Model Structure

A previously published population-based, multi-cohort, Markov-type model [19, 20] was adapted to estimate the lifetime impact of the RSVpreF vaccine on public health and economic outcomes in a hypothetical Japanese population aged ≥ 60 years. Health and economic outcomes were projected for the model population from model entry through the end of life (i.e., maximum age of 99 years), based on age, comorbidity status, RSV incidence rates, general population mortality rates, RSV case-fatality rates (CFRs), vaccination status, and time since vaccination, accounting for monthly variation in the timing of vaccination and rates of RSV illness.

Health outcomes comprised RSV cases, stratified by care setting (hospitalization, emergency department [ED], and outpatient), as well as RSV-related deaths. The likelihood of RSV-related death and non-RSV-related death depended on the age and comorbidity profile. The indirect effects of declining RSV infection rates due to the herd effect from increased RSV vaccination were not considered. The economic outcomes assessed in this study were medical costs, vaccination costs, and costs related to productivity losses. Cost-effectiveness was based on incremental cost-effectiveness ratios (ICERs) of a single dose of the RSVpreF vaccination versus no vaccination.

Model InputsTarget Population

Japanese adults aged ≥ 60 years (N = 43,732,000), stratified into four age groups (60–64, 65–74, 75–84, and 85 years or older), were included in the model (Table 1). The population size was obtained from the 2023 Population Estimates in Japan [21]. The population within each age group was further stratified by the comorbidity profile, classified as high risk or low risk, based on the presence or absence of chronic or immunocompromising medical conditions. The definitions of high-risk conditions and the detailed estimation procedure are included in the section Supplementary methods.

Table 1 Model input values for base-case analysisRSV Incidence Rates and Mortality

Annual RSV incidence rates, adjusted for age and care settings, were derived from a previous health and economic evaluation of the adjuvanted prefusion F protein vaccine (RSVpreF3 OA, GSK, Philadelphia, PA, USA [22, 23]), which was based on a prospective study in Japan [24]. Detailed calculation procedures are described in Table S1.

Since RSV seasonality data for adults were not available, data for the pediatric population were used. Moreover, to avoid the impact of infection control measures during the Coronavirus Disease 2019 (COVID-19) pandemic on RSV incidence rates, the average seasonality data from 2017 to 2019 were employed for the base-case analysis (Table S2) [25]. The most recent pediatric RSV seasonality data from 2024 were used for the scenario analysis (Table S3) [25, 26].

The CFR for RSV hospitalization was based on a multicenter prospective cohort study in Japan [27]. No excess mortality was assumed for the RSV ED visit and the RSV outpatient visit. General population mortality rates across different age groups were taken from the Japanese life tables [28] and adjusted by Population Estimates [21]. The allocation of the mortality rate by calendar month was based on data from Vital Statistics 2023 (Table S2) [29].

Effectiveness of RSVpreF vaccine

Model inputs for the effectiveness of the RSVpreF vaccine were based on the results of seasons 1 and 2 of the RENOIR trial [15, 16, 30]. The efficacy endpoints in the RENOIR trial were defined based on symptoms rather than care settings (hospitalization, ED visit, and outpatient visit). Therefore, vaccine effectiveness (VE) against RSV hospitalization and RSV ED visit was approximated using the RENOIR efficacy against medically attended RSV LRTD with three or more symptoms. Similarly, VE against RSV outpatient visit was approximated using the RENOIR efficacy against medically attended RSV ARI. Initial VE against RSV hospitalization or RSV ED visit was assumed to be 84.6% (95% CI, 32.0–98.3), lasting 7 months (season 1), then linearly declining to 72.0% (95% CI, 33.4–89.8) by Month 16 (season 2) [31]. The rate of linear decline observed from season 1 to season 2 was then assumed to persist through Month 41 (season 4) to reach 0% from Month 42 and beyond (Fig. S1a) [32]. Similarly, initial VE against RSV outpatient visit was assumed to be 65.1% (95% CI 35.9–82.0), persisting for 7 months, then linearly decreasing to 47.0% (95% CI 22.7–64.1) by Month 16 (season 2) and assumed a continued linear decline to reach 0% from month 41 and beyond (Fig. S1a) [16]. VE was assumed to be similar across age groups and comorbidity risk profiles, consistent with the RENOIR trial [15, 16].

In the scenario analyses, both conservative and optimistic VE waning were considered. In the conservative scenario, the rate of linear decline was assumed to persist through Month 30 (end of season 3) and was truncated at Month 31 (i.e., assumed to reach 0% by Month 31) (Fig. S1b). In the optimistic scenario, the observed slope between Months 7 and 16 was assumed to persist until effectiveness reached 0% (for VE RSV hospitalization/RSV ED, Month 70) (Fig. S1c).

