Data is one of the most important tools in responding to the COVID-19 pandemic. Citizens need accurate data to understand the virus and access vaccinations while public health officials need up-to-date information to track COVID-19 and evaluate immunization effectiveness. MeriTalk spoke with Jim Daniel, who leads state and local public health for Amazon Web Services (AWS), to learn how access to public health data has improved and how it may change in the future.
An epidemiologist by training, Daniel previously served as the director of public health innovation for the Office of the Chief Technology Officer at the U.S. Department of Health and Human Services (HHS).
MeriTalk: What successes – and challenges – have public health organizations experienced as they use data to respond to the pandemic?
Jim Daniel: Several states had been in the process of improving their immunization information systems prior to the pandemic. They have been the most successful. In many other states, systems were not ready to take the huge volumes of data coming in, especially as vaccination data becomes more important, and we’re looking at vaccine effectiveness. Some states are taking advantage of cloud technologies to build data lakes that allow them to better link data using machine learning and other capabilities to bring together all of that data across different systems.
MeriTalk: What cloud and data analytics capabilities are states using to meet constituents’ needs during the pandemic?
Daniel: Cloud technologies are very helpful as states try to build new infrastructure to do analytics in new ways. Certainly, AWS’s data lake and lake house architectures are ways that states can build that type of infrastructure quickly and have the compute power as well as the tools that they need to do those linkages and data analytics.
Scalable, cloud-based call centers have been incredibly useful in pandemic response. Many public health departments did not previously have this capability. When states opened mass vaccination clinics the first time back in February and March, the websites were confusing. States realized they needed to quickly deploy call centers to answer constituent questions.
And those call centers, built on the cloud, are able to scale quickly and help answer hundreds or thousands of calls per day. We’ve heard stories of call centers being turned on and within the first few hours 300,000 calls are coming in. As states move into third doses and pediatric doses of vaccines, they will need those call centers again.
MeriTalk: Public health organizations are still encouraging people to get immunized and to get their booster or publicizing that they can now get their children immunized. Are other IT challenges arising as they try to get the word out?
Daniel: One of the things that worries me is the dramatic increase in the number of queries and the volume of immunization information in the next few weeks as states start doing third doses and pediatric vaccinations at the same time. Constituents who want booster shots will query state immunization information systems to learn their current history; they want to confirm that the appropriate timeframe has passed since the second dose. Those queries will hit the immunization information systems at the same time as state are reporting new doses for children.
The volume is really going to stress the immunization information systems. Information systems that are running in the cloud and are scalable are going to be successful, and the others will still have a lot of challenges. A complicating factor is that states don’t really know at this point where kids will get their vaccinations. Children’s hospitals, pharmacies, community health centers, and schools could all play a role, but those institutions don’t necessarily have the means to report a vaccine administration to an immunization information system. States need to make sure that the infrastructure is in place to enable that reporting.
MeriTalk: Over the last six months to a year, have you seen substantial movement toward cloud-based systems? Or are states in a holding pattern – just dealing with what they have?
Daniel: Unfortunately, once the pandemic started, public health resources were so stretched. When states didn’t have vaccines to administer, the immunization teams were brought over to help with contact tracing, surveillance, and other issues. They didn’t necessarily have time to prepare for vaccinations.
I did not see a lot of migration to the cloud once the pandemic started, but a few states did. They did not try to switch to a new product, but they did successfully migrate their current system to the cloud. It was definitely worth the investment for them.
MeriTalk: How are artificial intelligence (AI) and machine learning (ML) helping states address pandemic challenges?
Daniel: AI and ML are definitely new in the public health space, and it’s challenging for public health to even think about new technologies right now as they’re preparing for another round of vaccinations. But AWS is trying to help educate public health customers about ways that AI and ML can help.
Immunization information systems have difficulties sharing immunization status with external partners like payers, which are very interested in knowing their members’ immunization status. Ordinarily, payers get information about a vaccination because there’s a claim associated with it. But COVID-19 vaccinations were administered through mass vaccination clinics underwritten by a Federal payer if there was no insurance. Because insurance information was often not collected, payers don’t have the claims to know their membership’s vaccination coverage. They need state immunization information systems to share that data, but the states don’t have the people power, and their immunization information systems don’t have the bandwidth to do linkages with huge datasets.
