Impact of flu vaccines on hospital burden

[Music] Welcome to Science Sessions, the podcast of the proceedings of the National Academy of Sciences, where we connect you with Academy members, researchers, and policymakers. Join us as we explore the stories behind the science. I'm Paul Gabrielson. Seasonal influenza produces a surge of hospitalizations and healthcare burdens that public health officials try to mitigate through distribution of flu vaccines. But with vaccine coverage far from complete and amid increased scrutiny of vaccine effectiveness, how well are flu vaccines reducing seasonal healthcare burdens? In a recent PNAS study, Lauren Ansel Myers of the University of Texas at Austin and colleagues studied the 2022-2023 flu season to understand the effectiveness of vaccines during a particularly severe flu season. They found that vaccination helped prevent hospitalization in people who received vaccines and helped them prevent transmission of flu to vulnerable populations. Improving vaccination rates among a critical demographic group, the authors say, could further reduce the public health burden during flu season. Lauren, let's start with some background. Can you introduce us to the health and hospitalization burden of seasonal influenza? For many people, seasonal influenza may feel like routine or no big deal, but it really is a major driver of seasonal illness every single year across the globe and in the United States. And in the US, typically, we see hundreds of thousands of people end up in the hospitals with flu, which causes substantial strain in our healthcare systems. And it varies from season to season. How do we define vaccine effectiveness? Vaccine effectiveness, we refer to it as VEE, is basically a way of saying how much lower is your risk of getting sick with flu or seriously sick with flu if you take a vaccine versus if you don't take a vaccine. In our study, we really think about two different kinds of risk. So if you take a vaccine, how much less likely are you to actually get infected in the first place than somebody who doesn't take a vaccine? We also think about vaccine effectiveness against severe disease. So if you take a vaccine, how much less likely is it that you will end up in the hospital with severe influenza versus if you don't take a vaccine? The infection blocking part is actually a really important part of the equation because every infection you prevent also prevents the chance that somebody would have spread flu on to somebody else who might end up in the hospital. Why did you choose the 2022-2023 year for studying influenza vaccine effectiveness? It really felt like an unusual season where it was worth taking the time to understand how big the burden of flu was and the extent to which vaccines may have reduced that burden. And why was it a weird year? Well, the preceding two seasons we saw almost no flu because we were in the heart of the COVID pandemic and we were doing things like wearing face masks and staying home and closing schools. All those things that were designed to slow COVID transmission also slowed and essentially stopped flu transmission. We got to the 2022-23 flu season. We had lifted all these restrictive measures. There wasn't a lot of immunity in the population from the preceding seasons and so there was the opportunity for flu to spread earlier, more quickly, more rampantly than it had. Certainly in the preceding years and possibly even before the COVID pandemic started. The dominant subtype that was spreading was the H3 and 2 subtype and that is one of the two subtypes that spread regularly that tends to cause higher hospitalization rights. This study was funded by a grant from the CDC and was inspired specifically by colleagues at the CDC asking whether we could help to quantify the benefit of vaccines in the 2022-2023 season. So a lot of the analysis we did very quickly to answer a pressing public health question. How did you estimate the number of hospitalizations prevented by vaccination? We built a complicated model that simulates flu transmission across the United States and we also built models for each of the 50 different United States. We calibrated the model to real hospitalization data from the 22-23 season and then we ran a bunch of what if experiments using the model. For example, if nobody in the country had gotten vaccinated at all or what if only one age group had gotten vaccinated, then we used all these different experiments or counterfactuals. We compared the results of them to estimate how many hospitalizations were actually prevented by the vaccines or how many more hospitalizations could have been prevented if more people had been vaccinated. And the main result is pretty striking. We estimate using this approach that flu vaccinations prevented about 70,000 hospitalizations during that 2022-23 season. And about half of that benefit came from blocking infections, reducing transmission, so not just from reducing severity once you become infected. How did people over 65 years of age benefit from vaccination of young adults? That's the age group that ends up getting hospitalized at the highest rate during influenza seasons. And one of the most surprising things we found was that vaccination in young adults and even children ended up indirectly protecting a lot of older adults. So specifically, vaccinating people between ages 18 and 49 had a big impact. So that is actually the age group that has the lowest vaccination rates. And that is also the age group that benefits the least personally from protection because they're not at very high risks. But it turns out by vaccinating them, we protect these other age groups. And why is that? Well, young adults make up a big proportion of the population. And young adults mix a lot socially with other age groups and within their age group at work across generations, etc. And if they're vaccinated, those vaccines then can have a big impact in breaking chains of transmission. So what we found was almost half of the hospitalizations that were prevented in people over age 65 actually came from vaccinating people in younger age groups, not from vaccinating those older people themselves. In 2020, the U.S. Department of Health and Human Services set a goal of trying to reach 70% vaccination coverage for influenza by 2030 as part of something called the Healthy People Initiative. So we ran simulations that helped us to quantify what things might look like if we actually achieved that aspirational 70% target. And the model suggests that increasing vaccination rates in particular age groups and across all age groups really could continue to lower the burden of illness even more and would be worthwhile. What are the takeaways from this study for future influenza seasons? The first is that blue vaccines have a large and measurable public health benefit. This is a year where vaccines were only sort of moderately effective, but we estimated that tens of thousands of hospitalizations were prevented and that U.S. states with higher vaccination coverage generally experienced lower hospitalization burden. The second interesting take home is that young adults are a surprisingly high-leveraged group. They have the lowest vaccination rates, so there's a lot of room for increasing vaccination rates in that group, but they're also the ones that mix the most and drive a lot of flu transmission. So boosting coverage or taking steps to try to encourage vaccination in that age group could really have an outside benefit. The third take home is that we really need better data on vaccine effectiveness and specifically on how well vaccines protect against infection. And we need better data more generally on all the things that go into our model including the level of immunity, how quickly immunity wanes, and how people come in contact with each other during the flu season. What are the caveats or limitations of the study? We had really limited data on how well the vaccine actually prevented infection in 2022, 2023. In fact, we didn't have any direct data from that season. And so we had to rely on extrapolating from previous seasons where H3N2 was the dominant subtype in which vaccine effectiveness against infection had been measured. And that's why we took the extra effort to measure how sensitive our results were to our assumptions about vaccine effectiveness against infection. Secondly, like any model, ours depends on a lot of simplifying assumptions. We focused exclusively on the H3N2 subtype. That was the dominant subtype, but it wasn't the only flu virus spreading that season. We represent immunity in our population at a population level rather than tracking individual immune history. And when you compare the model predictions to what really happened in 2022, 2023, we don't capture the exact timing and the shape of the epidemic peak. And some of these choices that we make, they make the model tractable, but they smooth over some of the real world complexity. We had a particularly hard time pinning down some key parameters that probably changed because of the COVID pandemic. Our social contact patterns changed dramatically during COVID, but a lot of the data we have on social mixing patterns came from the pre-COVID era. But even with those limitations, our findings are consistent with estimates from prior season, than even an estimate that came out from the CDC of the 2022-2023 season. So we feel pretty confident, at least in the qualitative results and the take-home messages of our study. Thanks for tuning into science sessions. You can subscribe to science sessions on iTunes, Spotify, or wherever you get your podcasts. If you like this episode, please consider leaving a review. and helping us spread the word.