Study design
We conducted a model-based economic evaluation to assess the cost-effectiveness of 21 vaccination strategies, including a no-vaccination scenario, in China, from a health-care system perspective. The model was developed using Microsoft Excel 2019 and the analysis was reported according to the Consolidated Health Economic Evaluation Reporting Standards statement[9].
Modelling
A decision tree model was developed to simulate disease and health outcomes within a targeted vaccination population, including individuals aged 60 years, those aged 60 years with high risk of cardiovascular or respiratory disease, and pregnant women (Figure 1).
We determined the total number of individuals requiring vaccination based on China’s seventh population census, assuming that the number of pregnant women equated to the number of newborns[10].
We calculated the incidence of respiratory infections within the vaccinated population by multiplying the incidence of GBD (Global Burden of Disease) - reported and (1-vaccine effectiveness)[3].
Natural background mortality was excluded from the model since the study period was limited to only one year. Considering that the follow-up duration of the clinical trial was approximately 6 months, it was assumed to represent the vaccination protection period, which aligned with the majority of published cost-effectiveness literature[11-13].
In addition, given the favorable safety profile of vaccines in clinical trials, we did not incorporate the impact of adverse events in our modelling[5].
We assumed that the upper respiratory infections caused by RSV could be resolved spontaneously or with outpatient treatment. Conversely, lower respiratory infections were deemed more likely to necessitate additional medical help or intensive care admission[14-17].
Model parameters were defined in Table 1.
Table 1 Parameters for Cost-Effectiveness Analysis Decision Model
Parameter
|
Base case
|
Range (distribution)
|
Sources
|
|
|
Queue population
|
|
|
|
|
Population of individuals aged 60
|
13280176
|
|
[10]
|
|
Population of individuals aged 60 at high risk
|
1819384
|
|
[18]
|
|
Annual number of pregnant women
|
9772477
|
|
[19]
|
|
Incidence rate
|
|
|
|
|
Upper respiratory tract infection
|
|
|
|
|
Population (<1)
|
46.33%
|
39.88%-53.18% (beta)
|
[3]
|
|
Population (60 years old)
|
38.62%
|
30.72%-46.83% (beta)
|
|
Lower respiratory tract infection
|
|
|
|
Population (<1)
|
0.78%
|
0.60%-1.03% (beta)
|
|
Population (60 years old)
|
1.21%
|
1.01%-1.45% (beta)
|
|
RSV-related respiratory tract infection
|
|
|
|
|
Population (<1)
|
19.50%
|
17.55%-20.48% (beta)
|
[20]
|
|
Population (60 years old)
|
16.00%
|
14.40%-16.80% (beta)
|
|
Mortality rate
|
|
|
|
|
Upper respiratory tract infection
|
|
|
|
|
Population (<1)
|
0.10%
|
0.05%-0.29% (beta)
|
[3]
|
|
Population (60 years old)
|
0.01%
|
0.01%-0.03% (beta)
|
|
Population (60 years old at high risk)
|
0.01%
|
0.01%-0.04% (beta)
|
[15]
|
|
Lower respiratory tract infection
|
|
|
|
|
Population (<1)
|
12.35%
|
10.85%-13.55% (beta)
|
[3]
|
|
Population (60 years old)
|
0.76%
|
0.71%-0.89% (beta)
|
|
Population (60 years old at high risk)
|
0.92%
|
0.86%-1.07% (beta)
|
[15]
|
|
Overview of treatment modality distribution
|
|
|
|
|
Upper respiratory tract infection
|
|
|
|
|
Outpatient treatment
|
|
|
|
|
Population (<1)
|
5.20%
|
4.68%-5.72% (beta)
|
[14]
|
|
Population (60 years old)
|
26.80%
|
24.12%-29.48% (beta)
|
[17]
|
|
Population (60 years old at high risk)
|
35.93%
|
32.34%-39.52% (beta)
|
[16]
|
|
Lower respiratory tract infection
|
|
|
|
|
Outpatient treatment
|
|
|
|
|
Population (<1)
|
26.80%
|
24.12%-29.48% (beta)
|
[15]
|
|
Population (60 years old)
|
17.39%
|
15.65%-19.13% (beta)
|
|
Population (60 years old at high risk)
|
28.57%
|
25.71%-31.43% (beta)
|
|
Hospitalization
|
|
|
|
Population (<1)
|
22.00%
|
|
|
Population (60 years old)
|
24.48%
|
0.43%-77.60% (beta)
|
|
Population (60 years old at high risk)
|
32.82%
|
23.49%-43.74% (beta)
|
|
Intensive care hospitalization
|
|
|
|
Population (<1)
|
8.60%
|
|
|
Population (60 years old)
|
5.01%
|
0.47%-37.36% (beta)
|
|
Population (60 years old at high risk)
|
26.74%
|
20.40%-34.22% (beta)
|
|
Vaccine efficacy
|
|
|
|
|
Arexvy, GSK
|
82.60%
|
57.90%-94.10% (beta)