AI and ML can help with linking membership files to state data, and AWS is developing some proofs of concept to show how that can work. Hopefully that can be a way that we can use AI and ML to help our public health departments.
MeriTalk: How is the average health consumer’s IT experience different now than it was before the pandemic?
Daniel: The biggest difference is people would now like access to their immunization data – and states are making it easier to obtain. When I worked for HHS, we had started a consumer access immunization data project – although we never saw this as a possible use case for it. States that implemented consumer access projects on AWS have been extremely successful, and more states have implemented similar programs.
Previously, people only needed that information for their children to go to school or camp. But now, travelers need to have a QR code with immunization data to fly to Singapore or Aruba. AWS is working with the Payment Card Industry Security Standards Council, which has put out a standard for QR codes to allow for travel to countries that are requiring it.
MeriTalk: What kinds of changes would you like to see in the future in terms of the use of healthcare data?
Daniel: I would love to see public health information integrated into the consumer’s personal health record. Right now, public health agencies are focused on just getting immunization data to them. But I’d love to see our consumer access partners help integrate that information into consumer portals that have patients’ entire health record.
Data sharing across jurisdictions would also help immensely. It’s something that everyone is working on.
I’ll tell you a personal story. I live in Washington, D.C., and I never made the lottery to get a vaccination. I knew there were plenty of appointments available in Maryland, so I drove a couple of hours into Maryland to get my first two immunizations.
When it came time for my booster, I had to decide whether to get it in D.C. or Maryland. Because so many states are not sharing data across jurisdictions, I made the decision to go to Maryland. That ensures that all of my data is in one place. If I need to travel internationally, all my doses would be in a single place, and a single QR code would have all of that information. Ideally, in the future all vaccination information will be in one place, regardless of where the immunization occurred.
MeriTalk: COVID-19 demonstrated that public health agencies needed to reach their constituents in new ways. How has AWS helped agencies respond?
Daniel: AWS has helped in many ways. Reaching constituents with call centers was the most important, especially during the first round of vaccinations, which focused on at-risk populations that could not navigate complicated websites. The call centers based on AWS were an amazing resource for public health to connect with their constituents. AWS is also helping our customers provide better dashboards that are used to communicate statistics and other useful information. Some jurisdictions have detailed information about vaccine breakthroughs based on which vaccines people had.
MeriTalk: Right now, what are public health officials most critical data needs?
Daniel: It’s becoming increasingly apparent that linking hospitalization data, immunization data, and case data in real time is critical to understand what’s happening across the U.S. A few states do it really well, and AWS is helping other states do it with data lakes and other cloud-based technology. But public health providers don’t have enough data to really figure out what’s happening. They get this data from the Centers for Disease Control (CDC) but because states are having difficulties linking all the different kinds of data, the CDC doesn’t have a complete picture.
Most states are pretty good about connecting case and immunization data – and they get de-identified hospitalization data. But they really need real-time hospitalization data with identifying information. That is a key part of a public health infrastructure and answering questions such as whether vaccinated people are hospitalized at a lower rate than unvaccinated people. There are major policy implications there.
MeriTalk: What advice do you have for states that are struggling with linking multiple sources of data?
Daniel: I recommend that states work backward a little bit and make sure that everyone’s on the same page with the questions that they want to answer. Then they can look at where that data resides and make the right policy decisions to answer those questions. Then they can build the technical solution that brings everything together into a data lake. The technical pieces are pretty easy to build, but there are often major policy issues around who can access that data.
MeriTalk: How can states learn more about increasing visibility of health data?
Daniel: Our webinar with state public health leaders in Louisiana, Massachusetts, and Minnesota provides great first-person perspective and lessons learned. And I would encourage everyone to join us for IMAGINE: The Modern Citizen Experience, a discussion series that explores the latest challenges and opportunities government organizations are facing as they seek to improve citizen services.
Content sponsored by AWS