Modeling the Cost-Effectiveness of the Next-Generation COVID-19 mRNA-1283 vaccine in the United States

Overview

A Markov model with monthly cycles, described in detail by Fust et al. (2025)¹⁶, was used to project the potential number of symptomatic COVID-19 infections and clinical outcomes (hospitalizations, deaths and cases of long COVID) and associated costs from the societal perspective and quality-adjusted life-year (QALY) losses expected in the target population in the US. Analyses were performed over a one-year time horizon from September 2025 to August 2026 with a single dose of mRNA-1283 (annual vaccination) compared to no vaccination,^3,17,18 consistent with the CDC cost-effectiveness analyses and Prosser et al. (2025). Cost-effectiveness, in terms of incremental cost per QALY gained, was assessed from the societal cost perspective. A scenario analysis was conducted in the target population where those 65 years and older were eligible to receive a second dose a minimum of two months after the first dose (semi-annual vaccination) considering current ACIP recommendations¹⁹ for 2 or more doses in 65 years and older. In addition, vaccination with mRNA-1283 was compared to vaccination with mRNA-1273 and BNT162b2.²⁰ As the value of COVID-19 vaccination in post-pandemic settings is overlooked, especially from stakeholders outside the healthcare sector, the benefit cost ratio (BCR) was calculated for all scenarios to inform debate about the value of COVID-19 prevention holistically. This was defined as the benefit of vaccination (cost savings associated with prevention of disease and the monetized health gains due to averted COVID-19 outcomes) divided by the costs of vaccination (costs and monetized health lost due to vaccine adverse events (AEs)). Health was monetized using various assumptions: a QALY gained was valued at the standard US willingness to pay for QALYs gained of $100,000 and $150,000;²¹ value per statistical life-year (VSLY) $604,000²² and value per QALY (VQALY) $717,000.²² A completed CHEERS checklist can be found in the Technical Appendix.

Model Structure

The cohort begins the model in the Well health state (Figure 1) in which each person is susceptible to symptomatic COVID-19 infection irrespective of their previous history of COVID-19 infections and vaccinations. In each month moving forward, a portion of those in the Well state may receive a vaccine and are then subject to vaccine-related AEs (AE rates, costs and QALY losses are described in the Technical Appendix Section 8). Each month, individuals in the Well health state are at risk of symptomatic COVID-19 infection. If they develop an infection, they move through the COVID-19 consequences decision tree (Figure 2) which estimates the subsequent clinical outcomes of infection and related costs and QALY losses. Those who die following their COVID-19 infection move into the Dead health state while all others return to the Well health state. Those who did not develop an infection remain in the Well health state.

• Download figure
• Open in new tab

Figure 1.

Illustration of the Markov model

• Download figure
• Open in new tab

Figure 2:

Illustration of the consequences of COVID-19 decision tree structure

The risk of infection and the risk of hospitalization once an infection has developed is reduced in those who are vaccinated. Starting from an initial VE, VE declines linearly on a monthly basis within the model. As a portion of the cohort can receive a new vaccine each month, the effectiveness calculation is a function of the fraction of the age group vaccinated each month and VE. The hospitalization VE values are adjusted so that the incremental protection against hospitalization above the protection against infection is applied to the probability of hospitalization. (See Technical Appendix Section 2).

Target Population

The size of the cohort was based on the US population²³ and FDA licensure for mRNA-1283.⁶ For ages 12-64, the cohort size eligible for vaccination is limited to high-risk individuals, estimated based on data from Kompanieyts et al. (2021)²⁴ for adolescents and Panagiotakopoulos (2025)²⁵ for adults 18-64 (See Appendix). A total of 214,034,195 people, including persons 65 and older in addition to eligible 12-64 year olds were modelled.

