2.1 Model Overview

A cost-effectiveness analysis was conducted comparing the SoC with nirsevimab for all Dutch infants during their first RSV season, incorporating the effects and costs related to RSV-LRTD and immunization, from a societal perspective. A static decision-analytic model was adapted for the Netherlands using Excel (Fig. 1) [12]. All infants are susceptible to acquiring RSV-LRTD, in which the risk is dependent on the combination of three factors: (1) proportion of RSV cases for a given month; (2) incidence rate per age (in months) and infant subpopulation; and (3) age at start of the season. In the Netherlands, the season can be defined as a 5-month period that starts in November, with a peak in January, based on data from the European RESCEU project in which age-specific estimates of RSV-associated hospitalization using multiple-season regression analyses were obtained [13]. The entire Dutch birth cohort of 2023 served as input for the model, for both interventions [14]. The time horizon is 1 year.

Fig. 1

Decision tree model, overview. * Complications: acute otitis media and recurrent wheezing. ED emergency department, GP general practitioner, MA medically attended, NMA non-medically attended, PICU paediatric intensive care unit, RSV respiratory syncytial virus

2.2 Immunization Strategies

An overview of the characteristics of nirsevimab and palivizumab can be found in Table 1. Both prophylactic treatment options are administered during the first RSV season for infants younger than 1 year.

For nirsevimab, each infant was administered a single dose. Given that nirsevimab offers 5 months of protection [15], administering it prior to the RSV season is crucial to ensure adequate protection throughout the typical 5-month RSV season of Dutch infants entering their first RSV season. This strategy was also adopted in the United States and Spain for the 2023/24 season, where the RSV seasonality is similar [10, 11]. Consequently, as the strategy is to optimize protection during the RSV peak period, the ‘in-season with catch-up’ strategy was selected as the base case. For the base-case strategy ‘in-season with catch up’, administration is either at the beginning of the RSV season for infants born out of season (April–October) or just after birth for children born during the season (November–March). Other immunization strategies that were included in our scenario analysis were (i) ‘in-season from October to April’, in which late preterm and term infants born within the season were immunized; and (ii) ‘year-round birth dose’, in which late preterm and term infants were immunized after birth, regardless of whether it is the RSV season or not. Regardless of the chosen strategy, all palivizumab-eligible infants received nirsevimab, including those who were born before the season (catch-up) and those born in-season (see Fig. S1 in the electronic supplementary material [ESM] for a graphical overview).

For palivizumab, monthly administration is needed during the RSV season of the eligible infants (born < 32 wGA or those with bronchopulmonary dysplasia in the previous 6 months or with congenital heart disease). Administration is at the beginning of the RSV season for those born outside the season (April–October) and at birth for those born within the season (November–March), with a maximum of five administrations in total [15]. However, given that the incidence data already reflect the use of palivizumab in the Netherlands in the palivizumab-eligible infants, no efficacy data was included for palivizumab in our model. In addition, no effectiveness data for palivizumab is available in the Netherlands. Given the absence of efficacy data for nirsevimab in this specific group, it was assumed that the effects of nirsevimab and palivizumab, including coverage rates, were equivalent in the palivizumab-eligible group. Nevertheless, the perceived efficacy of nirsevimab is higher in the late preterm and term infants compared with palivizumab in those at high risk [6, 16]. Since an (in)direct comparison of the effectiveness of nirsevimab and palivizumab was unavailable for the specific group of palivizumab-eligible infants, we conservatively assumed equal effectiveness and only considered the immunization costs incurred for both nirsevimab and palivizumab in this specific group. In addition, we did not alter the immunization strategy for the palivizumab-eligible infants.

2.3 Input Parameters

All data inputs are shown in Table 1. For a more detailed description and overview of the input, see ESM, Tables S1–S4. Costs were based on 2024 price levels, using the consumer price index (CPI) rate [17]. In the base case, discounting has no effect on the outcome, given that the model only accounts for 1 year and no mortality was included in the base case (no future quality-adjusted life-year [QALY] losses). In addition, no long-term consequences following an RSV hospitalization were included in the base-case analysis.

An expert panel was assembled, with all experts selected based on their experience and expertise in RSV. The elicitation process involved presenting the panel with the model structure and input, along with specific questions related to the study. Responses were documented, and further clarification was sought through follow-up questions via email. The experts' opinions were incorporated into the analysis to support the assumptions.

