Announcer: This podcast is intended solely for educational purposes and presents information of a general nature. It is not intended to guide or determine any specific individual situation and persons should consult qualified professionals before taking specific action. The views expressed in this podcast are those of the speakers and not those of Milliman.
Grace Bello: Hello and welcome to Critical Point, a podcast brought to you by Milliman. I’m your host Grace Bello.
Today, we’re on the case, solving the mystery of the missing influenza season.
Around fall of 2020, due to the Covid-19 pandemic and the looming flu season, public health experts feared the possibility of a “twindemic”—a one-two punch of Coronavirus…plus influenza.
At the time, Covid-19 vaccines were still in clinical trials. And because of the threat of the virus, patients were more reluctant than ever to seek in-person preventive care, including getting a flu shot.
Public health experts predicted that Covid-19 cases plus the onslaught of flu cases would overwhelm an already strained U.S. healthcare system.
But here’s what happened in 2020: Flu cases hit historic lows. The so-called twindemic? It never happened.
At Milliman, we wanted to find out why the volume of flu cases decreased. And, more broadly, what can the epidemiological data tell us about future airborne disease preparedness--and prevention.
This line of inquiry led to our newly released report, “The Mysterious Case of the Missing Influenza Season.”
Here with me today is the lead author of the paper, Ellyn Russo, Healthcare Consultant at Milliman. Ellyn specializes in researching healthcare services and has experience studying vaccines and influenza. Hello, Ellyn.
Ellyn Russo: Hi, Grace. It's good to be here with you. Thanks for inviting me.
Grace Bello: Yeah, thanks for being here. So this is a really interesting paper. I'm really excited to delve into it. I wonder what prompted the study.
Overview
Ellyn Russo: Sure. So as you know and many of our listeners will know, too, Milliman's an actuarial firm. My path, though, is a little bit different than the typical. Before joining Milliman, I was engaged with research on influenza in particular. We were studying the burden of infection in the Veteran's Health Administration population as well as the comparative effectiveness of different influenza vaccines. So I come with a more health services and clinical research background as opposed to actuarial science. I also have a background slightly in the virology space, so I was always very interested in how viruses compete and sort of that evolutionary game of what influenza might do with a cousin or another virus around. So there's just a lot of talk that was going on. Some folks were interested in really mapping out what health care providers, health care administrators, insurers could consider in terms of the impact--that twindemic impact--if it truly came to light. And we just were monitoring the situation. We looked at, you know, all of the news around the southern hemisphere not having a lot of cases, what else we might consider as flu season starts and kind of how we might monitor it. So that's what really set the background stage for this. As the winter picked up and 2021 came about, there were even more questions in wondering as we saw the cases were not picking up. We didn't hear about it in the news. Flu really wasn't happening. And then we realized we had a great data set to use to actually look at, you know, A, is this true; and B, why might that be happening?
Delving into 2020 flu data
Grace Bello: What were your theories in terms of why the amount of flu cases was so low?
Ellyn Russo: So one of my original theories was that yes, there was just competition and the flu virus just could not grasp hold of infecting humans because COVID was doing such a good job, unfortunately. But, you know, when it comes to measuring what you know, I don't know how to measure that. I'm not a virologist. I'm not a microbiologist, so focused more on the burden to the healthcare system.
Hypothesis 1: Vaccination increased
Ellyn Russo: What you mentioned also in the introduction was that the general population might have had some fears or some hesitation of getting the influenza vaccine purely because we were sort of in the mindset of staying away from the healthcare system so that individuals who really needed care who were suffering from COVID could get the care they needed. I sort of had an opposite thought process of, you know, we're talking so much about vaccine in the context of COVID. Maybe that will actually help raise awareness of the importance of vaccination overall and people might seek vaccination against influenza more. So that was one theory, that maybe vaccination against flu increased.
Hypothesis 2: Testing decreased
Ellyn Russo: Another one, and this gets back to kind of what was going on worldwide, you know, testing—my colleague brought up testing. Well, maybe we're just not testing for flu and we're just missing cases. We're so focused on COVID, are providers thinking about flu? Is it on the short list of the differential diagnoses for symptoms? Symptoms do differ. They certainly are different for COVID and influenza.1 So maybe, you know, as a provider, you're thinking COVID first, and the symptoms don't really look like flu, so you're not even testing for it.
