Latest / Public Health Review Morning Edition / 1096: How the CDC Is Forecasting the Next Outbreak
Transcript
- 0:00This is Public Health Review Morning Edition for Tuesday,
- 0:02March 31st, 2026. I'm John Sheehan with news from
- 0:06the Association of State and Territorial Health Officials
- 0:09today, how the CDC is using a new generation of data tools to
- 0:13forecast the next outbreak and is transforming how we detect
- 0:17and respond to infectious diseases.
- 0:19Jason Asher, director of the Center for Forecasting and
- 0:22Outbreak Analytics at the CDC, joins us to explain how for
- 0:26decades, public health has relied on lagging data telling
- 0:29us what already happened. But now, Asher and his team are
- 0:33building systems that turn real life data streams into
- 0:36actionable forecasts, simulations and data making
- 0:39tools for health departments across the country.
- 0:43You know I invite you to consider just how much on a
- 0:46daily basis, we each rely on predictions to make decisions.
- 0:50When was the last time you drove somewhere far away without using
- 0:53a navigation app? It's probably been a while.
- 0:55You rely on it to adjust your route and departure to make sure
- 0:58you arrive on time. Imagine you're packing for
- 1:01travel in the week ahead. You'd almost certainly check the
- 1:04weather forecast for your destination.
- 1:06Perhaps you'll need some warmer clothes or a raincoat.
- 1:08If you buy something online, you likely use the anticipated
- 1:11delivery date as a factor in your decision.
- 1:14Each of these are examples of predictive analytics.
- 1:17They help us anticipate what's coming and make better decisions
- 1:20about our time and resources. But historically, public health
- 1:24systems weren't built that way. They were designed to tell us
- 1:27what has already happened, not necessarily what is about to
- 1:30happen next. That means many high stakes
- 1:33decisions during outbreaks are made using incomplete or delayed
- 1:36data. So at the Center for Forecasting
- 1:39and Outbreak Analytics, we're working to change that.
- 1:42Our goal is to close a few key gaps, moving from lagged data to
- 1:47real time signals, from static reports to interactive decision
- 1:51tools, and from broad averages to insights that help guide
- 1:55action closer to where outbreaks are actually happening.
- 1:58One reason the federal government can play a role here
- 2:01is that we're able to bring together multiple streams of
- 2:03data that aren't available in anyone place at the state or
- 2:06local level. And when we combine those
- 2:08signals, we can generate forecasts, models and
- 2:11simulations that give a clearer picture of where things may be
- 2:14headed. And we see this work as
- 2:16returning something of value to the state and local health
- 2:19departments and increasingly, healthcare systems that are
- 2:22sharing data with CDC. So the goal is to turn that data
- 2:25into insights that support the decisions they are making every
- 2:28day. CFA is working towards the
- 2:31future, where a state health department can see early signals
- 2:34that an outbreak may be accelerating in a particular
- 2:36region, and where a hospital system can anticipate a surge in
- 2:39respiratory illness before it arrives so that it can adjust
- 2:42staffing and other resources ahead of time.
- 2:45When we set up CFA just a few years ago, we were essentially
- 2:49starting from scratch. But we've made significant
- 2:51progress in a short time, and the capabilities I'm talking
- 2:54about aren't decades away. With sustained effort, this kind
- 2:58of predictive insight could become a routine part of public
- 3:00health and health system decision making within the next
- 3:03few years. Jason, that's incredible.
- 3:06And as you sort of describe, the implications for are so immense,
- 3:09being able to take data streams from across the nation.
- 3:13What kinds of things are you working on now at the center and
- 3:16what's top of mind? Yeah.
- 3:18So let me share a few examples of the work we're doing right
- 3:21now. One place this work is already
- 3:23making a difference is in measles outbreak response.
- 3:26So we've been working directly with states that are responding
- 3:29to measles outbreaks. For example, in South Carolina,
- 3:32we've been partnering closely with the state health department
- 3:35to better understand how the outbreak is evolving and where
- 3:38it may be headed. We recently completed what we
- 3:41call a scenario assessment for that outbreak.
- 3:44In plain terms, instead of predicting a single outcome, we
- 3:47lay out several plausible paths the outbreak could take,
- 3:50depending on things like vaccination coverage and the
- 3:53strength of intervention measures.
- 3:55That kind of analysis gives leaders a way to step back from
- 3:58the day-to-day response and think more strategically about
- 4:01where the outbreak may be headed.
- 4:03And importantly, we're turning what we learn from working
- 4:06closely with individual states into tools that others can use,
- 4:09whether they're preparing for a potential outbreak or actively
- 4:12responding to 1. So alongside supporting
- 4:15individual jurisdictions, we're also working to make modeling
- 4:18more accessible. One example is our interactive
- 4:21measles simulator. It's web browser based, so you
- 4:24don't need specialized software or a background in modeling to
- 4:27use it. The simulator let's health
- 4:29department's explore what if scenarios for their own
- 4:32communities. They can adjust things like
- 4:34vaccination coverage or how quickly cases are identified or
- 4:38how well isolation guidance is followed, and then see how those
- 4:41changes might affect the size or duration of an outbreak.
