Newer posts are loading.
You are at the newest post.
Click here to check if anything new just came in.

March 09 2012

OK, I Admit It. I have a mancrush on the new Federal CTO, Todd Park

I couldn't be more delighted by the announcement today that Todd Park has been named the new Chief Technology Officer for the United States, replacing Aneesh Chopra.

I first met Todd in 2008 at the urging of Mitch Kapor, who thought that Todd was the best exemplar in the healthcare world of my ideas about the power of data to transform business and society, and that I would find him to be a kindred spirit. And so it was. My lunch with Todd turned into a multi-hour brainstorm as we walked around the cliffs of Lands End in San Francisco. Todd was on fire with ideas about how to change healthcare, and the opportunity of the new job he'd just accepted, to become the CTO at HHS.

Subsequently, I helped Todd to organize a series of workshops and conferences at HHS to plan and execute their open data strategy. I met with Todd and told him how important it was not just to make data public and hope developers would come, but to actually do developer evangelism. I told him how various tech companies ran their developer programs, including some stories about Amazon's rollout of AWS: they had first held a small, private event to which they invited people and companies who'd been unofficially hacking on their data, told them their plans, and recruited them to build apps against the new APIs that were planned. Then, when they made their public announcement, they had cool apps to show, not just good intentions.

Todd immediately grasped the blueprint, and executed with astonishing speed. Before long, he held a workshop for an invited group of developers, entrepreneurs and health data wonks to map out useful data that could be liberated, and useful applications that could be built with it. Six months later, he held a public conference to showcase the 40-odd applications that had been developed. Now in its third year, the event has grown into what Todd calls the Health Datapalooza. As noted on GigaOm, the event has already led to several venture backed startup. (Applications are open for startups to be showcased at this year's event, June 5-6 in Washington D.C.)

Since I introduced him to Eric Ries, author of The Lean Startup, Todd has been introducing the methodology to Washington, insisting on programs that can show real results (learning and pivots) in only 90 days. He just knows how to make stuff happen.

Todd is also an incredibly inspiring speaker. At my various Gov 2.0 events, he routinely got a standing ovation. His enthusiasm, insight, and optimism are infectious.

Todd Park

When Todd Park talks, I listen. (Photo by James Duncan Davidson from the 2010 Gov 2.0 Summit.

Many will ask about Todd's technical credentials. After all, he is trained as a healthcare economist, not an engineer or scientist. There are three good answers:

1. Economists are playing an incredibly important role at today's technology companies, as extracting meaning and monetization from massive amounts of data become one of the key levers of success and competitive advantage. (Think Hal Varian at Google, working to optimize the ad auction.) Healthcare in particular is one of those areas where science, human factors, and economics are on a collision course, but virtually every sector of our nation is undergoing a transformation as a result of intelligence derived from data analysis. That's why I put Todd on my list for of the world's most important data scientists.

2. Todd is an enormously successful technology entrepreneur, with two brilliant companies - Athenahealth and Castlight Health - under his belt. In each case, he was able to succeed by understanding the power of data to transform an industry.

3. He's an amazing learner. In a 1998 interview describing the founding of Athena Health, he described his leadership philosophy: "Put enough of an idea together to inspire a team of really good people to jump with you into a general zone like medical practices. Then, just learn as much as you possibly can and what you really can do to be helpful and then act against that opportunity. No question."

Todd is one of the most remarkable people I've ever met, in a career filled with remarkable people. As Alex Howard notes, he should be an inspiration for more "retired" tech entrepreneurs to go into government. This is a guy who could do literally anything he put his mind to, and he's taking up the challenge of making our government smarter about technology. I want to put out a request to all my friends in the technology world: if Todd calls you and asks you for help, please take the call, and do whatever he asks.

June 10 2011

Challenges aired at Health Data Initiative Forum

A major bash by the Department of Health and Human Services and the Institute of Medicine--together with the NIH, EPA, and others--drew hundreds of people yesterday in Washington, DC to discuss the use of government data in health care. By placing "challenges" in the title to this article, I'm indulging in a play on words. Most readers will come expecting me to talk about problems that require grappling and resolution. But the "challenges" I'm talking about are contests held by numerous institutions to produce applications that consume health care data and, ultimately, benefits for patients, doctors, policy makers, and the public at large.