Infection Incidence

The model required the incidence of symptomatic COVID-19 infection in the target population when no one receives the vaccine. As these data are not collected by the CDC, this input was approximated based on estimates of observed hospitalization rates per 100,000 by age group for the 2023-2024 and 2024-2025 seasons from COVID NET²⁶, using data on vaccine coverage and effectiveness and the conditional probability of hospitalization given infection in a modified version of the model, as described in the Technical Appendix. Given the uncertainty in COVID-19 incidence in the post-pandemic setting, similar to approaches for seasonal influenza,²⁷ an average of the estimated incidence rates for September 2023 to August 2024 and September 2024 to August 2025 were used to project the COVID-19 incidence for season 2025/2026 for the base case in this analysis. Use of the values from the individual seasons was tested in sensitivity analyses. For high-risk patients, the incidence of symptomatic COVID-19 infection was assumed to be equivalent to the incidence of COVID-19 infection for the general population, as it is assumed that high-risk patients are more likely to experience more severe COVID-19 outcomes but are not at increased risk of developing COVID-19 infection.³

Vaccine Coverage

The vaccine coverage rate (VCR) of the first seasonal dose of the vaccines (annual vaccination) for season 2025/2026 was assumed based on data from COVIDVaxView.^28,29 Data from September 2024 to July 2025 were available, with no further uptake assumed in August 2025 (See figure in the Technical Appendix Section 4). Data for those 12-17 years was taken from the pediatric database, while data for the other age groups from the adult database. For the scenario where those aged 65 years and over received a second seasonal dose (semi-annual vaccination), in the absence of CDC-reported data on second dose coverage for the 2024/2025 season, the same coverage as observed during 2023/2024 season was used as an approximation of the 2024/2025 estimates. The use of the 2023/2024 data likely reflects a conservative estimate of second dose coverage, as data from Los Angeles county suggest second dose coverage rates of approximately 22% in those ≥65 years who already had received a first those.³⁰ In absence of CDC data, the VCR for the high-risk population was conservatively assumed to be equivalent to the general population.

Vaccine Effectiveness

The VE values used for the model analytic time horizon were derived based on the age-specific rVEs of mRNA-1283 compared to the mRNA-1273 vaccine from the pivotal phase 3 randomized clinical trial (RCT) NextCOVE^8,31 and the 2024-2025 US real-world VEs of mRNA-1273.712 targeting KP.2.³² Estimates for high-risk populations in RCT and real-world effectiveness (RWE) studies were used for those aged 18-64 years with high-risk in the model. mRNA-1273 VEs were based on a large, US nationwide RWE study providing interim VE estimates of mRNA-1273. For those aged 12 to 17 years, the 2024-2025 VEs for mRNA-1273 against infection and hospitalization were assumed to be the same and were based on MacNeil’s estimate of the real-world VE of 2024-2025 COVID-19 vaccines against emergency department and urgent care visits in children aged 5-17 years, as these were the only available US VE estimates applicable to mRNA-1273 for this age range.³³ rVE values for mRNA-1283 compared to mRNA-1273 were applied to the mRNA-1273 VEs to calculate the initial VEs for mRNA-1283 as displayed in Table 1 (details provided in the Technical Appendix). The 95% confidence intervals (CIs) for the VE of mRNA-1283 against infection and hospitalization were derived based on Monte-Carlo simulations considering the 95% CIs of the VE of mRNA-1273 and of the rVE estimates of mRNA-1283 vs mRNA-1273.

For the scenarios, where mRNA-1283 was compared to mRNA-1273 or BNT-162b2, the initial VE estimates shown in Table 2 were used. To estimate the VE of BNT162b2, the rVE estimates derived in an indirect treatment comparison (ITC)¹⁴ were applied to the VE of mRNA-1283. For the 65 plus and the high-risk 18 to 64 years of age populations the rVE estimate of mRNA-1283 vs BNT162b2 against symptomatic COVID-19 infection of 22.8% (95% CI: 3.7%; 38.1%) and 19.0% (95% CI: 4.9; 31.0%) were assumed. For the rVE against hospitalization, the overall population estimate of 44.1% (95% CI: 3.2; 67.7%) was assumed; this number was derived applying the same methods as outlined in Beck et al. but using the NextCOVE estimate of the rVE of mRNA-1283 vs mRNA-1273 (38.1%, 95% CI: −6.7; 64.1%) against FDA defined severe COVID and the Kopel et al.³⁴ estimate of the rVE of mRNA-1273 vs BNT162b2 (9.8%, 95% CI: 2.6; 16.4%) against COVID-19 hospitalization.³⁵ For ages 12 to 17, BNT162b2 VEs against infection and hospitalization were assumed to be the same as mRNA-1273 as the values were based on a mixed population that received either of the two vaccines.

