Nirav Shah, M.D., M.P.H., is the New York State Commissioner of Health.
Travis Korte: Tell me about some of New York State’s health data initiatives.
Nirav Shah: Two years ago, when Governor Cuomo started in New York State, we were facing a fiscal cliff, with $53.5 billion Medicaid program that was growing at a rate of 13 percent every year. It was using up all the tax dollars we had, crowding out education and other important sectors. What we did in two months was get all the stakeholders together—hospitals, CEOs, unions, doctors, advocates, legislators—and came up with 4,000 ideas on how to fix Medicaid. We boiled them down to 250, and two years ago in April, the legislature passed 78 of those ideas. Within the first year we saved $4 billion; we’re on track to save $34 billion in our state Medicaid program over five years.
One of the fundamental anchors was real-time data-sharing with all the relevant stakeholders. We have monthly reports on our Medicaid website that break out by sector, to the dollar, how much we’re spending in Medicaid. In real-time, we can look at a hospital and see that they have a spike in Medicaid claims for sepsis because we’re paying for it. In near real-time we can call up the hospital and say “What’s going on in terms of quality?” Sometimes before they even know it, we can call them out on it. That kind of real-time monthly data-sharing among stakeholders shows the power of data in New York State. Policymakers all talk about data but this is real data in practice: empiric findings to guide policy.
We also saw real-time use of data on a procedure called percutaneous coronary intervention (PCI). Two years ago, in New York State, 23% of elective PCIs were inappropriate, according to American Heart Association guidelines. We publicized that, we shared data down to the individual provider level, and we threatened to not pay them for inappropriate cases. To date, the inappropriate PCI rate has dropped to 8%, and Medicaid has saved $4 million. By freeing the data and using it to guide policy, we achieved incredible improvements in quality and patient safety, while bending the cost curve and making improvements in population health for New Yorkers.
Korte: Talk a little about the role of public-private partnerships in the New York Department of Health’s data strategy.
Shah: Basically there’s a “5-50 problem” in health care in New York and everywhere else. 5% of the patients consume 50% of the healthcare dollars: these are sick people with two or more chronic conditions. How do you help them? We know you can help these high-cost patients if only you did simple things like reconcile medications between hospital discharge and primary care. But nobody does it because no one has the data. Instead, there’s a faded copy from the hospital discharge that lists your meds, and there’s the meds your primary care provider thinks you’re taking. Then CVS has a different list.
In New York, we have the SHIN-NY (Statewide Health Information Network – New York), with data from all those different sources. We’re connecting all of the electronic health record (EHR) data from all the different systems across the state in one statewide network. This is extremely helpful. If you show up in a Brooklyn ER and you’re from Buffalo, they can pull up your allergies, your medication list, and your clinical diagnostic tests in real-time no matter what EHR system they use. That incredible interoperable system is the result of a consortium led by New York State. Since it was launched, 19 other states and many major vendors have joined in to set interoperability standards.
In New York, we also have apps being developed and funded by a program called the Digital Health Accelerator. In the first round, 250 companies applied, eight were selected, and one has already been bought out. Strategic investors committed venture capital, while 23 large health care systems, federally qualified health centers, and doctor’s offices offered to mentor these companies. They told developers, “This will help us. This won’t help us. This is how it would help us.” Then they came up with programs that providers would actually use. Seventeen have actually been implemented in these provider organizations. Hospitals are using the apps that were created by these companies to do real-time medication reconciliation (safety checks on medications ordered). That’s an example of a public-private partnership where now we’re on track to get $150-$200 million in venture capital to create the next Silicon Valley in New York State around health data apps. We’re also on track to create 1,500 jobs over the next three to five years. These are just primary jobs that do not include all the support jobs that will result.
Korte: You’ve told me about the technical vision and the economic vision for New York State health data, but I’m still wondering about the intellectual vision. What do you hope can be learned from this data?
Shah: In New York State, we publish open government health data on health.data.ny.gov. We’re releasing dozens of high value data sets on a Socrata open government platform, including things like restaurant inspection data for the State of New York. Why should the data sit on a server in the Department of Health? Citizens paid for us to collect it. We just published all of that data on our website, and now Yelp is going to put all that data into their reviews of New York State restaurants.
We also have weekly nursing home census data. When Hurricane Irene hit, we had this data up. It was during a weekend, and dozens of nursing homes had to evacuate. But they didn’t have to call us up to find out where to place evacuated patients. They just went online to see the weekly nursing home bed census data, found the closest empty beds, and moved their patients out of harm’s way. There’s incredible power to freeing all this data.
We have over 300 data sets in the Department of Health including data about Medicaid data, claims, public health, water quality, and air quality. You name it, we have it.
Student body mass index (BMI) is another example. For 680 school districts across the State of New York, we have BMI, height and weight on every student in second grade, fourth grade, seventh grade and 10th grade. Imagine understanding how healthy a community is at a school district level based on rates of pediatric obesity. That one measure of BMI can tell you about the availability of fresh fruits and vegetables, the walkability of the streets, and the safety of the parks, combined into one high quality variable that you can use for planning and policy. When you mash it up with other data sets, it’s incredibly powerful.
People are creating apps in real time because our health data site has an open application programming interface (API) that everyone knows. And we’re federating local and federal data into that state site as well. It’s a one-stop shop. What we’ve created is an ecosystem of data that’s highly accessible, of high value, with really well-documented metadata. We’re now creating app hackathons, codathons and challenges through the fall and into next year, where we will pay people and create prizes to create apps off of this data.
Korte: In your recent talk at the Bipartisan Policy Center, you mentioned your goal of having an “anticipatory health system.” Can you go into a little more depth on that?
Shah: Right now health care is extremely reactive. You get congestive heart failure, you end up in the hospital, they put you on medication. What about anticipating your needs? If you’re a slender 43-year-old woman who lives in Albany and spends 60 hours a week behind a desk in an office job, the number one requirement for your health isn’t waiting for the standard, age-appropriate bone density scan when you’re 65, but being screened earlier and starting Vitamin D. If we actually had this vast amount of data on where people live and what their risks are, the data would tell us that today. It would personalize care at a level that has never been done before.
We talk about genetics as the next big revolution. But genetics only results in 30% of preventable mortality. Health care quality only results in 10%. The rest, fully 60%, is determined by social and behavioral and environmental choices. Your zip code is more predictive than your genetic code of whether you are going to live or die. We know this. It’s been ignored because it’s been hard to do anything about, since we haven’t had the data.
You have to start with security and privacy of the data. People own the data. No one’s going to get any data that you don’t want them to get. But once you share the data you will get real value from it that is pertinent to your lifestyle and your behaviors. We’re going to pay to create that value by hot-spotting what we know works. Within the next year, we will have that big data platform in New York State for health-related data that includes public health and healthcare. It will unleash the creativity and power of the private sector while safeguarding security and privacy with strict penalties and a regulatory framework that only government can provide.