Hypothesis 3: Transmission avoided
Ellyn Russo: And then the third one that sort of, I think, encompasses all of them, the other two as well, is this idea that transmission was avoided. Meaning, you know, life as we knew it was completely disrupted in 2020. We all adapted to having to wear masks, to washing our hands more, to not going to certain places. And that's exactly how we prevent flu, right? It's transmitted very similarly. Again, we knew there were low rates of it worldwide. We weren't seeing it in the Southern hemisphere, which is where the flu season typically starts. So in our summer in the northern hemisphere, it's winter in the southern hemisphere and that's where flu is typically starting. It wasn't happening. There were barely any cases. And more on the virology of it, the unique aspect of the basic reproduction number—which is how many cases one individual can transmit is actually substantially lower for flu than for COVID2 and for some other vaccine preventable diseases.
"[To prevent the spread of COVID-19,] we all adapted to having to wear masks, to washing our hands more, to not going to certain places. And that's exactly how we prevent flu."
Grace Bello: Right. And so to clarify that last point, you're basically saying that COVID is more contagious than flu and so if people are protecting themselves against COVID, they're very much protecting themselves against flu.
Ellyn Russo: Right, flu really can't even be there because COVID is the one that's being transmitted a lot easier. But if we’re preventing that, it's preventing how any disease is transmitted, any sort of respiratory illness that's transmitted in a similar way.
The role of MedInsight
Grace Bello: Right, right. So how did you and your team conduct the study? What data were you relying on?
Ellyn Russo: So this study was a collaboration across several of us here at Milliman. So we actually were able to use our Milliman MedInsight Emerging Experience research data set. It's a suite of healthcare analytical tools that helped healthcare entities leverage their data. So really studying and identifying trends and anomalies in data that might happen over time. It's more real-time than some of our other data sets. So we had data through December of 2020 as compared to some of our other data sets only went up to 2019. So we were actually able to see 2020 experience data from healthcare entities that submit their data to us. Med Insight differs from other data sources. So as I mentioned earlier, we knew flu wasn't circulating because entities like the Center for Disease Control and Prevention (CDC), the World Health Organization (WHO) who typically monitor the presence of flu and the burden of the season, they were already presenting data, right? We knew that the CDC, they've got systems to look at both how many people have flu--so really, the timing and intensity of influenza activity, how many people are being diagnosed on a weekly basis—that was already being studied. We knew those rates were low. They also looked specifically at what types of influenza viruses are causing infection, and this is where we get back to that virology theme. That's contributing to how we design and tweak the influenza vaccine, so it offers maximal coverage and it's optimized to preventing the infection. The CDC looks at outpatient visits, hospitalizations, mortality through a very intricate and detailed sampling method where, yes, certain practices, outpatient practices or hospitals are contributing all of their data at any given time for how much flu is happening and what those viruses are. But they're not necessarily available at that level so at the practice level, at the hospital level, at the region level, they're not as actionable for all users of the data. So, like we saw with COVID overtaking our healthcare system, the more information we have, the better we're able to allocate resources to help with the burden of disease if pockets of disease erupt and it's really overwhelming certain areas. And like I said, modifying the vaccine. So really, it's kind of more of a population focus. Which is similar to what Med Insight does as well, but there, you know, CDC and WHO are looking very, very high level to help prevent disease and address symptoms and the burden of it when it's really bad.
"The more information we have, the better we're able to allocate resources to help with the burden of disease."
