What the Two-Season Evidence Actually Shows
The two-season evidence on adult RSV vaccination addresses a narrow but essential question: did one dose reduce RSV-associated hospitalization among adults aged 60 years or older, and what does the resulting percentage actually mean? Reading the answer well requires more than repeating the headline estimate. The study population, comparison group, vaccination window, statistical conversion, and uncertainty determine what readers can responsibly conclude about protection against serious disease.
The study starts with hospitalized adults, not the whole population
The evidence comes from the Investigating Respiratory Viruses in the Acutely Ill (IVY) Network, a multicenter surveillance network operating at 26 hospitals in 20 US states. It included adults aged 60 years or older who were hospitalized with acute respiratory illness during one of two periods: October 1, 2023, through March 31, 2024, or October 1, 2024, through April 30, 2025. To be included, patients needed respiratory-virus testing close to illness onset and hospital admission. Their nasal swabs were also tested at a central laboratory for RSV, RSV subtype, SARS-CoV-2, influenza, and human metapneumovirus.
The design produced two groups. Patients who tested positive for RSV alone became the case group. Patients who tested negative for RSV, SARS-CoV-2, and influenza became controls. Of 6,958 participants, 821 were RSV-positive cases and 6,137 were controls. The controls were not a random sample of all older adults. They were people sick enough to be hospitalized with an acute respiratory illness but whose testing did not identify the viruses used to define the comparison.
That choice is important. Vaccinated and unvaccinated people can differ in ways that affect both their chance of being vaccinated and their chance of being hospitalized. They may differ in healthcare access, preventive habits, chronic disease, or willingness to seek care. A simple comparison of everyone who received a vaccine with everyone who did not would therefore mix the vaccine question with many other differences. By comparing RSV-positive admissions with similar respiratory-illness admissions that were negative for the target viruses, the researchers tried to make the groups more alike in their reason for coming to the hospital.
The controls also help hold several background conditions closer together. Both groups had crossed the threshold of having an acute respiratory illness serious enough to prompt hospital evaluation, and both were observed during the same broad respiratory-virus seasons. This does not remove every difference: a person with a non-RSV illness may still differ from a person with RSV in symptoms, timing, exposure, or care-seeking behavior. But it is a more relevant comparator than a healthy person who never needed medical attention. The design asks whether vaccination is especially underrepresented among the admissions attributable to RSV, rather than simply asking whether vaccinated people look healthier in general.
The exclusions served a specific purpose. An RSV-positive patient who was simultaneously positive for SARS-CoV-2 or influenza was excluded from the case analysis because the RSV vaccine would not be expected to prevent hospitalization caused by those other viruses. RSV-negative controls with SARS-CoV-2 or influenza were excluded because vaccination behaviors can be correlated across vaccine-preventable respiratory infections. Leaving those patients in the control group could distort the comparison.
The resulting question is more precise than “Does the vaccine work?” It is: among adults at least 60 years old who were hospitalized with acute respiratory illness, were patients with RSV less likely to have been vaccinated than comparable patients without RSV, after accounting for important differences? That is a patient-relevant question, but it is still narrower than measuring every infection prevented in the community.
What counted as vaccination
The study defined vaccination carefully. A vaccinated participant had received one RSV vaccine dose at least 14 days before illness began. The unvaccinated group included people with no prior RSV vaccination or people vaccinated after illness onset. Participants who had received two doses were excluded, as were those vaccinated fewer than 14 days before illness onset.
The 14-day boundary protects the comparison from a basic timing problem. A dose given immediately before symptoms cannot reasonably be expected to provide the full protection being measured, and illness may already have begun before vaccination. Excluding that interval does not prove that protection starts exactly on day 14. It creates a consistent study definition that separates the exposure from the illness being evaluated.
Vaccination status was not taken from one source alone. Researchers used electronic health records, state or city vaccine registries, and self-report in a hierarchical verification process. Even so, the study remains observational: people were not randomly assigned to receive a vaccine. The vaccinated participants may have differed from the unvaccinated participants in ways the available data did not fully capture.
In the final sample, 63 of the 821 RSV cases had been vaccinated, compared with 966 of the 6,137 controls. Those raw proportions—7.7% of cases and 15.7% of controls—show the pattern the model analyzed, but they are not themselves the vaccine-effectiveness estimate. They do not account for age, geography, calendar time, or other differences, and they do not represent the vaccination rate among all older adults in the country. Keeping the numerator and denominator in view is a useful guard against treating a hospital study as if it were a population survey.
The analysis compared the odds of vaccination among cases and controls. It adjusted for age, sex, race and ethnicity, geographic region, and calendar month and year. Those calendar variables matter because RSV activity and vaccination uptake both change over time and across places. The adjusted odds ratio was then converted into an estimated vaccine effectiveness using the formula:
Vaccine effectiveness = (1 − adjusted odds ratio) × 100%
In plain language, the analysis looked for whether vaccination was less common among people hospitalized with RSV than among people hospitalized with a similar acute respiratory illness without RSV, then adjusted the comparison for the listed factors. The percentage is therefore a relative estimate produced by this study design. It is not the percentage of vaccinated individuals who were guaranteed to avoid hospitalization, and it is not a direct count of infections prevented in the entire population.