The monthly VE waning of mRNA vaccines against infection for Omicron variants relative to those who had never been vaccinated was estimated in a meta-analysis by Higdon et al., (2022)³⁶ to be 4.75% (95% CI 3.05%; 6.75%). We considered this to be an appropriate source for VE waning against infection because there is minimal to no protection against COVID-19 infection a year after receiving vaccination, which implies that waning of annual vaccination VE would be similar to the waning of VE of those who received initial doses of COVID-19 vaccine. However, as protection against hospitalization is longer lasting, a study by Andersson et al., (2024)³⁷ which examined the effectiveness of XBB.1.5 containing COVID-19 mRNA vaccines was used to estimate the base case rate of waning of VE against hospitalization of 2.46% per month, also considering these findings to be consistent with findings for JN.1 and KP.2 updated vaccines.^3,38 In sensitivity analyses, the estimate from Higdon et al, against hospitalization (1.37%), as well as an alternate source which examine the durability of the mRNA-1273 XBB.1.5 vaccine (3.87%)³⁹ were tested.

Probabilities for the COVID-19 Consequences Tree

Key probability inputs are summarized in Table 3 and the text. Further details are provided in the Technical Appendix. Patients with symptomatic COVID-19 infections are divided by their highest level of medical care received: 1) No formal health care (not hospitalized nor was outpatient care received); 2) Outpatient care (not hospitalized but did receive outpatient physician visits or emergency department visits without hospital admission); and 3) Inpatient care. These age-specific probabilities were estimated based on database analyses and adjusted to represent the high-risk population for ages 12 to 64 years as described in the Technical Appendix (Supplemental Table 10). Hospitalized patients are further stratified by location of care, including general ward (no intensive care unit [ICU] or ventilation required), ICU (excluding extracorpeal membrane oxygenation or invasive ventilation), or ICU with mechanical ventilation based on data from the COVID-Net surveillance system.²⁶ In absence of data, the proportions for the high-risk populations were assumed to be equivalent to the general population.

Hospitalized patients may die, be discharged and then readmitted, or simply discharged. Estimates of in-hospital mortality for the general population⁴⁰, stratified by age and in-hospital location of care, were based on COVID-NET data from March 2022 – October 2023 as analyzed by the University of Michigan COVID-19 Vaccination Modeling Team.⁴⁰ Based on Joshi et al. (2025) ⁴¹, it was conservatively assumed that the relative risk for in-hospital mortality for high-risk patients relative to non-high-risk patients was 1.23, considering an estimate derived for patients with diabetes. The COVID-NET surveillance system tracks only in-hospital deaths and not post-discharge mortality. The 30-day readmission and post-discharge mortality probabilities were estimated from a meta-analysis for the general population.⁴² The readmission rate for the high-risk populations was calculated using the RR for patients with diabetes (RR=1.27) from Joshi et al. (2025)⁴¹ and a weighted²⁵ average of high-risk and general population of 9.35% (range: 4.1%, 9.28%) was used. A post-discharge mortality of 7.87% (95% CI: 2.8% to 13.0%) was assumed to be the same for both the high-risk and general populations and applied to all age groups.

The probability of long COVID and infection-related myocarditis is described in the Technical Appendix Sections 7 and 9.

Quality of Life

QALY losses due to COVID-19 related outcomes and AEs were assumed to be the same for the general population and high-risk patients. They are summarized in Tables 25 and 26 of the Appendix.

QALY losses associated during the acute infection phase were estimated for consistency with data presented to ACIP by the University of Michigan COVID-19 Vaccination Modeling Team.⁴⁰ Caregiver QALY losses were included for patients <18 years based on caregiver spillover data reported by Mercon et al. (2025) ⁴⁶; caregiver QALY losses were not included for adult patients. Patients receiving no formal care were assumed to experience QALY losses equal to those seeking outpatient care, as the health states described in Mercon et al. (2025) ⁴⁶ do not include requirements for medical attendance.