2.3.1 Clinical Input

For both treatment arms, the total Dutch birth cohort was taken into account. In 2023, 164,487 children were born in the Netherlands [14], of whom 7% were born preterm [18] (born < 37 weeks’ gestational age) and 1.83% (3015 infants) of these preterm children were palivizumab eligible, as they were born < 32 weeks of gestational age [19]. The number of palivizumab-eligible infants was estimated based on the average number of infants who received palivizumab from 2019 to 2021 (2512), assuming a coverage rate of 80% based on expert opinion [17, 20]. For all health outcomes (e.g., hospitalization, pediatric intensive care unit [PICU] admission, specialist visit, general practitioner [GP] visit, and mortality), an overall incidence rate for these subgroups was used in the analysis, as no distribution of the incidence among palivizumab-eligible, late preterm, and term infant is available [1]. Hospital admissions were derived from a multiseason regression analysis of weekly hospitalizations in the Netherlands from 2013 to 2017, corrected based on the observed number of hospitalizations (see ESM, Table S2) [13]. The percentage of PICU cases was estimated from a recent nationwide, prospective, observational multicenter study conducted between September 2021 and June 2023. An average of both seasons was estimated for those aged 0–5 months (0.226%) and 6–11 months (0.022%) (see ESM, Table S3) [22]. For the GP visits, it was assumed that 5 and 12.5 visits occur per RSV hospital admission of those aged 0–5 months and 6–59 months, respectively [7]. The incidence of specialist visits was estimated by applying a 10% rate following GP visits [23]. Additionally, 40% of infants were assumed to visit a specialist after a PICU admission, with 28% of these infants requiring a neuropsychological assessment [22]. In the case of emergency department (ED) visits, a rate of 21% after GP visits was considered [23]. In our base-case analysis, mortality was not included, nor were longer term effects resulting from RSV infection. It is unclear whether complications such as recurrent wheezing or mortality can be prevented by nirsevimab. Therefore, these complications were included as scenario analyses. Utilities associated with the aforementioned health effects, including QALY loss of the parent, were obtained from a prospective study conducted in Europe [24]. This study assessed the economic burden and health-related quality of life among infants with RSV across multiple European countries.

2.3.2 Costs

To estimate the total cost per hospital admission, general inpatient hospital costs per day [25] were multiplied by the mean length of stay (LOS) of 4.04 days for an RSV infection (see ESM) [26]. PICU costs include hospitalization, covering both PICU admissions and general ward readmissions, as well as transportation expenses. The LOS for a PICU admission, estimated at 5 days, was derived from an observation study focusing on infants with RSV admitted to the PICU [22]. Daily costs were calculated based on the 2023 DBC (Diagnosis Related Group) tariffs of a general PICU admission (declaration code 190151) [27]. On average, 4% of patients were readmitted after their PICU stay, with an average LOS of 4 days [22]. The proportions of patients requiring mobile intensive care unit (MICU) transport or ambulance transport were derived from the same observational study, with rates of 46% and 49%, respectively. Additionally, ambulance transport at discharge was included (75.4%) [22]. MICU transport costs were obtained from DBC tariffs, incorporating average costs for transport < 2 h (declaration code 190132) and > 2 h (declaration code 190133) [27]. For ambulance transport, costs were derived from Dutch reference prices [25]. GP visit costs include home visits (10%), standard visits (90%), and costs for a prescription drug to alleviate respiratory discomfort (with costs for salbutamol and administration costs) [15, 29].

Direct non-medical costs include the travel costs of parents, in which the mean distance that they have to travel to and from a hospital is 189.9 km (9.7–885.1) [30], with costs per kilometer of €0.21–0.33 [28] for public transport or personal vehicle and parking costs of €4.97 [28]. In addition, the percentage that travelled by public transport was assumed to be 50%.

For productivity losses of a parent, the lost days of work were estimated per inpatient and outpatient RSV episode. Because of the 3-month maternity leave in the Netherlands, productivity loss by parents/caregivers due to RSV illness in infants during their first 3 months of life was assumed to be 0. Therefore, a correction on the total lost days of work was made. For the infants in hospital, the average hospital LOS for children aged < 1 year (4.04 days, see ESM Table S4) [26] was multiplied by the probability of inpatient hospitalization of those aged 3–11 months. For outpatient visits, the reported work loss by parents was 0 for children aged 0–3 months, 0.67 days for 3–5 months, and 1.6 days for 6–11 months, which were weighted using the probability rates of GP visits [7].

The acquisition costs of nirsevimab were estimated in the model for both preterm and term infants. For the palivizumab-eligible infants, costs were derived from a Dutch database that provides detailed information on medication usage, namely, on the total number of infants who received palivizumab (2512), the number of administrations per patient (5 doses), and the cost per dose (€653.40) [19, 31].