MedInsight has--and what Milliman uses for our data sources--is predominantly claims data, so we're looking at a variety of insured populations, commercially insured individuals, government insured individuals: Medicaid and Medicare, Medicare Advantage. It's a different look at the same population level of detail, but it's for a different population in the sense that you might look at your total insured population in an area who may or may not be seeking health care. So you're able to see different things going on with your patient population in addition to what CDC and WHO are reporting. And that's why we had a particular interest in doing this. We thought we could look at a slightly different data source that sort of complements what's already being done but also allowed us to look at and contribute knowledge about insured populations, again where there's access to healthcare insurance and therefore healthcare claims that we could see. The overall goal really is to look at, you know, were we seeing the similar decreases in the rate of illness? You know, what happens in a healthcare claim in terms of what gets coded on that claim may or may not line up with the actual results of a test or kind of the downstream effect of what testing happens and what happens for that care for that patient. We took a different approach to look at...were we seeing the same trends and what else might we learn in the context of vaccination, testing, and transmission avoidance?
Grace Bello: So it sounds like your data was a little more granular—and, actually, if you wanted to, you could drill down by geography, by location.
Ellyn Russo: Exactly, exactly. And also have these other points of, you know, one individual connecting their vaccination status, their testing status—looking at sort of those interplays. We didn't exactly do that in the study, but that's what this data offers.
Flu transmission and flu-associated pediatric death
Grace Bello: So, let's zoom out a little bit and talk more broadly about influenza. So I know that flu is really important to study for a number of reasons and one is the potential impact on kids. So one part of it is I read somewhere that kids can be a major vector of the flu and kids, young kids in particular, are very prone to the flu. So how did that figure into your research?
Ellyn Russo: Right. So getting on to that point, yes, children are a very big vector of the transmission for influenza virus in particular. They're known to have the disease for, or infection for, a little bit longer than other age groups, so they're able to transmit it a little bit longer and therefore expose others. We did, we gave some thought to this in terms of, you know, not wanting to limit to certain risks or age groups for flu. For example, in the season prior to the current one, so in 2019-2020, hospitalization rates among 18 to 49-year-olds were actually higher than during the 2009 H1N1 pandemic. We don't want to forget about the burden across the entire population. But yes, there are definitely differences within season-to-season virulence of who is being impacted the most by the virus. I think we're sort of accustomed to an annual burden of flu. Meaning, you know, every year, every season there's a couple hundred individuals under the age of four that do pass away from flu. We know infection from flu is a major contributor to death among those 65 and older and among those with high-risk conditions. So it's something we felt very important-- was a very important topic, meaning, you know, continuing to understand the dynamics of influenza infection, what sequelae are associated with it on our…on the population, the health of our population on the use of healthcare resources. It's important to understand what dynamics are affecting that so that, you know, there's a holistic view of, every single year we know this is coming, how do we plan for it, how do we account for it and what do we do if it's different--if it's either really bad or really good, as it was for the past year. So that all played a role. We wanted to not necessarily take away from the fact that there was this very devastating other pandemic going on, right, that COVID-19 is very much a focus and needs to be, but flu continues to plug along and is sort of always there.
"Every season there are a couple hundred individuals under the age of four that do pass away from flu. We know infection from flu is a major contributor to death among those 65 and older and among those with high-risk conditions."
Demographics of influenza transmission
Grace Bello: Right. And you bring up a good point, which is that in 2020, COVID-19 took over the headlines, but flu is an annual problem, an annual challenge in the U.S. and elsewhere. So let's talk a little bit more about the demographics. You were mentioning how young kids in particular, and older adults are really prone to influenza. So can we talk a little bit about those vulnerable populations and what the impact is from flu?
Ellyn Russo: So, yes, let's talk about those that are at highest risk for complications from infection due to influenza. So similar to when you think about, you know, what we hear in the news about COVID, it's not so much the actual virus that's contributing to sort of bad outcomes, it's how your body responds to it and what your immune system does as a result. It sort of gets overwhelmed with this infection. So when we talk about the burden of disease, we're talking about what happens, who's at risk for getting those complications and ultimately suffering, either sequelae, meaning they're in the hospital for longer because they can't breathe very well or it's affecting other organs. For flu, a lot of the populations, as I mentioned, so under four years of age is sort of a very high-risk group, over 65 is another very high-risk group, and then individuals with high-risk conditions. So some of these might be pregnancy, heart disease, other respiratory conditions, any immunocompromising state. They're sort of well-known to be at highest risk for getting those complications. We know it leads to death, as I said, in the underage, under four years of age group and over 65 years of age. It sort of leads to death at higher rates than other age groups. COVID on the other hand, as we know, is disproportionately impacting the elderly population, so over 65 years of age. Not so much in the younger kids, but that's where flu kind of has a lot of its impact.