Post-COVID QALY losses were estimated based on data from Sandmann et al. (2022) ⁴⁷ and the “Post-HOSPitalization COVID-19 study” and applied to patients who sought outpatient care and patients who were hospitalized, respectively; patients who did not seek formal health care were assumed to have no post-infection QALY loss (post-COVID QALY losses were only applied to those that survived a medically attended acute phase of illness). The post-COVID QALY losses applied to those who received outpatient care are assumed to reflect the 6-month period post-infection, while the estimates for hospitalized patients reflect the 1-year period post-discharge; these estimates exclude the acute phase of care. It is assumed that the post-COVID QALY losses capture long-term health impacts associated with not only long COVID, but also exacerbations of underlying medical conditions and any other long-term impacts of COVID-19 infection. No additional QALY losses specific to long COVID are applied.

The loss of QALYs due to infection-related mortality includes losses over the patient’s entire life span, not just the months lost within the one-year time horizon. QALYs lost were calculated as the discounted present values (at a rate of 3% per year)⁴⁸ of the lost utility over the remaining expected lifetime, based on age-specific life expectancy at the time of the event ⁴⁹ and standard baseline utility values.⁵⁰

Healthcare Costs

All cost estimates are presented in 2025 US dollars and are summarized in the Appendix with key inputs described below.

Hospitalization costs, stratified by location of care within the hospital setting, for the general population were estimated from Yehoshua et al. ⁵¹, a database analysis designed to examine healthcare costs and resource utilization for patients with a hospital admission with an ICD-10 code for COVID-19 (U07.1) between February 4, 2023 and February 29, 2024 (XBB and JN.1 Omicron lineages). The cost of inpatient care for high-risk patients was estimated using a similar approach to the probability of hospitalization given symptomatic infection and inpatient mortality. It was assumed that the RR for increased costs for high-risk patients was 1.07 based on Kapinos et al.⁵² Weighted averages of the high-risk and general population estimates of hospitalization costs were calculated as shown in the Technical Appendix Table 21 to be $51,415 for patients who received ventilator support, $22,449 for those treated in the ICU without ventilation and $51,415 for all other patients. It was assumed that patients experiencing hospital readmission would be readmitted to their original location of care, and readmission costs were assumed to be equivalent to the initial hospitalization costs.

A post-COVID cost toll, estimated separately for outpatients and hospitalized patients, is applied to those who survive the acute phase of COVID-19 infection. This toll is designed to capture medical costs related to long-term health effects of COVID-19 infection in the year following COVID-19 diagnosis, including, but not limited to, long COVID. Post-COVID costs for those ages ≥65 years were based on Chambers et al. (2023)⁵³, and high-risk estimates were based on Scott et al. (2024) ^54,55; the model uses a weighted average of the high-risk and general population in all analyses. To avoid double-counting with the post-COVID estimates, long COVID is associated with productivity losses in model analyses only, and no additional cost impact specific to long COVID is applied. 2025/2026 season private sector list prices were assumed for the unit cost of the vaccines: - $177.12 for mRNA-1283, $141.80 for mRNA-1273 and $147.69 for BNT162b2.⁵⁶ An administration fee of $20.05 was included in base case.⁵⁷ A scenario analysis examining a vaccine administration fee of $40 based on the Centers for Medicare & Medicaid Services (CMS) national payment allowance in 2024-2025 for COVID-19 vaccine administration.⁵⁸

Lost Productivity

Non-health care associated costs for COVID-19 patients included lost productivity from time missed from work due to receiving a COVID-19 vaccine and AEs, COVID-19 infection and long COVID, and infection-induced myocarditis. Caregiver productivity loss was also included to account for time loss from work to care for patients with COVID-19. It was assumed that all patients under 18 years of age required caregiver time for vaccination, AEs associated with vaccination, and during the acute infection period; it was further assumed that adults would not require a caregiver for vaccination, AEs, or during the acute symptomatic infection period. It was assumed that all patients with long COVID (both <18 and ≥18 years of age) would require a caregiver based on data from Kwon et al. (2024)⁵⁹. Lost productivity estimates are provided in the Technical Appendix; general population estimates were used for all model age groups, with the exception of the percentage in the labor force, which was adjusted for the high-risk population.