|
[5]
|
|
Abrysvo, Pfizer
|
|
|
|
Population (<1)
|
81.80%
|
40.60%-96.30% (beta)
|
|
Population (60 years old)
|
66.70%
|
28.80%-85.80% (beta)
|
|
Duration, months
|
6.00
|
2.00-10.00 (gamma)
|
|
Cost, US$
|
|
|
|
|
Vaccine/dose (US)
|
|
|
|
|
|
|
Arexvy, GSK
|
270.00
|
|
[21]
|
|
Abrysvo, Pfizer
|
200.00
|
|
|
Vaccination/dose
|
3.36
|
3.03-3.53 (gamma)
|
[22]
|
|
Treatment
|
|
|
|
|
Outpatient treatment
|
16.02
|
14.42-16.82 (gamma)
|
[23]
|
|
Hospitalization
|
654.61
|
589.15-687.34 (gamma)
|
|
Intensive care hospitalization
|
1053.12
|
947.80-1105.77 (gamma)
|
|
Productivity losses due to deaths (<1)
|
339682.27
|
12074.77-2390362.39 (gamma)
|
Calculated
|
|
Utility, YLDs or YLLs
|
|
|
|
|
No treatment or outpatient treatment, YLDs
|
0.0037
|
0.0033-0.0039 (beta)
|
[24]
|
|
Hospitalization, YLDs
|
0.0102
|
0.0092-0.0107 (beta)
|
|
Death (<1), YLLs
|
17.39
|
13.4-80.79 (gamma)
|
Calculated
|
|
Death (60), YLLs
|
12.88
|
12.57-38.64 (gamma)
|
|
Intervention strategies
We defined the absence of vaccination as the status quo in China because the vaccines were unavailable in China. A total of 21 vaccination strategies were investigated in our study, comprising a combination involving three potential vaccinated population with two vaccines, each at three different vaccination coverage levels (25%, 50%, and 75%). We assumed that vaccination was implemented when the cohort entered the model, with pregnant women receiving vaccination during labor, providing six months of protection for the infants.
Data analysis
The costs of treatment encompassed the costs of diagnosis, hospitalization, and medicines (Table 1). The costs of vaccination encompassed the vaccine price (25%/50%/75%/100% of US prices), the cost of staff services during vaccination, and the cost of advocacy for vaccination. All costs were converted from Chinese RMB to US dollars ($1=¥7.1254, applicable in 2023)[25]. To estimate the lost productivity from early infant mortality, we employed established formulas[26].The discount rate and wage growth rate were set at 3% (0-6%) and 4.9% (0-8%), respectively[27, 28], and the natural mortality rate assumed remains relatively stable in the future, derived from population census[10].
Notes: r=discount rate, g=wage growth rate, pi=probability of surviving to year i, Mi= productivity loss for surviving to year i, N=natural mortality rate, GDP2022=China's GDP per capita in 2022. The retirement age in China was about 60 years old, so we had only calculated the productivity loss due to the early deaths of infants.
We selected Disability-Adjusted Life Years (DALYs) as the measure of effectiveness, and DALY was the summation of Years Lost Due to Disability (YLD) and Years of Life Lost (YLL). We obtained YLD from published literature (Table 1) and calculated YLLs using the formula.
The expected DALYs and costs for each strategy were derived from the model. We calculated the incremental costs and Averted DALYs for each vaccination strategy compared with a scenario of no vaccination across four price settings. We identified the cost-effectiveness frontier and calculated the incremental cost-effectiveness ratio (ICER), defined as the incremental cost per DALY averted for each strategy on the cost-effectiveness frontier compared with a lower-cost non-dominated strategy to ascertain the most cost-effective strategy. We adhered to the criteria of cost-effectiveness from WHO (classified as highly cost-effective, cost-effective, or not cost-effective based on an ICER <1, 1–3, or >3-times the per-capita gross domestic product [GDP]; Chinese per-capita GDP was $12741 in 2023)[29].
The selection of distribution for all parameters was based on consideration of the properties of the parameters and data informing the parameters (Table 1).
Univariate sensitivity analysis was conducted for each parameter within its respective ranges to pinpoint the most sensitive parameters. Probabilistic sensitivity analysis was performed through 10,000 simulations to establish the probability of cost-effectiveness for each intervention strategy compared with all others. The reporting of methods and results conformed to the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) (as shown in the Supplement Table 1)[9].