- 4:45That kind of rapid exploration can be especially helpful early
- 4:48in an outbreak when decisions need to move quickly and
- 4:51information is still evolving. Stepping back, health
- 4:54departments also tell us that they need tools to help them
- 4:56understand what's happening right now and where things may
- 4:59be headed next. There's two capabilities we've
- 5:02expanding there now. Casting an RT estimation.
- 5:06These tools help answer 2 fundamental questions during an
- 5:09outbreak. What's happening right now?
- 5:11And is transmission increasing or decreasing?
- 5:14So let me start with now casting.
- 5:16Simply put, now casting means estimating what's happening
- 5:19right now, even before all the data have been fully reported.
- 5:23Because reporting delays are common, the most recent numbers
- 5:26are often incomplete. Now Casting helps fill those
- 5:29gaps so decision makers don't have to wait days or weeks for a
- 5:32clearer picture of current conditions.
- 5:35For example, we've been working with state officials in South
- 5:37Carolina to develop NOW Casted estimates during the measles
- 5:40outbreak there, sharing updates one to two times a week
- 5:44throughout the outbreak. We've also integrated NOW
- 5:46Casting international surveillance systems like
- 5:49Respnet to improve the timeliness of available
- 5:51hospitalization data for respiratory illnesses.
- 5:54The goal is straightforward, to reduce blind spots during fast
- 5:57moving situations. And there's another benefit.
- 6:00When we begin with a more accurate picture of what's
- 6:03happening right now, other modeling work becomes more
- 6:06reliable. So better understanding today
- 6:08leads to better decisions tomorrow.
- 6:10Another tool we've been expanding is something called
- 6:12RT, or the Time varying Effective reproductive number.
- 6:16RT tells us how quickly infections are spreading,
- 6:19somewhat like a speedometer in a car.
- 6:21RT can tell us whether infections are increasing,
- 6:24decreasing or staying the same. Used alongside forecasts and
- 6:29surveillance data, RT gives a quick read on transmission
- 6:32trends. So just a couple of years ago,
- 6:35routine state level RT estimates weren't widely available.
- 6:38Now today we produce weekly state level estimates for
- 6:41COVID-19 and during the respiratory season for influenza
- 6:44and RSV as well. But the next step is increasing
- 6:48geographic detail, moving towards sub state or county
- 6:51level estimates. For jurisdictions, that means
- 6:53more localized insight into where transmission may be
- 6:56accelerating, so decisions about outreach and interventions can
- 6:59be more precisely targeted. That's incredible.
- 7:03It sounds it sounds almost like a like a weather forecasting
- 7:07map, but for diseases for the country.
- 7:09Yeah, that's right. I think the the metaphor of, you
- 7:12know, making infectious disease forecasting as commonplace as
- 7:15weather prediction, that's a major goal of ours.
- 7:17Wow. And obviously it sounds, it
- 7:20sounds as you describe it to be such a useful array of tools
- 7:25that it may and it makes complete sense that agencies
- 7:28would use them. Do you develop them in concert
- 7:31with member organizations? Member.
- 7:33Agencies, yes. So our our tools, they only have
- 7:36impact if they work for decision makers, especially state and
- 7:39local public health leaders. So that's why we actively seek
- 7:43feedback from state, territorial, local and tribal
- 7:45partners, often in collaboration with ASTO and other public
- 7:49health networks. For example, last spring, more
- 7:52than 140 public health professionals shared feedback on
- 7:55our Respiratory Disease Season Outlook, an early assessment
- 7:58designed to give health leaders a sense of what might be coming
- 8:01during respiratory season. You told us that understanding
- 8:04combined hospital demand for COVID-19, RSV and influenza
- 8:08during the fall and winter was the most useful piece of
- 8:11information. So this year we made that
- 8:13prediction more prominent in the outlook, including in the final
- 8:15update that we released earlier this month.
- 8:18You also asked for RT estimates for RSV and more plain language
- 8:22interpretation of what those numbers mean.
- 8:24So we built that capability. Now we're working on an extra
- 8:27quest, more localized insights at the sub state level and
- 8:30partnering with states like New Mexico to evaluate how useful
- 8:34these estimates are during the earliest stages of an outbreak.
- 8:37We're also improving the measles simulator that I mentioned
- 8:40earlier based on early user feedback.
- 8:43A new version released last month includes A baseline
- 8:45immunity calculator, the ability to model smaller populations,
- 8:49and downloadable results for planning discussions.
- 8:52These are based directly on feedback from state and local
- 8:54partners. They didn't just come from us
- 8:56sitting in a room, they came directly from people.