Vaccination Strategy

The vaccine coverage was assumed to be 50% based on the influenza vaccination rate in adults aged ≥ 65 years in 2018 and 2019 before the COVID-19 pandemic [33]. Monthly vaccine uptake was assumed to begin 2 months earlier than the RSV seasonal distributions, ensuring that individuals are vaccinated and protected ahead of the RSV season (Table S2) [22, 26].

Medical Costs and Costs Related to Productivity Losses

RSV-attributable medical costs for episodes of RSV hospitalization, RSV ED visits, and RSV outpatient visits were based on a retrospective study using a large Japanese claims database [34]. In the Japanese healthcare system, vaccines are not reimbursed, and their prices are not officially determined, nor are the wholesale prices disclosed. Therefore, the vaccine cost was assumed based on the previous cost-effectiveness analysis of the RSVpreF vaccine for maternal use in Japan [35], which was below the cost-effectiveness threshold of Japanese yen (JPY) 5 million per quality-adjusted life-year (QALY). Given that the RSVpreF vaccine has dual indications (maternal use for prevention of infant RSV disease and use in adults aged ≥ 60 years), the price was not varied among these indications. The cost of vaccine administration was calculated based on a fee-for-service price list (2024) by the Ministry of Health, Labour and Welfare, including costs for initial consultation, procedures, and biologics premiums (Table 1) [36].

The costs of morbidity- and mortality-related productivity losses due to RSV were estimated based on the number of work-loss days due to RSV-related diseases, workforce participation rates, and the average daily wages for both patients and caregivers (Table 1). The patients’ work-loss days for the RSV hospitalization were estimated based on a retrospective study in Japan [11], and those for the RSV ED visit and the RSV outpatient visit were estimated based on an economic model that calculated costs for RSV-related diseases in the United States [37]. The workforce participation rates and daily wages for both patients and caregivers were estimated from the Labor Force Survey 2023 [38] and the Basic Survey on Wage Structure 2023 [39].

Utilities and Disutilities

Age-specific utility values were based on the Japanese Population Norms of the 5-level EuroQol 5-Dimension (EQ-5D-5L) and the Health Utilities Index Mark 3, as reported in a study by Shiroiwa et al. [40] and adjusted by the Population Estimate 2023 [21]. The disutility values in the base-case analysis were based on a Japanese time trade-off study [41]. Detailed disutility information is presented in Supplementary methods.

Ethical Approval

This study was based on previously published data and does not contain any new studies that required human participants or animals.

Model AnalysesBase-Case Analysis

The base-case analysis compared a single dose of the RSVpreF vaccine with no vaccination from the Japanese payer and societal perspectives, employing a lifetime horizon with a discount rate of 2% [42]. In this study, a cost-effectiveness threshold of JPY 5 million per QALY was applied [43, 44].

Scenario and Sensitivity Analyses

Scenario analyses were performed for various age groups, narrowing down the age ranges to ≥ 65 years, ≥ 70 years, ≥ 75 years, ≥ 80 years, ≥ 85 years, and ≥ 90 years. Additionally, analyses were conducted for three subgroups, including adults aged 65 years and those aged 60–64 years at high risk, adults aged ≥ 65 years and those aged 60–64 years at high risk, and adults aged ≥ 75 years and those aged 60–74 years at high risk. The first two subgroups align with the population covered by the Japanese NIP for other infectious diseases such as pneumococcal diseases, herpes zoster, influenza, and COVID-19 [45]. The last subgroup is recommended for RSV vaccination in the United States [46] and Canada [47]. Alternative parameter values (such as the proportion of patients at high risk, RSV incidence rates, RSV hospitalization-related CFR, disutilities, monthly distribution of RSV cases and vaccine uptake, and VE) were also considered (Tables S3 and S4).

To identify model drivers and examine key areas of uncertainty, a one-way deterministic sensitivity analysis was conducted considering changes (± 25% of the base-case value) in selected variables: disease incidence rates (RSV hospitalization, RSV ED visit, and RSV outpatient visit), costs, VE, utilities and disutilities, and RSV hospitalization-related CFR.

A probabilistic sensitivity analysis (PSA) with 1,000 iterations was performed to explore uncertainties around key model parameters. The distribution used for the PSA is described in detail in Table S6. Typical probability distributions were used in the analyses, following the guidance of Briggs et al. [48].

Threshold Analysis

A threshold analysis was conducted, which determined the vaccine price excluding administration costs, with the ICER being just below the cost-effectiveness threshold of JPY 5 million/QALY.