The comparison can be summarized as follows:
The design does not make the groups identical, but it makes the central comparison more disciplined than a broad vaccinated-versus-unvaccinated count. Its strength is that it focuses on a concrete clinical event and uses testing to define the outcome. Its limitation is that the estimate still depends on who entered the hospitals, how accurately exposures and outcomes were measured, and how completely the adjustment variables captured meaningful differences.
What the headline estimate means
Across the two RSV seasons, one dose was associated with an estimated 58% effectiveness against RSV-associated hospitalization. The 95% confidence interval ran from 45% to 68%, and the median time from vaccination to the relevant illness was 223 days, with a wide range across participants. The most defensible reading is that, in this hospitalized surveillance population and under this test-negative design, vaccination was associated with substantially lower odds of RSV-associated hospitalization than the comparison pattern would predict without vaccination.
That sentence contains several limits worth keeping visible. “Two seasons” means the estimate combines the 2023–2024 and 2024–2025 surveillance periods; it does not mean every future season will produce the same result. “Adults aged 60 years or older” describes the enrolled age range; it does not directly answer how well vaccination works in younger adults. “Hospitalization” is the measured endpoint; it does not stand for all infections or all respiratory symptoms. “Estimated effectiveness” identifies an adjusted observational association, not a randomized guarantee.
The confidence interval is part of the result, not an optional footnote. It describes the range of values compatible with the data and the model under the study’s assumptions. The estimate is encouraging because the entire interval is above zero, but it is not a promise that the true value is exactly 58% in every setting. A different season, population, product mix, vaccination pattern, or circulating-virus context could produce a different estimate.
Nor does 58% mean that a vaccinated person’s personal risk falls by 58 percentage points. Vaccine effectiveness is relative. To understand absolute benefit, one also needs the baseline chance of the outcome. If two people have different underlying risks of RSV hospitalization, the same relative effectiveness can translate into different numbers of hospitalizations avoided. That distinction is central to risk-stratified vaccination: the expected value of a dose depends on both how likely severe disease is and how much protection the dose provides.
The study also examined what happened after hospitalization. Estimated effectiveness was 73% against acute respiratory failure and 73% against acute organ failure. It was 67% against intensive care unit admission and 72% against the combined outcome of invasive mechanical ventilation or death. These results suggest that the association was not limited to milder admissions; vaccination was also associated with less progression to several severe in-hospital outcomes.
Those outcomes are related but not interchangeable. Hospitalization asks whether the patient entered the hospital with RSV-associated illness. Respiratory failure, organ failure, ICU admission, and invasive ventilation or death describe increasingly serious consequences during the hospital stay. A vaccine-effectiveness estimate for one endpoint should not be silently substituted for an estimate for another. The severe-outcome findings strengthen the clinical importance of the pattern, but they do not convert the study into a direct measurement of mortality prevented across the community.
The last estimate needs especially careful handling. Its confidence interval extended from 7% to 91%, much wider than the headline interval. That width reflects limited precision for a less frequent and more serious composite outcome. It does not invalidate the result, but it means the exact magnitude is less certain. The same principle applies throughout the analysis: a clinically important direction of effect and a precisely measured effect are not the same thing.
What the estimate cannot establish
First, this study estimates effectiveness against RSV-associated hospitalization, not protection against every RSV infection. A person could be infected without being hospitalized, and that event would not appear as the primary outcome unless they entered the study through the defined hospital pathway. The findings also do not show that vaccination prevents all severe respiratory illness from every virus. The controls’ virus testing and the study exclusions were designed precisely to keep the outcome focused on RSV.
Second, the study does not have the causal certainty of a randomized clinical trial. The authors note that only 15.7% of controls were vaccinated. Early adopters of a newly available vaccine may differ from the broader eligible population, including in preventive behaviors that were not measured. Mask use, healthcare-seeking patterns, access to vaccination, and other differences could leave residual confounding even after statistical adjustment. The analysis also did not account for a recent prior RSV infection, which might provide some protection independently of vaccination.
Third, the study does not supply a personal forecast. It does not say what will happen to a particular adult with a particular combination of age, immune status, heart disease, living situation, and prior infection. Nor does it establish that a single effectiveness estimate applies equally to all products, all age groups, or all future seasons. It enrolled adults aged 60 years or older; it therefore cannot directly quantify effectiveness for younger adults covered by a later policy decision.
These limits do not cancel the observed benefit. They define its proper scale. The study is strong evidence that one dose provided meaningful protection against RSV-associated hospitalization and several severe outcomes across two seasons in a large, geographically diverse hospitalized cohort. It is not evidence that protection is identical for every person, that hospitalization risk is the same before vaccination, or that the current one-dose policy has already answered every question about duration and repeat doses.
The result is best treated as a well-measured piece of an evolving decision, not as the final word. It supports confidence that adult RSV vaccination can prevent serious disease in the studied population while leaving room for population differences, unmeasured factors, future seasons, and later evidence to change the estimate. That combination—meaningful benefit with bounded certainty—is exactly what makes subgroup analysis necessary.
The next layer of the evidence is therefore about variation rather than the headline average. The study did not find one uniform response across its participants. Effectiveness was lower in adults with moderate or severe immune compromise and in the analyzed group with cardiovascular disease, while the chronic-lung comparison looked different. Understanding those differences is essential: a risk-stratified recommendation is meaningful only if readers can see why the same dose may have a different expected value across people.