- 8:59People who may be listening to the podcast right now telling us
- 9:02what would make tools more useful in real world decision
- 9:05making. And are you finding that
- 9:09agencies now like getting to use these tools are are finding
- 9:13their their? Operations opened up.
- 9:15That they can increase their capacity now thanks to these
- 9:18more efficient tools. Yeah.
- 9:20So one of the most significant ways we support state and local
- 9:23health departments is through Insight Net to CFAS, national
- 9:26network of innovation hubs. It includes nearly 130 partners
- 9:30across 30 states spanning the academic, public and private
- 9:34sectors. So through Insight Net, we're
- 9:36developing new forecasting and modeling tools while also
- 9:39helping to build that capacity you mentioned within state and
- 9:42local public health departments to identify and respond to
- 9:45outbreaks earlier. Across the network, partners are
- 9:48building tools that help health departments understand
- 9:51vaccination coverage, identify outbreak risk and simulate how
- 9:55diseases might spread in their communities.
- 9:57So some examples are Clemson supporting the ongoing measles
- 10:02outbreak response in South Carolina with real time modeling
- 10:05and weekly situational updates, School level vaccination
- 10:08dashboards developed by UNC for North Carolina, National risk
- 10:12mapping tools developed by Johns Hopkins, and localized outbreak
- 10:16simulators used by cities like Austin, TX, just to name a few.
- 10:20But we've also heard clearly the building tools alone isn't
- 10:23enough. Health departments want support
- 10:25and understanding how to apply them and confidence in using
- 10:28them. So we've also focused on
- 10:30creating plain language resources designed specifically
- 10:33for public health professionals who may not have formal training
- 10:36and modeling. One example is our Modeling
- 10:39Handbook, which explains what different types of models do,
- 10:42how they work, and how to interpret their results.
- 10:45We also publish practical guides and short explainers through our
- 10:49Behind the Model series, where we walk through how our
- 10:52forecasting and modeling methods work in plain language.
- 10:55The goal across all of this work is to make modeling information
- 10:58easier to understand, easier to access, and easier to use.
- 11:02Because capacity building isn't just about developing new tools,
- 11:05it's about making sure people feel equipped and confident
- 11:08using them in real world decision making.
- 11:11You mentioned that it's only been in the last few years that
- 11:13these data streams have been able to, you know, be compiled
- 11:17and and open up these new insights.
- 11:20What do you see moving forward? As you know, computing power
- 11:24gets gets stronger and more and more data streams open up.
- 11:29Yeah. So we see a future where disease
- 11:31forecasting is faster, more localized and easier to
- 11:36integrate into everyday public health work.
- 11:38In many ways, we think about this, like I said earlier, like
- 11:41weather forecasting, it's not just about high performing
- 11:44predictions. It's about giving you advance
- 11:46notice so you can prepare, providing insights specific to
- 11:50your community and updating information in real time as
- 11:53conditions change. Looking ahead, our focus is on
- 11:56making updates more timely, increasing geographic detail,
- 12:00expanding coverage across multiple pathogens, and
- 12:03continuing to strengthen partnerships with states,
- 12:06territorials, and local health departments.
- 12:09But the goal isn't sophistication for its own sake.
- 12:12It's really usability tools that fit into real world situations
- 12:16and help you act earlier and more confidently.
- 12:19As we talked about earlier, predictive tools are already a
- 12:21part of how we make decisions in everyday life.
- 12:24We're working towards bringing that same kind of anticipatory
- 12:27insight to public health. Jason Asher is director of the
- 12:32Center for Forecasting and Outbreak Analytics at the CDC.
- 12:36Pull up a chair and join ASTO for a fireside chat with a
- 12:40subject matter expert in radiation readiness
- 12:42communications. In this webinar, titled Past,
- 12:45Present, and Future Reflections from a Radiation Readiness
- 12:48Professional, Jessica Wieder will share reflections on her
- 12:52career and insights from her experience in radiological
- 12:54preparedness and emergency communications.
- 12:57This webinar will include discussions around key
- 13:00considerations or effective radiological risk communication
- 13:03from a public health preparedness perspective,
- 13:06including coordinating messages across partners and addressing
- 13:09public concern and false and misleading information.
- 13:13The conversation will also highlight practical tools and
- 13:16readiness resources that can. Support state.
- 13:18Local and territorial partners, the Public Health Infrastructure
- 13:23Grant, or FIG, is driving progress in strengthening public
- 13:26health systems nationwide. Want to stay in the loop?
- 13:29The FIG National Partners Connections newsletter is your
- 13:32one stop source for funding and engagement opportunities,
- 13:36program updates and success stories, events, training and
- 13:39webinars, tools, resources and more.
- 13:42Subscribe now at the link in the show notes and stay connected to
- 13:45the national conversation shaping the future of public
- 13:48health. This has been Public Health
- 13:50review Morning edition. I'm John Sheehan for the
- 13:52Association of State and Territorial Health Officials.