as an end in itself, rather than as a means to actually improving, in this case, quality of care and access to patient records. To what extent did you consult veterans, the end users, to make sure that this was the right move?
We had several methods at the time, which I think they still use. One thing that was always great about being at the VA is that while most other organizations strive to meet having a certain percentage of veterans, we always had around 40 or 50% of the staff that were veterans in some way. We could holler down the hall and say, “Hey, Jerry, come look at this. Does this make sense?” We had some of those informal conversations with groups of staff, and some of them were actually QA testers from My HealtheVet anyway. They’re veterans, they use the product, and their job is to make sure it works.
At the time, our CTO was also the open government chief for the VA, and we tried to use as many tools as possible to get the word out there. We used things like Facebook when we had our first soft launch of the product, when it wasn’t quite an alpha but it was definitely headed towards a beta phase. We said, “Hey, we’re telling you about this. Anybody who is interested, give us feedback.”
It was great, because we had a lot of people who really were excited. To this day, I still talk to some of them. One gentleman from Missouri, Randy, who I talk to all the time, he’ll still message me and say, “Hey, I know you’re not still working on this, but can you help me out? It’s not working today.” That’s the sort of relationship that we built over time. We heard from folks who wanted the process to be better. They wanted access to their records in a better way and easier way. They stayed engaged.
We could actually lean on them to do some of our smoke tests. With later iterations, when we were heading toward production for the next round of functionality, we could say, “Hey, can you be around on a Sunday morning to log in and make sure things are working?” Because we had some folks who were statisticians on the team, they always used surveys as part of the My HealtheVet website, just to say, “What’s going on? Is this working for you? Do you want to improve function?”
We had a very qualitative and quantitative approach to be able to improve our functionality over time. I can’t say whether… you know, did veterans see the prototype? Folks who worked at the VA saw it. Over time, we got better and better at actually reaching out to more and more folks to say we needed their help.
For example, toward the later part of Blue Button, we were getting ready to think about how to handle images. You clearly can’t put images in an ASCII file. You probably don’t even want to put them in a PDF because you’ll have severe degradation in the image quality. We were thinking about different ways of [getting feedback.] One of the approaches was to use a pilot program. That way, you can have groups at the local level say, “Oh, this works. That doesn’t work. Don’t make that. That’s not the right shade of blue, because I can’t read the text against it.”
On the Blue Button project, at least, we tried to have as much feedback as possible to make sure we were getting what folks needed.
I want to ask you about hacking the bureaucracy. I think those of us who work in that space all have these experiences of finding those folks who share the culture of “yes” that you describe. I’m wondering if there’s some techniques for increasing the probability that people in government will have that culture. Specifically, I wanted to ask you about something you and I talked about the last time we spoke, which is “Midwestern nice.”
I know “Midwestern nice,” because my mother is from Michigan. I’m particularly curious about how “Midwestern nice” was a tool in your arsenal, or something that helped you achieve your broader goal.
I’m sure, in your experience, part of being “Midwestern nice” is that we just talk to everybody. I’ll be in a store, and my grandma will strike up a conversation with somebody as if she’s known them for 30 years, and she definitely hasn’t. I think, if you talk to some of the folks I worked with at the time, I unknowingly had that characteristic too. I was completely willing to have any and every conversation with anybody, even if it started off with shouting.
You know, [some of them] weren’t happy that I was calling them because they didn’t understand why I could possibly need to talk to them about this, that or whatever. I was just trying to learn what was going on, so that if there were barriers, we could break them down. I think it really helped to be willing to have a conversation with somebody that I have known for three minutes as if I’ve known them my whole life.
What I found out over time is, a lot of the people who are like me in other agencies and other areas all kind of come from the Midwest, or we went to college in the Midwest, and it rubs off on you or something. I think that there’s something definitely about the Midwestern personality type… the willingness to have conversations, sit down with people, and break bread together, so that you can overcome personal differences. When you’re sitting there having dinner, you can’t just get up and walk away: you really have to have a conversation about what’s going on.
During Blue Button, [that affect] really helps because, by the end, when we were troubleshooting some infrastructure and software problems, we were all in a big conference room eating barbecue as a family, because we were in Texas. We were eating barbecue family style, because we became this family.
You do have to have that person who can nurture relationships and say, “You gotta go make up with your sister now.” You’ve got to work with your IT folks. If you don’t understand what they’re saying, you’ve got to ask questions. You’ve got to work together to make sure requirements are fulfilled. You can’t just say, “Oh, well…” You really have to all work together. We’re all in this together, thinking about whether we’re actually meeting the intent of what we wanted to do, whether we’re coming together to fulfill our design and desires for the product.
What did that look like in practice as the project evolved? You mentioned starting small, working on layups until you can get to the big things. How did that manifest itself in organizational change?
I feel we were very lucky at the VA, again, because our leadership was saying, “Hey, we need to make some changes here. We have to get away from our standard way of doing business.” From an IT perspective, our standard development cycle was 10 years, where you designed requirements for like three years. By the time you actually went to develop it, it took another five years. And it was completely useless by the time you got to testing, because it definitely didn’t meet what you needed anymore.