"We've seen the data on COVID disproportionately affecting different minority groups, different socioeconomic statuses. It's not unique to COVID. It's not unique to influenza, unfortunately."
And we also know that there's some socioeconomic differences in terms of who's affected by it. So similar to how we're describing, you know, kids staying home from school, any time you've got people in close proximity, that's how respiratory illness transmits. So when we're thinking about people that might be of lower income or living in different areas that are very heavily densely populated, any time you're sort of living in a household with a lot of people, you're more at risk. Any time there's somebody who in that household then gets sick, it's easier transmitted because it's hard to leave your house just because one person is ill, so a lot of people end up getting sick. And also along that line, people who are in some of these lower income brackets might also not have been able to avoid schools or daycares that were still open for their kids because of their job and the need to work and stay employed. We saw that happen with COVID as well. We've definitely heard and seen the data on it being disproportionately affecting different minority groups, different socioeconomic statuses. And it's not unique to COVID. It's not unique to influenza, unfortunately.
Inequities in healthcare
Grace Bello: Right. And that goes back to something we've been exploring more broadly at Milliman which is the inequities in the healthcare system, who gets sick, who gets the most sick and who needs the most care, right?
Ellyn Russo: And building up that a little bit, too, Grace--you know, this idea that you might start getting symptoms from the flu or COVID even, but you still might not be able to go to the doctor or have a quick visit, telehealth or whatever, and sort of say, "I'm having these symptoms." There are treatments for flu, at least, that you can start very early on. But if you're not able to even take the time to have that visit, you might end up having, you know, a more severe case because, you know, number one, maybe you didn't get vaccinated because you didn't have time to get vaccinated either. Number two, you couldn't get in to your doctor right away and get that antiviral medicine, so you end up with a more severe case. And that's where we see the sort of burden happening the most when it comes to the inequities.3
Grace Bello: Right. You weren't able to get time off work to get the vaccination. Maybe it's hard for you to get that preventive care and once again get time off work, right. So I think what's interesting is that the way that your team looked at the data, it was much more granular and really was looking at the different ways that influenza affected populations I was wondering if you found any other kind of demographic information that you could compare and contrast.
Ellyn Russo: So we actually did not do too much of that, just to sort of keep this very high level. But what I'll say--and maybe walking through some of the findings will be helpful--we, you know, we certainly right off the bat, we saw the dramatic decrease. So rates of influenza office visit, emergency department and hospital inpatient utilization were anywhere from 86 to 94 percent lower for the season. So when we say dramatic, that is about as dramatic as you can get. Personally, in my experience, I've never seen anything like that before. So that's a huge decrease. Flu still was being recorded as a diagnosis. That's why we're still finding it somewhat; it's not zero. It was still being recorded. Either it was suspected or it did end up being an infection for that person. But that was across the board. We definitely saw, you know, the highest rate which is, you know, very slightly higher than the rest was in the Medicaid population. So as compared to what you might consider a more healthy, less socioeconomically burdened population of the commercially insured individuals, Medicaid had a slightly higher rate of influenza infection.4
Did influenza vaccination increase?
Ellyn Russo: Vaccination against influenza—the big takeaway here is, when you look at the average that we were able to find in our data, you know, we're only seeing one in five were vaccinated across the board, you know, across the entire population we studied, one in five received vaccination. That's pretty low. If you look at the CDC rates, it's closer to 50 percent; it's more like one in two are receiving a vaccination.
"Rates of influenza office visit, emergency department and hospital inpatient utilization were anywhere from 86 to 94 percent lower for the season."
Ellyn Russo: We certainly know a limitation of our data is that we might be missing some vaccination, especially if it happens at work, so your employer might offer a free influenza vaccination clinic and they're not going to run it through your insurance. So we know we're missing some, but to see a wide distribution in terms of, you know, 1 in 10 in the Medicaid population being vaccinated. Closer to three out of every four in the Medicare Fee-For-Service population receiving vaccines. That's a pretty substantial variation in terms of like we just said, getting the preventive measure that has been proven effective to reduce the burden of infection. Even if you still get the flu, having had the vaccine is going to reduce how bad or how severe your case is. So, so there's a lot of variation in there, and we definitely see very low rates in the Medicaid population.