In the base-case analysis, the productivity losses due to COVID-19 premature mortality were included, using a human capital approach based on age-stratified market productivity loss estimates from Grosse et al. (2019).⁶⁰ The inclusion of productivity loss due to premature COVID-19 mortality is consistent with recent health economics analyses performed by ACIP and the CDC for RSV and pneumococcal vaccines.^61–63 A scenario analysis where non-market productivity losses were also included was performed; additionally, a scenario analysis excluding all premature mortality-related productivity losses (both market and non-market) was performed. Sensitivity Analyses

Deterministic sensitivity analyses (DSAs) performed to assess the impact of key variables on estimated symptomatic infections, COVID-19 hospitalizations, deaths are presented in the technical appendix. Deterministic sensitivity analyses were also designed to assess the robustness of the cost-effectiveness results. Several of these have been described above in the text and include parameters related to incidence of COVID-19 infection, probabilities of infection consequences, VE and vaccine coverage. In addition, COVID-19 related costs and QALY losses were also varied. Where available, estimates were varied according to 95% confidence intervals (CIs) or reported ranges. For all other DSAs, an alternative data source was used or parameters were varied by ±10% or ±25% of their base-case value.

Scenario Analyses

Several scenario analyses described below were conducted (further detail on model inputs are presented in the Technical appendix). Two scenario analyses were conducted for VE: 1) rVE against hospitalization for mRNA-1283 versus mRNA-1273 were set to be equal to the rVE against infections; and 2) VE against symptomatic COVID-19 infection for adults ages 18 years and older from ICATT study.⁶⁴ The model cannot estimate any benefit associated with changes in transmission, which may include the reduction in secondary transmission when a vaccinated person develops an infection. This effect was demonstrated by the CDC RIGHT study³³, which estimated that if infected people had been vaccinated within the prior 6 months, their risk of transmitting to household contacts was reduced compared to the who had not been vaccinated (adjusted RR 0.55; VE: 45%). These results were therefore used to create an indirect benefit scenario.

Vaccine coverage rates in September were increased either by 5% or 10%, for all ages and for the ≥65 years subgroup. The percentage of patients seeking no formal care for COVID-19 related illness was assumed to be equal to CDC influenza rates from the 2023-2024 season. Several scenario analyses were also performed specific to high-risk individuals, including assuming that 50% of 12-17 year-olds are considered high-risk, that the likelihood of receiving outpatient care is equivalent to the general population, that the RR of inpatient mortality is 1.62 based on cardiovascular disease⁴¹, and using the percentage hospitalized based on the RR derived from Wilson et al. (2025) (RR=1.48). Scenario analysis considering the healthcare perspective as well as the societal perspective including also non-market productivity losses associated with premature mortality, excluding all productivity losses (market and non-market) associated with premature mortality, and excluding caregiver productivity losses and by considering opportunity costs (Supplemental Section 12).

The number of antibiotic prescriptions that can be avoided through vaccination with mRNA-1283 was investigated (further details are provided in the Technical appendix section 18).

For the analyses comparing mRNA-1283 to mRNA-1273 and BNT-162b2, a range of rVE values were used to calculate the ranges used for sensitivity analyses as shown in Table 2; scenario analyses examining COVID-19 incidence and vaccine coverage were also performed for these comparisons.

Base case analysis (compared to no vaccine)

As shown in Table 4, use of an annual dose of mRNA-1283 is estimated to reduce the number of symptomatic infections, hospitalizations, and deaths COVID in the target population by 2.2 million, 137,391, and 17,960 compared to no vaccine, respectively (additional clinical outcomes are presented in Table 4). The corresponding numbers needed to vaccinate (NNV) to avert one COVID-19 symptomatic infection, outpatient episode, hospitalization, death and long COVID were estimated at 27, 90, 432, 3,301, and 908 respectively.

