This study adds to the recent literature on the characterization of RSV burden, with the most up-to-date clinical data. It estimates the impact of universal infant immunization with nirsevimab on RSV-related health outcomes, all-cause LRTD hospitalizations, expenditures and QALY loss, based on a previously published static decision-analytic model adapted for the UK setting [19, 20]. The model predicts that infants currently ineligible for palivizumab (term and preterm) account for 98% of the hospitalization burden, despite having a lower risk of severe events compared to those infants eligible to receive palivizumab. These results demonstrate that immunization of otherwise healthy infants born at or close to term (approximately 96% of cohort) can avoid the majority of poor health outcomes and expenditures associated with RSV, despite premature infants with CHD or CLD being at greater risk of serious outcomes.

The results show that immunizing all infants with nirsevimab could reduce RSV-related LRTDs by 60% and associated costs by 66%. The estimated reduction in all-cause LRTD hospitalizations (excluding RSV events) was 21%; this additional benefit may be key in relieving pressure on the health system during winter. Immunization of infants with nirsevimab born before the RSV season can avoid 134,329 total RSV LRTDs (64%). For infants born during the RSV season, immunization with nirsevimab is estimated to avoid 74,293 RSV LRTDs (36%). This distribution of RSV LRTDs occurring in infants born during the RSV season versus those born before the RSV season aligns with previous estimates [8, 37, 38]. Infants born before the RSV season make up the majority of the annual birth cohort (i.e., 7 months of the year; 60%) [29, 30], explaining the higher absolute number of health events and costs versus those born during the season. Similarly, as the majority of the infant population (96%) is born full-term, most health events occur among the term Infant population, born both during and before of the RSV season. This highlights the need to immunize all infants to reduce the overall RSV burden, including those born full-term and those born before the RSV season. The sensitivity analysis lends additional credibility to these results which show that parameters related to term infants have the strongest impact on the model results.

RSV seasonality informs a critical component of any immunization strategy. A minimum 5-month duration of protection of nirsevimab suggests that the optimal immunization timing for infants born before the RSV season is October, thus providing protection throughout the RSV season. This reinforces health equity by offering optimal coverage for infants born before the season, while also protecting the youngest and most vulnerable infants born during the season. However, immunizing a substantial portion of the infant population in October (approximately 62% of the UK’s annual infant birth cohort) could be logistically difficult and reduce real-world coverage. Therefore, the strategy evaluated in our study aligns immunization with nirsevimab in the term infant population with the UK NIP. The difference between the October catch-up and the NIP strategy is marginal in terms of reduction in hospitalizations and direct costs for the entire infant population, considering most infants are born at term (Table S7).

By integrating nirsevimab prophylaxis within the UK NIP, the health system burden of immunization for infants born pre-season is spread out across July, August, and September, instead of concentrating all immunizations in October, thereby maximizing coverage, particularly in infants less vulnerable to severe RSV-related health events. If the 150-day nirsevimab duration of protection is strictly assumed, then integrating nirsevimab within the UK NIP could leave a small portion of older infants with limited protection toward the end of the season when RSV is still in circulation. However, these infants will be less vulnerable to severe RSV-related MA LRTDs as they are older in age; plus, there is evidence that some infants may experience residual protection beyond 150 days.

Another option considered by the JCVI involves immunizing all infants at birth [15]. Existing and expected data suggesting prolonged protection with nirsevimab[23] could render this strategy pragmatic. Sufficient evidence to properly model this duration and waning of protection is not fully available at this time as a strong correlation between protection with respect to serum concentration has not yet been established. A scenario in which all infants receive administration with nirsevimab at birth was tested under the assumption of a linear decay in protection between months 6 and 12. The outcomes of this strategy were similar to the NIP strategy. A final scenario tested was the immunization of the infants born during the season only. This strategy would result in most infants experiencing their first RSV season unprotected, with only a third of health events avoided compared to all-infant protection.

The primary strength of our model is its granularity compared to existing studies in infant RSV [17, 20, 39]. In addition to considering differentiated risks of RSV-related health events by infant subpopulation, the use of hospitalization rates by wGA (from [8]) allows for stratification of subpopulations by monthly age. Our model can therefore evaluate the potential impact of nirsevimab independently for infants within each subpopulation, in addition to those born during versus before the RSV season. Disaggregating the population granularly allows multiple strategies to be evaluated and trade-offs assessed, weighing up pragmatism versus outcomes to determine the most impactful intervention and inform the decision-making process. Another strength of our model is that the analysis is based on the most up-to-date data for the risk of hospitalization due to RSV among infants in the UK, collected shortly before the COVID-19 pandemic and so not confounded by the influence of social distancing on circulating respiratory viruses [40, 41].