We were fortunate that our Assistant Secretary for IT, Roger Baker, came in and said: “We are going to move towards something like an agile environment.” It’s not strictly agile for the purists out there, but we had to deliver something every six months, whether it was documentation or new functionality, something that was tangible. You could put either your hands on it, or see it working on the screen, or understand that you’d spent time refactoring, things like that.
Then the project sponsor, the CTO, said, “I think we can do better than that. I think we can do a three month release schedule. We should take a small bite [for each release]. It doesn’t have to be a lot of functionality, but I think if we can keep the iterative cadence of churning out something new, we’ll get better at doing this. We’ll understand better what we can and can’t deliver. That way, people know we’re serious about it, [and] know that we’re actually going to be able to deliver real product to people, and improve on it.”
Sometimes, government is really good at initial delivery: [people say] “Wow, that’s so shiny and great and we love it.” Then you don’t hear anything about it ever again. It just sits out there and is, like, going to die on the vine because the budget got cut or whatever. After our initial release in Fall 2010, right away we said, “What’s the next thing that we could do?”
We started figuring out which pieces of the record made sense to deliver. [Including] lab results made sense, because duplicative lab results are the worst, right? You don’t want to get stuck with a needle twice in a row, and you don’t want to get stuck with a bill twice in a row. We started thinking about the discrete pieces of data that we could start pulling from the record, one section at a time.
It was always a big negotiation, because there were a lot of people who had to provide input. We had folks who looked at it from the clinical perspective, like the doctor. If a veteran comes in and he has demographic information, his lab information and maybe [an] emergency contact, is that good information for a clinician to have? Then, we looked at it from the veteran’s perspective. Is this data appropriate for a veteran to have? We pushed ourselves to do all of this quickly, because the faster we could get things out, the greater opportunity we had to learn.
For example: “Oh, that wasn’t the right data this time. Some people didn’t really care for that. Now they’re angry and they’re telling us on Facebook.” The first couple of three-month cycles were really rough. Then the longer we went on, it became like a well-oiled machine, where we knew exactly how we all fit together. We knew exactly what problems to escalate up the chain. We knew we would not be able to remove those barriers, but that our senior leadership could.
Initially, it was a steep learning curve for us. But kudos to all of my teammates for figuring out how to make that all work over time.
It sounds like there was a great combination of that high level ability to clear the brush, plus some really hard work and good feedback loops.
Blue Button turned five years old in 2015.
I know. That’s wild.
I saw that there’s now over 500 companies pledging to make their data available in a way that is compatible with the Blue Button. At Code for America last year, Peter Levin, the CTO, said that now over 150 million people will be able to gain access to at least some of their health data with the Blue Button.
Can you talk about the role that outside companies, technology companies and others, played in coalescing behind this initiative? What did that process look like?
This is one of the parts that’s always interesting. Again, because we were in the spirit of collaboration and open government, this wasn’t just internal collaboration. We asked: who else out there can help us? One of the first things that happened for Blue Button, in the year that it launched for the VA, was that a contest was issued around how could you use data like this. It didn’t necessarily have to be functional.
We had a huge opportunity there to begin the conversation of what sort of improvements we could make. What can you do with this data? We didn’t have the Apple Watch or anything like that yet, but what other devices might be useful? If you had access to your lab data on a regular basis, and you created a parser for it, is that useful for a patient? Is that useful for a veteran to actually have?
Issuing those sorts of questions and challenges to our colleagues in the private sector really allowed us to say, “Listen, we’re turning the data fee, but we’re not telling you we have the answers on how to make this pretty.” In fact, if you look at government websites over time, they’re not always the prettiest. Some of them are gorgeous, and some of them aren’t great. We are not saying that we are the artists that can make this an amazing experience. We’re just trying to make sure that the data is available to veterans, that they can download it in an easy way, and that when they need it in an emergency situation, they could print it off and hand it, or their loved ones could print it off and hand it to the emergency room doctor.
I do remember early conversations with a couple of different partners in which they, probably not surprisingly, were a little shocked and a little mad that we picked ASCII. Because they’re like, “But there’s a lot of formats you could use, because now we really have to do some hard work in thinking about how to parse this data.” It’s like, “Well, that’s what we had to do too, right? We’re asking you take a little bit of a leap of faith with us to do some of that hard work. The rest of that leap of faith is thinking about what could be done with this data now that it’s liberated, so to speak.”
We always had that connection through the Markle Foundation, who is pretty good about bringing multiple sides to the table to think about these things. I remember that during pretty much all the big conferences after that always had a Blue Button conversation. My favorite time was when I walked into one of these Blue Button conversations to literally deliver lapel blue buttons, just to say, “Hey you guys, now you can blue button yourself. Now we’re all blue buttons.”
We brought a lot of folks to the table, just to brainstorm how to implement it elsewhere. What does [Blue Button] look like, not just in healthcare, but in other spaces? Does it make sense in the energy space? Should people have a Blue Button, which I think they wanted to call Green Button, for their electricity bills or their utility bills? Those sorts of partnerships had to go not just to government, but the private sector too, because we couldn’t bring all the answers. There’s just no way. We knew what we could do, which was free the data, and let other smart folks think about what to do with it.