The majority of severe COVID-19 cases, defined as hospitalizations and deaths, were prevented among adults aged 65 years and older, accounting for 123,324 hospitalizations and 16,367 deaths. These figures represent 89.8% of all hospitalizations and 91.1% of all deaths, respectively. Second dose coverage has been low, but use of semi-annual dosing in those 65 years and older is estimated to increase the number averted by an additional 37,939 symptomatic infections, 1,632 hospitalizations, and 217 deaths in the overall target population of 12 to 64 high-risk and 65 years and older compared to the base-case strategy of annual vaccination for all ages.

For the broader public health impact, mRNA-1283 compared to no vaccine was estimated to prevent 252,892 hospital bed days during crowded conditions and the number of antibiotic prescriptions avoided in the outpatient setting with use of mRNA-1283 compared to no vaccine is 167,896, with 83% of the antibiotic prescriptions avoided being among those 65 years and older.

Given this clinical impact, annual dosing with mRNA-1283 in the target population (base-case analysis) is expected to lead to a gain of 171,150 QALYs compared to no vaccine. Although vaccination costs $12 billion, mRNA-1283 prevents $9.3 billion in costs due to reductions in COVID-19 infection-related costs and lost productivity savings compared to no vaccine. The incremental cost per QALY gained for mRNA-1283 compared to no vaccine is therefore $16,247 (Table 6). In the subgroup of adults aged ≥65 years, mRNA-1283 is associated with 132,619 fewer QALYs lost and $91.1 million in cost savings compared to no vaccine, resulting in mRNA-1283 dominating no vaccine in this population, while the incremental cost per QALY gained for high risk 12 – 64 years olds was $74,530 (See Supplement Table 34). In the scenario of the target population including semi-annual dosing for those aged ≥65 years, vaccine costs increase to $13.2 billion; mRNA-1283 is associated with $3.7 billion in additional costs and 173,188 QALYs gained for an ICER of $21,459 per QALY gained relative to no vaccine.

Vaccinating with mRNA-1283 generates at least a return on investment of $2.16 per one dollar spent. The BCR increases to $9.74 per dollar spent for mRNA-1283 COVID-19 vaccination as the method of monetizing health benefit is varied. (See Supplement Table 50 for all BCRs for base case and the scenario analyses with semi-annual dosing.)

Deterministic Sensitivity Analyses (mRNA-1283 compared to no vaccine)

The impact of deterministic sensitivity analyses (DSAs) on the number of symptomatic infections, hospitalizations, and deaths prevented is summarized in the Technical Appendix (Supplemental Table 37); the impact of key DSAs on cost-effectiveness results is presented in Figure 3 (details in Supplemental Table 38). Varying the COVID-19 incidence has the greatest impact on the ICER; varying COVID-19 incidence to the lower bound estimates based on the 2024-2025 season yields an ICER of $51,113 per QALY gained relative to no vaccine (214.7% change from the base-case ICER), while varying to the upper bound estimates based on the 2023-2024 season leads to mRNA-1283 dominating no vaccine. Varying the proportion of patients with symptomatic infection who are hospitalized also has a large impact on the ICER; varying to −25% and +25% of the base-case values yields ICERs of $31,826 (95.9%) and $6,100 (−62.5%) per QALY gained, respectively. Varying post-discharge mortality to the 95% CI upper bound decreases the ICER to $8,784 (45.9% decrease), while the 95% CI lower bound increases the ICER to $29,794 (83.4%). Varying the mRNA-1283 initial VE against hospitalization based on the 95% CI lower and upper bounds led to ICERs of $28,493 (75.4% increase) and $8,301 (48.9% decrease), respectively. Other key variables impacting the cost-effectiveness results by >20% include the vaccine administration cost, the mRNA-1283 waning rates against infection and hospitalization, post-discharge costs, mRNA-1283 initial VE against infection, and hospitalization costs. Vaccine coverage also impacts cost-effectiveness results by 20%, as VCRs impact clinical outcomes due to the heterogeneity in model age groups (i.e., changing coverage rates in model age groups will affect clinical outcomes).

