Comparison of Strategies and Incidence Thresholds for Vi Conjugate Vaccines Against Typhoid Fever: A Cost-effectiveness Modeling Study

Abstract

Background: Typhoid fever remains a major public health problem globally. While new Vi conjugate vaccines hold promise for averting disease, the optimal programmatic delivery remains unclear. We aimed to identify the strategies and associated epidemiologic conditions under which Vi conjugate vaccines would be cost-effective.

Methods: We developed a dynamic, age-structured transmission and cost-effectiveness model that simulated multiple vaccination strategies with a typhoid Vi conjugate vaccine from a societal perspective. We simulated 10-year vaccination programs with (1) routine immunization of infants (aged <1 year) through the Expanded Program on Immunization (EPI) and (2) routine immunization of infants through the EPI plus a 1-time catch-up campaign in school-aged children (aged 5-14 years). In the base case analysis, we assumed a 0.5% case-fatality rate for all cases of clinically symptomatic typhoid fever and defined strategies as highly cost-effective by using the definition of a low-income country (defined as a country with a gross domestic product of $1045 per capita). We defined incidence as the true number of clinically symptomatic people in the population per year.

Results: Vi conjugate typhoid vaccines were highly cost-effective when administered by routine immunization activities through the EPI in settings with an annual incidence of >50 cases/100000 (95% uncertainty interval, 40-75 cases) and when administered through the EPI plus a catch-up campaign in settings with an annual incidence of >130 cases/100000 (95% uncertainty interval, 50-395 cases). The incidence threshold was sensitive to the typhoid-related case-fatality rate, carrier contribution to transmission, vaccine characteristics, and country-specific economic threshold for cost-effectiveness.

Conclusions: Typhoid Vi conjugate vaccines would be highly cost-effective in low-income countries in settings of moderate typhoid incidence (50 cases/100000 annually). These results were sensitive to case-fatality rates, underscoring the need to consider factors contributing to typhoid mortality (eg, healthcare access and antimicrobial resistance) in the global vaccination strategy.

Figure 1.

Calibration of transmission model to low-, moderate-, and high-endemicity settings, with model schematic. The model was calibrated to age-stratified annual incidence in 3 scenarios: low endemicity (10 cases/100000; A), moderate endemicity (50 cases/100000; B), and high endemicity (200 cases/100000; C). The model-predicted (gray) and observed (orange) incidences of typhoid cases are shown by age group. Note the different scale for the y-axis on each plot. D, The model structure for typhoid transmission.

Figure 2.

Effectiveness of Vi conjugate vaccination strategies on typhoid fever incidence. We modeled the effectiveness of routine immunization of infants (through the Expanded Program on Immunization [EPI]) and routine immunization through the EPI with 1 catch-up campaign in school-aged children (EPI+Catch-up) on the incidence of typhoid fever over a 10-year period. We modeled settings of low endemicity (10 cases/100000; A), moderate endemicity (50 cases/100000; B), and high endemicity (200 cases/100000; C). Note the different scale for the y-axis on each plot.

Figure 3.

Scenario analysis for the cost-effective incidence threshold under changing case-fatality rate estimates. This scenario analysis tested the effect of increasing the case-fatality rate, which may result from increasing prevalence of antimicrobial resistance to Salmonella enterica serovar Typhi.

Figure 4.

Cost-effectiveness acceptability curves assessing the probability either strategy falls below the willingness-to-pay threshold. We used a probabilistic sensitivity analysis varying all cost, disability, and epidemiology model inputs to compute the probability of each strategy being highly cost-effective at a specified willingness-to-pay threshold ($/averted DALY). Curves are provided for 3 settings: low endemicity (10 cases/100000; A), moderate endemicity (50 cases/100000; B), and high endemicity (200 cases/100000; C). The vertical lines indicate 2 commonly used willingness-to-pay thresholds of 1 times the gross domestic product (GDP) per capita ($1035/averted DALY; left) and 3 times the GDP per capita ($3105/averted DALY; right). A negative incremental cost-effectiveness ratio (ICER) reflects a cost-saving strategy. Note that the low-incidence setting (A) is graphed across a wider x-axis scale as compared to settings of moderate and high endemicity.

Figure 5.

One-way sensitivity analysis of key model parameters to assess changes in the incidence threshold for either strategy. This analysis tested the effect of varying single model parameters for cost, disability, and epidemiology on the cost-effective incidence threshold for vaccination through the Expanded Program on Immunization (EPI) relative to no vaccination (A) and routine immunization through the EPI with a school-based catch-up campaign relative to routine immunization through the EPI alone. The horizontal axis depicts the incidence (cases/100000) at which the strategy was highly cost-effective (willingness to pay of $1035/averted disability-adjusted life-year [DALY]). Each model parameter is presented with the base case value, and lower and upper range in parentheses.