The challenges to which so many developers have responded draw on data that HHS has loaded in a programmable format onto Health.Data.Gov--and that they continue to release at a fast clip (see for instance the Health Indicators Warehouse). The zeal with with HHS has organized its data under convenient APIs puts it among the foremost of agencies in the open government movement.

Todd Park, HHS CTO and hero to everything in the open government movement, described the "datapalooza" as follows: "Today is just a day of massive celebration." And there was certainly a lot of mutual back-slapping among Obama Administration leaders on the stage. But the strength of the new applications shown off at the conference proved there was indeed reason to celebrate. The release of data and its use by developers ranging from college kids to international corporations is a non-partisan cause for exultation.

By bringing us to a new plateau, these apps also show some of the next steps we need to take in health care to extract real improvements in care and costs savings from the data on which the apps are based. So there are actually "challenges" of the problematic type implicit in the conference after all. I'll start by describing the progress shown by some of the apps I saw; a briefer but more comprehensive list can be found on the blog of my colleague, Brian Ahier. Then I'll move on to next steps.

A sample of recent health-care apps and challenges

In the interest of full disclosure, I should mention that I submitted some ratings of apps as a judge in the Health 2.0 Developer Challenge. Some of the interesting projects I discovered at the conference include:


This won the top prize in a special contest run by the Institute of Medicine and the National Academy of Engineering for college students, Go Viral to Improve Health. The app was developed by two students, one of whom said she never gets enough sleep because she goes to school full-time and works two jobs. How, one might ask, did she have time to write a winning app? I hope that after winning this prize she can quit one job.

The app makes it extremely easy to measure sleep and compare different nights. You just press a button that says "I'm going to sleep!" and the app records the current time. If you press your button in the morning to turn off your alarm on the mobile device, the app records you as waking up. You can display your sleep times for multiple nights and save the data for running statistics. I think this app would make an excellent complement for the Zeo sleep monitor, which I've covered in other blogs.

Organized Wisdom

This site connects doctors and patients, somewhat like HealthTap, which I covered two months ago. The data aspect of Organized Wisdom that I find interesting is their display of common ailments--all the usuals you'd expect, such as asthma and diabetes--associated with each geographic area you search for. This data comes straight from You can also click on each ailment to pull up an NIH page about it. This data could be used by doctors to intervene more effectively in the locations they work in, or by patients to choose where to live or prepare for the risks where they do live.

I should note that the conference also highlighted a journalist's site called


This patient portal, run by the New York-Presbyterian hospital, illustrates the evolution of current clinical institutions toward patient-centered care, such a focus of health activists. By exposing patient data on a portal, a hospital can turn its EHR (electronic health record) to a PHR (personal health record). MyNYP also resembles many hospitals in exposing their data in the lithe CCR format and providing an interface to Microsoft HealthVault, so that patients can mingle their hospital data with other data they accumulate. A web developer from MyNYP told me they provide a limited way for a patient also to upload data. I expect this feature will expand as patients demand ways to provide their doctors with observations of daily living.

The interface to MyNYP is clean and readable, but one aspect shows the complexity of demands for usable interfaces. Each hospital visit is a separate item, providing its own button for you to upload its data to HealthVault. For the neediest patients--those likely to care most about their data--this could mean having to click buttons dozens of times to upload a few months' worth of data. A simple list of checkboxes on the side and an "All" button, such as Gmail provides, could let patients select data more easily.

Lumeris Maestro

This is heavy-duty software for helping hospitals and clinics meet the requirements of "accountable delivery," which could be the formation of an Accountable Care Organization (one of the hot trends in current movements to lower health care costs) or some other type of integrated care. For instance, it can help you check areas of variability (are one set of doctors treating a condition substantially differently from another set?). The software helps organizations integrate clinical, pharmacy, and other data. It is designed to let organizations keep their current EHRs and run the accountability functions on top of them.