• Download figure
• Open in new tab

Figure 3:

Tornados: Impact of deterministic sensitivity analyses on the base-case cost-effectiveness results (mRNA-1283 relative to no vaccine)

Scenario Analyses (mRNA-1283 compared to no vaccine)

The results of the impact of scenario analyses on infections, hospitalizations, and deaths in the base-case target population are shown in the Technical Appendix; the economic impact in the comparison of mRNA-1283 to no vaccine is depicted in the table below (Table 6). Premature mortality productivity losses were shown to have a large impact on the ICER; including both non-market and market productivity losses leads to mRNA-1283 dominating no vaccine, while excluding both market and non-market premature mortality productivity losses leads to a 79.5% increase in the ICER (compared to the base-case) to $29,171 per QALY gained. An analysis performed from the healthcare perspective (excluding both patient and caregiver lost productivity) yielded an ICER of $30,886 per QALY gained, representing a 90% increase from the base-case. Assuming the rVE for hospitalization is equivalent to the rVE against infection (i.e., there is no additional benefit against hospitalization) increases the ICER by 60% to $26,018 per QALY gained. Basing the proportion of patients with no formal care for COVID-19 related-illness on CDC burden of illness influenza data from the 2023-2024 season leads to a 56% reduction in the ICER; this is largely due to the percentage with no formal care being lower in the influenza data compared to the base-case analysis. In the scenario analysis examining indirect benefit, the ICER is reduced to $9,653 (40.6% decrease).

Comparison against originally licensed mRNA COVID-19 vaccines mRNA-1273 and BNT162b2

Clinical outcomes for the comparisons between mRNA-1283 and mRNA-1273 and mRNA-1283 and BNT162b2 are presented in the Technical Appendix (Supplemental Tables 32 and 33).

In the target population, mRNA-1283 is expected to avert an additional 427,337 symptomatic COVID-19 infections, 36,263 COVID-19 related hospitalizations and 4,725 deaths compared to mRNA-1273. This impact on clinical outcomes leads to reduced total costs for the mRNA-1283 vaccination strategy compared to mRNA-1273, and fewer QALYs lost, leading to mRNA-1283 dominating mRNA-1273 in the base-case analysis for both the target population and the ≥65 years subgroup (Table 6; Supplemental Table 35). In the 12-64 high risk group the incremental cost per QALY gained was $14,206. Compared to BNT162b2, mRNA-1283 is expected to avert 606,942 symptomatic infections, 46,176 hospitalizations, and 6,023 deaths, leading to reduced total costs and fewer QALYs lost for the mRNA-1283 vaccination strategy compared to BNT162b2. As a result, mRNA-1283 dominates BNT162b2 in the base-case analysis for both the target population and the ≥65 years and 12-64 high risk subgroups (Table 6; Supplemental Table 43). The BCR for the base case target population for mRNA-1283 relative to mRNA-1273 ranged from 3.14 to 14.11, while the BCR for mRNA-1283 relative to BNT162b2 ranged from 4.81 to 20.91.

Scenario analyses for the comparisons between mRNA-1283 and mRNA-1273 and mRNA-1283 and BNT162b2 in the target population are presented in the Technical Appendix (Supplemental Tables 32 and 33). Varying the COVID-19 incidence to the 2024-2025 season yielded an ICER of $20,397 per QALY gained for mRNA-1283 compared to mRNA-1273. Varying the rVEs for infection and hospitalization to their respective lower bounds led to ICERs of $9,907 and $77,473 per QALY gained for mRNA-1283 relative to mRNA-1273, respectively. In all other scenarios, mRNA-1283 dominated mRNA-1273, consistent with the base-case results. In the comparison against BNT162b2, varying the rVE of hospitalization to the lower bound led to an ICER of $483, per QALY gained for mRNA-1283 relative to BNT162b2. In all other scenarios, consistent with the base-case analysis, mRNA-1283 dominated BNT162b2.