Did influenza testing decrease?
Ellyn Russo: Testing-wise, in terms of the amount of testing that was occurring for influenza, we did see a reduction, and this was similar to what was published by the CDC as well in terms of, you know, there weren't a lot of--there was not a lot of influenza testing at the tail end of last year's season. So once COVID started picking up in the February-March-April timeline, testing for influenza definitely decreased. It remained low throughout the summer when we don't typically see influenza in the northern hemisphere. And then it started to pick up again, so it wasn't completely low for the start of the season. It was a little bit lower, but one thing that outside of what we've studied also contributed to this is the testing, the diagnostic assays for testing for COVID and influenza were actually made into one test. What became available was the ability to test for different types of influenza and COVID all in one test which made it a lot easier to sort of, you know, check the box of getting the testing done for that. So that might be why we started to see rates come up a little bit in terms of testing and certainly, you know, in 2021, we don't have the 2021 data here in this article, but certainly testing, I think, probably remained elevated if I had to guess because of that reason of easier access to testing.
Did we avoid transmission due to COVID-19 precautions?
Ellyn Russo: We didn't have measures in our data set of, you know, the adherence to mask wearing. We didn’t try to actually look at, you know, geographic variation in terms of where we knew there were very strict stay-at-home or mask-wearing protocols. We didn't get to that level in this paper, but you know, the general premise is, well, if, you know, the tool we know prevents disease--which is vaccination--didn't change. It's not like it increased substantially. We weren't missing cases because testing was still occurring. You know, maybe it's at lower rates, yes, but it was still occurring. That kind of left us with, well, this points to it really being that transmission and the fact that flu is transmitted less frequently, less easily than COVID, so having an additional barrier to further prevent that transmission seems like it is most likely affecting the reason or the decrease in those rates. You know, as I sort of said earlier, flu was still circulating a little bit, so it's not that it was completely wiped out. But having that extra barrier, taking those extra precautions a little bit further or just being a little bit more cognizant of washing hands, of covering your mouth and your nose only helped, certainly only helped reduce flu in this case.5
"Across the entire population we studied, one in five people received a [flu] vaccination. That's pretty low."
Continued mask-wearing and vaccination for flu prevention
Grace Bello: Right. So basically the COVID-19 prevention helped with flu prevention. And so we're seeing that flu vaccines work and mask wearing works and social distancing.
Ellyn Russo: Vaccination against influenza has been and continues to be proven to be a very effective measure for preventing infection. Coverage remains suboptimal, so if we, you know, if we try to look at how many people should be vaccinated, we're not quite there on an annual basis, which has always been the sort of working hypothesis to why we still see flu circulating. Flu, also back to my theme of virology, it undergoes antigenic drift, so its genetic makeup is actually changing at a very rapid pace and that's also why it's hard to control and get rid of flu completely. But we know vaccination works and we know we're supposed to do it every single year because of those changes in the virus' genetic makeup, but we don't do it. So that is a risk, is that if we don't achieve optimal coverage through vaccination, you know, there still is infection happening. We know mask-wearing works. That is certainly proven effective for COVID. You know, all of our efforts to stay at home, to change our lives, to adapt to working from home and just very dire circumstances this past year, worked. We saw the reduction in cases. We certainly saw how we changed the curve for COVID-19. And it works. It's working, too, against flu.
"We certainly saw how we changed the curve for COVID-19. And it works. It's working, too, against flu."
Grace Bello: It's a cultural change, just like populations that were impacted by SARS then continued the mask-wearing even for prevention of colds and flu.
Ellyn Russo: Between vaccination and mask wearing or some of our changes around what's normal and the importance of doing it might change. It might just become a little bit easier to do.
Assessing specific patient populations
Grace Bello: Right. And to loop back around to the data that your team was looking at, it seems like it really is important to look at your specific patient population.