2.4 Analysis

The outcomes were expressed in health effects, costs, and the economically justifiable price (EJP) of nirsevimab. We compared different nirsevimab strategies for all infants, including in-season with catch-up (base case), in-season only, and year-round administration, against the standard of care (palivizumab) in the palivizumab-eligible population. The EJP was calculated using a willingness-to-pay threshold of €50,000 per QALY from a societal perspective. This threshold is based on a report published in late 2023, which recommended applying a €50,000 threshold for public health interventions, including vaccines [38]. Although nirsevimab is a passive immunization rather than a traditional vaccine, we extended the application of this threshold to nirsevimab [38]. Probabilistic sensitivity analysis (PSA) was based on the calculated acquisition price at the Dutch threshold of €50,000 per QALY.

The following scenarios regarding the input data of the model were performed (see Table 3 for an overview of the input parameters):

1.

Scenario 1: GP visits from Dolk et al. [39].

2.

This scenario incorporates Dutch-specific GP RSV incidence data. This data is based on influenza-like-illness GP cases in children aged 0–4 years. The number of cases were distributed by age [23].

3.

Scenario 2: Including mortality rate or risk of death for RSV inpatient hospitalization. The number of deaths per year was used to calculate the incidence of those aged 0–5 months (0.97 deaths) and 6–59 months (0.7 deaths) [7], in which health effects were discounted by 1.5%.

4.

Scenario 3: Including recurrent wheezing. Recurrent wheezing emerges as a prolonged complication following RSV infection. Our scenario takes into account the associated costs and QALY losses attributed to recurrent wheezing, in children up to 3 years of age [7, 28, 40], in which health effects and costs were discounted with 1.5% and 3%, respectively. Costs for recurrent wheezing were based on five GP visits and prescription drug costs as presented in Table 1.

5.

Scenario 4: Including acute otitis media.

6.

Acute otitis media is a well-known complication after RSV infection, however, the effect of nirsevimab on this complication remains unclear. Our scenario only includes acute otitis media after hospitalization [41] including related costs [42] and QALY losses [24].

7.

Scenario 5: Including all-cause LRTD hospital admissions. Nirsevimab has demonstrated efficacy in reducing all-cause LRTD hospital admissions [8]. Consequently, a scenario was considered that incorporates all-cause LRTD cases, in which all-cause LRTD cases, excluding those related to RSV, were added [13]. The length of stay of a LRTD hospital case unrelated to RSV was 2 days [26].

8.

Scenario 6: PICU incidence and LOS from Linssen et al. [21]. he percentage of PICU cases was estimated from the average number of PICU RSV bronchiolitis in children over the years 2010–2016 [21]. RSV-related PICU admissions can be nosocomial (1.5%) or community-acquired, for which the LOS is 8 and 13.5 days, respectively. To calculate the overall costs, LOS was multiplied by the PICU costs per day [22].

9.

Scenario 7: Additional hospital bed days after PICU admission. In this scenario, it was assumed, based on expert opinion, that an additional 7 hospital bed days are required following PICU admission.

10.

Scenario 8: Parent/caregiver QALY loss. In this scenario, we applied QALY losses of the parent (5.48 × 10-4 instead of 0), which is based on the same source as our base case, however, from a different approach [24]. Rather than relying on QALY data specific to the Netherlands, we incorporated the mean QALY loss of the parent from all the European countries included in the cohort study, to explore the impact of parental QALY loss on the outcome [24].

11.

Scenario 9: QALY losses from Hodgson et al. [43]. An alternative data source for the QALY losses was included [43], given the challenge in determining QALY losses in infants, which introduces some uncertainty. For medical and non-medically attended cases, the QALY loss was 3.823 × 10−2 and 3.024 × 10−2, respectively.

12.

Scenario 10: Excluding palivizumab-eligible infants. In this scenario, the palivizumab-eligible group is excluded from the analysis.

13.

Scenario 11: Immunization of the palivizumab-eligible group only. Our last scenario compares the intervention exclusively within the preterm and term infant groups, excluding the palivizumab-eligible group from the analysis.

A deterministic sensitivity analysis (DSA) was performed to assess the impact of individual parameters on the EJP. A PSA with 1000 simulations was conducted while all input variables were simultaneously varied over their distributions. For the input ranges, 95% confidence intervals (CIs) were used when available. In cases where CIs were not provided, a 20% range was applied. Costs were distributed using a gamma distribution. Efficacy, end of protection, and coverage rates were modelled with a normal distribution. RSV incidence and disutilties were distributed using a beta distribution. The DSA was visualized in a tornado diagram. The PSA was visualized in a cost-effectiveness plane and a cost-effectiveness acceptability curve (CEAC) to describe the impact of the combined parameter uncertainty on the model outcome. An overview of all included input parameters for the sensitivity analysis (PSA and DSA) can be found in Supplementary Table S5 (see ESM).