I held a special regard for this project because the presentation went very fast and raced through a number of buzzwords. I trust that experts in the field appreciated the importance of the goals and design, but the general public could easily have bypassed the project.

Feedback Disease Outbreak Investigation System

If you can figure out from the name what this project does, you don't need to read this blog. The project accumulates data about outbreaks of food-related illness and data that traces the origin of food delivered to various stores and restaurants. It provides elegant and animated visualizations that help an epidemiologist quickly track down the origin of food poisoning, hopefully in time to prevent further victims.

As with Lumeris Maestro, the presentation of this project at the conference was rather pro forma. The presenters didn't even bother to update their prepared script to mention the recent spectacular and widely reported outbreak of E coli deaths in Europe. But the demo persuasively conveyed the value of this software under crisis conditions.


This site also excited a special sympathy in me because it's heavily based on taxonomies. When searching for in-depth data about a heart attack, it's important to know that clinicians refer to it as a myocardial infarction. Healthline correlates all these different forms of terminology (along with diagnostic codes) and offers rich information to the general public, including beautiful anatomical diagrams.

From the app to the ecosystem

The notion of an "ecosystem" for health care apps was invoked at least twice during yesterday's conference, by Park and by Risa Lavizzo-Mourey, President of the Robert Wood Johnson Foundation. I believe they were conceiving this ecosystem as a business environment that made software development lucrative enough to attract more apps. Kathleen Sebelius, HHS secretary, indicated in her opening remarks that the contestents for this year's challenges had to demonstrate sustainable business models as well as good applications. And an initiative called StartUpHealth was announced by Organized Wisdom, planning to provide funding and mentoring over a series of years to promising health-related enterpreneurs. All well and good. But I will use "ecosystem" in this section to denote a rather different set of supports and usage.

The first part of the ecosystem is getting accurate and useful data for the apps to play with. Opening government databases is critically important, but improving them is the next step. Much of the practice required by the government under "meaningful use" criteria is directed at making the data submitted by doctors is complete, consistent, and truly reflective of what they do in the clinic. Just one common example of dirty data is provided by diagnostic codes, which researchers complain are usually chosen for billing for the highest possible payment rather than indicating what's really wrong with the patient.

In short, we may have to do the best we can for years to come with incomplete or misleading data. And we need to create better standards (as I described in my blog from the Massachusetts Governor's Health Care Conference as well as incorporate them into electronic products clinicians use every day.

App developers should also coordinate with each other to use data formats in common. Letting a thousand flowers bloom should entail letting a thousand XML schemas proliferate. As we've seen in this article, itself offers some degree of standardization, and so do industry formats such as the CCR.

More standards will need to be introduced so that consumers can easily plug machines such as blood pressure monitors into commodity cell phones, and let the apps recognize and communicate with such external devices.

Once the app is outfitted and running, we need to incorporate its results into actual changes in patient care. This reflects a question I asked of several app developers: your tool will be eagerly adopted by patients who are responsible, tech savvy, and conscious of their health options. But those are the patients already most likely to take care of their health. What about the famously unresponsive patient? (This population was highlighted at the conference by a project called ElizaLIVE.) A useful app will be able to integrate with an EHR or PHR, along with devices consulted by the patient and any other interventions planned by the doctor or by public health officials.

Thus, the ecosystem I envision looks like this figure:

A health care app at the center of an ecosystem

So the health care field is left with plenty of challenge, after all, to handle its challenges. A lot is going on in this area. The National Cancer Institute, Walgreens, and Sanofi-aventis all announced new challenges. The University of Michigan announced a graduate program in health informatics, and the Robert Wood Johnson Foundation announced a Health Data Consortium (of which O'Reilly Media is a member) for pushing the incorporation of data into more apps and into health initiatives in general.

Older posts are this way If this message doesn't go away, click anywhere on the page to continue loading posts.
Could not load more posts
Maybe Soup is currently being updated? I'll try again automatically in a few seconds...
Just a second, loading more posts...
You've reached the end.

Don't be the product, buy the product!