We considered a plausible range of $2–$100 for VCPI to account for vaccine dose, wide distribution and administration, and wastage based on the estimates for other flavivirus vaccines (29). Our results show that for a sufficiently low VCPI in this range, a single-dose vaccination program is cost-saving for all countries studied (Figure 1, green). The lowest VCPI was found for Costa Rica, where the vaccine was cost-saving with a probability of >90% for VCPI up to $10, derived from the cost-effectiveness acceptability curve (Appendix Figure 5). With the same probability, the highest VCPI under which the vaccine was cost-saving was $25 for Guatemala and Panama. The highest values of VCPI for a cost-saving scenario in other countries were $14–$24.

For positive ICER values, we considered the average per capita GDP of each country in 2015 and 2016 as the threshold for cost-effectiveness (30). For this threshold, the vaccine is very cost-effective with a probability >90% at VCPI of <$16 in Costa Rica (mean incremental cost of $7,352/DALY averted; 95% CI $1,280–$9,234/DALY averted) and <$47 in French Guiana (mean incremental cost of $14,475/DALY averted; 95% CI $10,016–$16,653/DALY averted), with other countries having the highest value of VCPI in this range (Figure 1, red). For the threshold of 3 times the per capita GDP, the vaccine is still cost-effective (with a probability of >90%) with VCPI up to $24 (mean incremental cost of $4,829/DALY averted; 95% CI $2,395–$6,068/DALY averted) in Nicaragua and $96 (mean incremental cost of $49,934/DALY averted; 95% CI $36,523–$53,661/DALY averted) in French Guiana, with other countries having the highest value of VCPI in this range (Figure 1, black). We determined the VCPI for scenarios that are cost-saving, very cost-effective, and cost-effective for each country (Table), the corresponding incremental cost per DALY averted with 95% CIs (Table; Appendix Table 7), and the associated cost-effectiveness acceptability curves (Appendix Figure 5).

We also calculated the reduction of fetal microcephaly during pregnancy by comparing the simulation scenarios in the presence and absence of vaccination. We found a marked reduction in cases of microcephaly, within the range of 74%–92%, attributable to vaccination; the median percentage reduction was >80% in all countries (Figure 2). This finding suggests that a Zika virus vaccine with a prophylactic efficacy as low as 60% could substantially reduce the incidence of microcephaly.

Given that the attack rates in future outbreaks may be different from those estimated for the 2015–2017 outbreaks, we further conducted cost-effectiveness analysis for 2 additional scenarios (Appendix Table 8). In the first scenario, we considered an increase of 4% in the estimated attack rate for each country. We found that vaccination was very cost-effective with a probability >90% at VCPI of <$20 in Nicaragua (mean incremental cost of $1,067/DALY averted) and <$50 or less in French Guiana (mean incremental cost of $14,914/DALY averted). The highest VCPI for other countries ranged between these values.

In the second scenario, we decreased the attack rates by 4%, with a lower bound of 1% for each country. The results show that vaccination was very cost-effective, with a VCPI of <$4 in Mexico (mean incremental cost of $3,054/DALY averted) and <$41 in French Guiana (mean incremental cost of $15,037/DALY averted), with other countries having the highest VCPI value in this range (summary of additional results of cost-effectiveness analysis in Appendix Tables 9, 10, and Appendix Figures 6, 7). The median percentage reduction of microcephaly in these scenarios was >75% with vaccination (Appendix Figure 8).

We determined the VCPI within the input range of $2–$100, for which vaccination is cost-saving (when ICER values are negative) and is very cost-effective (when ICER values are positive, below the threshold of the per capita GDP) for 18 countries in the Americas. Although several factors (e.g., the level of preexisting herd immunity, attack rate, costs associated with the management of Zika virus infection and its outcomes, and the willingness to pay) are critical in determining VCPI for cost-effectiveness, our results show that targeted vaccination of women of reproductive age would be cost-effective, and even cost-saving, in all countries studied if VCPI is sufficiently low. Furthermore, vaccination with a protection efficacy of 60%–90% notably reduces the incidence of microcephaly, with a median percentage reduction >75% in simulated scenarios.

Previous work suggests that a prophylactic vaccine with a protection efficacy of 75% reduces the incidence of prenatal infections by >94% if 90% of women of reproductive age are vaccinated (31). These estimates are slightly higher than what our model predicts (with a median percentage reduction of 75%–88%) in similar scenarios, which is expected given the deterministic nature of the model used in the previous study (31). Nevertheless, the findings indicate that targeted vaccination is a noteworthy preventive measure for mitigating the impact of Zika virus infection in future outbreaks.

Considering direct medical costs associated with short- and long-term Zika virus infection outcomes, our study provides a cost-effectiveness analysis of a Zika virus vaccine candidate from a government perspective. Several recent modeling studies also evaluate cost-effectiveness of a Zika virus vaccine (20,32). However, these studies have either considered only a few countries in Latin America or relied on homogeneous models. The strength of our study relies on the evaluation of cost-effectiveness for countries affected by Zika virus with estimated attack rates >2% within a single modeling framework. We based our analysis on an individual-level stochastic approach, accounting for parameter uncertainty and heterogeneities in disease transmission. Because of its dynamic nature, the simulation model also considers the accruing herd immunity during the epidemic that results from the indirect protection effects of naturally acquired immunity in the population.

Our results should be considered within the context of study limitations. First, we note that we based our analysis on estimates of attack rates during the 2015–2017 Zika virus outbreaks in Latin and South America countries (9,17,18), and these attack rates were regarded as the levels of preexisting herd immunity in the simulations. Should these levels change as the result of a decline of herd immunity or accumulation of new susceptible persons at the time of vaccine availability in future outbreaks, the expected changes in the VCPI range for cost-effectiveness require further analysis. Second, although the initial phase of clinical trials indicates high levels of neutralizing antibodies (14,15), the range of vaccine efficacy has not been ascertained; our estimates rely on the assumption that a single dose of vaccine would provide a protection efficacy of 60%–90%. We assumed that during the epidemic pregnant women are vaccinated (with a coverage of 80%) early in their first trimester, because the highest risk of microcephaly occurs then. However, we understand that because of various factors, including access to healthcare resources and late recognition of pregnancy, vaccination may not occur before any potential Zika virus infection during pregnancy. The risk for microcephaly was not altered if infection occurred following vaccination, but the disease was considered to be asymptomatic. The validation of these assumptions requires efficacy data from clinical trials, which are currently lacking. In our model, the risk of sexual transmission was included only during the infectious period. Although this risk may continue for several days or weeks following recovery (33,34), our simplifying assumption is justified because of uncertainty in the duration of sexual transmission at the individual level. Despite these limitations, which warrant further investigation as relevant information and data become available, this study provides estimates for Zika virus vaccine cost-effectiveness to inform decision makers for the implementation of the VTPP strategies in an outbreak response scenario.

Additional information about the model for cost-effectiveness of Zika virus vaccine.