Which are the initiatives that you’re most excited about, in this vein of unleashing personal data from its proprietary, often nearly useless context?
I think, for me, it’s always been enabling research at a personal level. Also, like many other people, I have a chronic disease. I have been searching for the answers for how I manage this, which medicine is the right medicine, and all that sort of stuff. It would be nice to know if I could just upload that to another site, and share that with other people. So, patients like me, and other websites like that, started to think: “Can you just upload that data instead of [manually] entering it in all the time?” Then, can you download it from those sites, so you can share it?
I’m not necessarily saying I want to be doing a lot of clinical trials for myself. It’s more of that personalized… like, “Here’s how my symptoms are. Here’s how I feel today. Here’s the tests I’ve already had done. Is anybody else like me? What are you doing to try and manage your own disease and manage your health care?” Going to eight doctors twice a year is never fun. Not that I’m going to eight [doctors] right now, but if I could get rid of some of those visits, I would just feel better mentally about the whole process.
For me, it was always very personal. This project was always personal anyway, but I think having ownership of your own personalized approach to medicine is what always excited me about Blue Button. It continues to excite me. There’s always some new device. I’m always excited about what that brings to the conversation, and how it might improve my own health and that of my friends and family.
I feel like it’s often rare in government, particularly at the federal level, that you can be a part of a project that so directly impacts people’s quality of life. What did that feel like? What would you recommend for other people who work in government, who are looking for that sort of project?
Working hard, being the best that you can, and contributing the best that you can always helps you. One of the big things is establishing that reputation of somebody who is willing to try. I think, often in bureaucracy, you always get those bureaucrats who are just “no” people. Some days it’s “no” because it’s a legitimate no, and sometimes it’s “no” because you’re in the way of Taco Tuesday or something like that, and it’s time to have tacos and not talk to you.
I think if you can establish yourself as one of those people who is willing to try, even if you’re not always the best at it, that helps. Another thing that I kind of lucked into, I think, is that I just was so darned excited about this open government thing. It was literally, I think, President Obama’s first executive order.
I was just so excited by this notion of like, “Finally, we’re going to make government be awesome. We’re going to actually do something cool.” I think when you get excited about those things, and you convince your colleagues to maybe get a little bit excited about those things, you become that person: “Yep, Rachel needs to help with that project because she’s got the passion. She’s got the knowledge.” Your colleagues recognize that you could be helping on those projects.
I actually remember… it wasn’t specific to Blue Button, but it was open government. At the time, I was not yet reporting to the CTO, but to the chief of staff for IT. She said, “Oh, we need somebody to volunteer to be on this.” I was immediately like, “Me. I’ll do it.” That’s kind of what set in motion for me being able to eventually be part of this Blue Button project that I was so enthusiastic, maybe a little crazy enthusiastic about. I want to participate. In some ways, I went out and sought my own adventure to get to the Blue Button.
Then as time went on, it just seemed like the right fit for me to continue working on Blue Button activities, just because I was so amped about it. I was at an open government meeting and it was time for me to talk. Somebody said, “Tell about the Blue Button.” I just, I guess, was so excited that I was talking so fast that they’re like, “Can you do that in half-time now?” I was so passionate and so thrilled that we were doing something that I knew was helping people out.
We’d heard from a veteran who said, “I was able to download my information, my demographics.” It had his list of allergies. He took that with him, and he had had a seizure. His mom or his wife, I can’t remember who, was able to pull that out. Reading that on Facebook and hearing that confirmation, that this was a product that really helped people, it made me feel great about this. Like, “Yeah, we need to keep going.”
That’s a really long-winded way to say: I think if you put yourself out there, and let people know what you’re about and what you’re excited about, that goes a long way towards being able to find those projects, and either get assigned to them, or selected for them, or volunteering for them.
That’s great. You’re now back in Iowa, where you’re from, and working with the Department of Public Health. What are you up to now?
I’ve taken a lot of what I learned while I was in the federal government, and my time in between the federal and state government. My job is really to help educate my colleagues on how you bring innovation into public health. Like: what sort of technologies can we procure, and what kinds of things can we build? Do we even have to build things? Do we have the tools onsite?
It’s really kind of a joy: I learned all these things while I was at the VA and felt like I was running crazy, like a chicken with my head cut off. I just didn’t know which way was up sometimes, because we were learning all together, and this was a new way of doing things. I now finally get to come up to folks here and say, “I’m not saying this is the right way, but it worked for us in the past. Here’s what you should think about. Here’s how you might want to bring innovation into the public health sphere.”
From my perspective, I started with the Blue Button, and that was on the consumer side [of healthcare]. Then, I did a small step into the provider side in a couple of different ways. Now, I’m taking one step up, looking at the population health level and saying, “What could we be doing that could really improve everybody’s health in a particular geographic area? What sorts of technologies can we use to support those initiatives?”
Excellent. Rachel, thank you so much for being on the show. Keep up the great work.
Absolutely. Thank you for having me.
Joshua Goldstein is the VP of Product at CityBase.