The results of our evaluation can be compared with recent studies assessing the impact of immunization on RSV LRTDs [16, 17, 20, 42]. The most recent Hodgson study uses a dynamic transmission model, leveraging an adapted version of the susceptible-exposed-symptomatic-recovered model structure which includes an asymptomatic state as well as the potential effects of maternal-protective antibodies [20]. The Hodgson study estimates fewer avoided RSV LRTDs associated with nirsevimab prophylaxis compared to the results of this analysis. While differences in model structures between the Hodgson and current study can explain some of the differences in results, the primary driver of the discrepancy is related to several other dimensions. First, Hodgson et al. assumed sterilizing immunity induced by nirsevimab, causing an increase of susceptible infants over the second RSV season, and therefore an increase in the number of cases in the 1–4 years age group. This assumption on the mechanism of action of nirsevimab is incorrect, as RSV exposure in nirsevimab-immunized infants is accompanied by subclinical manifestations of disease, indicating that sterilizing immunity is not induced by nirsevimab [43]. This is backed by new data showing no increase in severity in the second season of nirsevimab-immunized infants [44]. The Hodgson model also underestimated key parameters like RSV hospitalization costs, where it is assumed that RSV-related hospital costs are the same across all age groups from 0 to 15 years, and that this is equivalent to the cost of hospitalized acute bronchiolitis in all individuals aged < 18 years. Given RSV disease severity is known to be greatest in infants aged under 1 year and decreases with age, Hodgson’s approach likely underestimates the cost of hospitalization in infants aged under 1 year. Furthermore, QALY loss inputs used in the Hodgson 2024 model are outdated and not specific to the target age group for immunization. Our model utilizes more recent, relevant and comprehensive QALY loss data from Mao et al. [45], a prospective observational cohort study conducted in the UK, Spain, Finland, and the Netherlands, which estimates QALY losses in infants aged under 1 year with a confirmed RSV case during the 2017–18, 2018–19 or 2019–20 season. Hodgson’s model also failed to account for the proven benefits of nirsevimab in reducing all-cause LRTD hospitalizations [22, 23, 32], and the direct impact on infants of replacing palivizumab with nirsevimab [24]. Finally, substantial evidence indicates nirsevimab will have greater real-world impact than maternal vaccination, through the protection of all infants, regardless of gestational age, with sustained efficacy over the RSV season and timely administration to allow protection throughout the RSV season. Early real-world evidence shows robust and consistent outcomes with a pooled effectiveness in preventing RSV LRTDs of approximately 84.4% (95% CI 76.8–90.0) [46]. Using Hodgson’s assumptions, our model results in 4178 fewer hospitalizations avoided, £69 million fewer direct costs avoided, and 1678 fewer QALY saved versus SoP, suggesting significant underestimation of the impact of nirsevimab in reducing RSV-related burden compared to the results presented here.

Furthermore, a formal model comparison was conducted to ensure the cross-validity of our model in accordance with guidelines for multi-model comparisons executed by the Respiratory Syncytial Virus Consortium in Europe (RESCEU) [47]. This study aimed to compare the outcomes of different available, model-based, analytical approaches designed to estimate the cost-effectiveness of RSV prevention in infancy and pregnancy, using a standardized set of input parameters across three static models—one produced by researchers at Antwerp University, the Novavax model, and the Sanofi-generated model presented here. Our model produced identical results to those generated with the same inputs in the Antwerp University model, for the overall infant population and per age group, supporting the validity of the results presented here.

Although a key strength of our model centers around the level of granularity available in modeling infant birth cohorts, one key limitation in this analysis is the lack of some granular source data. In these situations, assumptions were applied to fit the available data. For example, the risk of A&E visits was assumed to be consistent across all subpopulations. Estimates for hospitalizations are also unavailable by month of age; therefore, the monthly trend for the overall infant population from Reeves and colleagues [8] is assumed to be applicable to all subpopulations.

Similarly, data on the route into admission of infants who are hospitalized is largely unavailable, particularly with respect to infants who are first admitted to A&E. As the only available source to inform A&E visits presents incidence rates only for infants who do not go on to be hospitalized [48], this analysis treats these health events as mutually exclusive and therefore cannot capture any impact the potential reduction in A&E visits has on the number of hospitalizations for this population of infants. For the same reason, the model also does not include NHS contacts which do not lead to hospitalization.

Assumptions for wheezing similar to those of a recent cost-effectiveness analysis on RSV prevention in the EU [49] are applied in our base case; an estimated avoided 12,493 cases, £3.5 million in direct costs and 476 QALYs saved were associated with wheezing over 3 years as a result of universal nirsevimab immunization. However, a key assumption in the model is the indirect impact of nirsevimab on the incidence of wheezing as a result of prevented hospitalizations. While prevention efficacy against RSV disease is proven, prevention efficacy against sequelae was not assessed in clinical trials and remains a point of conjecture.

Finally, the model adopts a static structure from a payer perspective, meaning potential indirect effects on the transmission of RSV are not captured. While no evidence shows an impact of nirsevimab on infection susceptibility, the potential effects of RSV antibodies on viral shedding, and therefore transmission, should be explored in order to further characterize the indirect effects of nirsevimab [50]. This analysis does not consider the lifetime lost earnings due to an infant RSV death or any other societal impact. This narrow, direct NHS-focused perspective may not fully capture the overall benefits of immunization; the impact of introducing nirsevimab from a societal perspective should be considered by decision-makers in particular.