Ellyn Russo: Absolutely. Thanks for bringing that back up. You know, I think it's something I didn't really dive into is the importance of the healthcare system and the environment overall. There's a lot of players in that system, you know, right down from if you start at the patient level—what are my actions, what are my behaviors, you know, do I know how to get the flu shot? Do I know how to get the COVID shot? Do I know, you know, am I wearing my mask? Am I being cognizant of that? You know, up to the provider team that's taking care of you. So anytime you need something for your health, whoever it is, right, what are they doing, what resources do they have available to help you? Up to the hospital systems, all of these provider systems that we're seeing more and more of cross-collaboration between different types of providers, infectious disease providers being more involved in other aspects of care for people. So that's part of the environment. And then we think really broad, really system-wide level of who's paying for care and who's watching the overall population across state boundaries, across regional levels, you know, worldwide. You know, that's where entities like insurance companies, the Centers for Disease Control and Prevention and the World Health Organization, that's where their-- You know, every action that they're taking as well plays into how we ultimately see population health improving or not improving.
So when you think about the MedInsight database, when you think about what different data sources can do, they're providing different views at different times into different population aspects and dynamics. It's helpful information I think to come back to, you know, looking at, you know, the insured population which as I said, some people are not going to be needing healthcare because it turns out, you know, vaccination rates might be high so they don't need that healthcare and that's very important at that particular time because there's another outbreak of COVID or something going on. But it allows you to be more nimble and react to different situations and make decisions very quickly if you're able to see data at that level--and in real time.6 So more up to date than we've seen before with these types of data sources. That's huge. That's definitely contributing to how we think about even going forward. We always keep talking about 2020 is just so different from any other year we've studied. Milliman excels in looking at trends over time and kind of what's coming up next based on what's happened before. But, you know, how do we respond to the next pandemic, to the next, you know, even chronic disease. You know, what's coming next in terms of what's affecting people? How do we adapt to it? How do we stay on top of using data to help us learn and sort of react and be proactive about what resources are there, how we can help people even more than we have before when we think about health and healthcare overall?
"How do we respond to the next pandemic, to the next chronic disease?"
Grace Bello: Right. And what's interesting is you brought up, I had brought up the individual level of the vaccination and the mask wearing, but you brought up the broader level of how can, you know, insurers and healthcare systems, et cetera support people staying well?
Ellyn Russo: I think there's long lasting implications for ease of vaccine, ease of anti-viral treatment early on for flu, those sorts of themes I think will continue on. And who's recognizing them? Certainly doctors and nurses and everyone taking care of people. But it becomes a lot easier when you've also got leaders and administrators of hospital systems and insurance companies also sort of seeing the picture, too, and helping to make it even easier. You know, everyone has different touchpoints, whether you're a PCP, you're a primary care physician, or your insurance company when you're talking about coverage or kind of understanding where you can go for certain things. Everyone's got a touchpoint to help make that as easy as possible and as seamless as possible for a patient, for an individual.
Grace Bello: Thank you, Ellyn, for joining us. You can read her team’s paper, The Mysterious Case of the Missing Influenza Season, on Milliman.com. To listen to other episodes of our podcast, visit us at Milliman.com or you can find us on iTunes, Google Play, Spotify, or Stitcher. See you next time.
1See the CDC’s resource on symptoms of COVID-19: https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/symptoms.html, updated on 2/22/21.
2"While the virus that causes COVID-19 and flu viruses are thought to spread in similar ways, the virus that causes COVID-19 is generally more contagious than flu viruses." https://www.cdc.gov/flu/symptoms/flu-vs-covid19.htm
3This is meant to serve as one example of a possible contributing factor for inequities in the setting of infection from influenza.
4Additional analysis is needed to confirm or refute this statement as the analysis described was not designed to account for differences in the age, sex, or geographic location of the study populations.
5Furthermore, venues such as schools, corporate offices, restaurants, and more were forced to close at least temporarily. The prevention of gatherings also contributed to the low flu transmission rate.
6Milliman’s MedInsight Emerging Experience research data is near-real-time, providing insights with minimal lag time.