Cost-Effectiveness of Alternative Uses of Polyvalent Meningococcal Vaccines in Niger: An Agent-Based Transmission Modeling Study

Abstract

Background. Despite the introduction of an effective serogroup A conjugate vaccine (MenAfriVac™), sporadic epidemics of other Neisseria meningitidis serogroups remain a concern in Africa. Polyvalent meningococcal conjugate (PMC) vaccines may offer alternatives to current strategies that rely on routine infant vaccination with MenAfriVac plus, in the event of an epidemic, district-specific reactive campaigns using polyvalent meningococcal polysaccharide (PMP) vaccines. Methods. We developed an agent-based transmission model of N. meningitidis in Niger to compare the health effects and costs of current vaccination practice and 3 alternatives. Each alternative replaces MenAfriVac in the infant vaccination series with PMC and either replaces PMP with PMC for reactive campaigns or implements a one-time catch up campaign with PMC for children and young adults. Results. Over a 28-year period, replacement of MenAfriVac with PMC in the infant immunization series and of PMP in reactive campaigns would avert 63% of expected cases (95% prediction interval 49%-75%) if elimination of serogroup A is not followed by serogroup replacement. At a PMC price of $4/dose, this would cost $1412 ($81-$3510) per disability-adjusted life-year (DALY) averted. If serogroup replacement occurs, the cost-effectiveness of this strategy improves to $662 (cost-saving, $2473) per DALY averted. Sensitivity analyses accounting for incomplete laboratory confirmation suggest that a catch-up PMC campaign would also meet standard cost-effectiveness thresholds. Limitations. The assumption that polyvalent vaccines offer similar protection against all serogroups is simplifying. Conclusions. The use of PMC vaccines to replace MenAfriVac in routine infant immunization and in district-specific reactive campaigns would have important health benefits and is likely to be cost-effective in Niger. An additional PMC catch-up campaign would also be cost-effective if we account for incomplete laboratory reporting.

Keywords: economic evaluation; meningitis; meningococcal; simulation; vaccine.

Figure 1:

Model structure depicting health states (ovals) and health events (rectangles) for our agent-based model of meningococcal epidemics. Natural death may occur in any state.

Figure 2:

Weekly confirmed meningococcal cases in Niger reported between 2002 to mid-2015, associated to serogroups A, C, W, and X (blue curve), and to serogroups C, W and X (red curve).

Figure 3:

The proposed agent-based model matches the key characteristics of meningococcal epidemics in Niger between 2002 to mid-2015 under the assumption of complete strain replacement. A) Age-distribution of meningococcal meningitis cases in Niger versus the age-distribution of cases generated by the model. B) Estimated meningococcal carriage prevalence in different age groups from carriage survey studies in the African meningitis belt [19] versus the age-specific average carriage prevalence obtained from the model. C) Average of confirmed weekly meningococcal cases observed from 2002 to mid-2015 versus those produced by the model. D) Cosine of the angle (θ)between the vectors of Fourier amplitude for observed and simulated meningitis time-series; cosine of 1 indicates total match in periodicity and cosine of 0 indicates no overlap between the significant periods of two time-series. See Figure S8 in Appendix for the fit of the model under the no strain replacement scenario.

Figure 4:

Comparing the time-series of meningococcal cases observed between 2002-2015 in Niger (black curve) with three simulated trajectories produced by the calibrated model (blue, green, and red curves) under “with strain-replacement” scenario. The periodicity at which simulated epidemics are occurring matches the periodicity of observed epidemics. Figure 3, Figure 5 and Figure S9 show that trajectories generated by our model also match other key properties of meningococcal epidemics in Niger (e.g., age-distribution of cases, age-specific carriage prevalence, average weekly meningococcal incidence at national and district level, and number of districts in each year between 2002-2015 where the threshold of 10 meningitis cases per 100,000 population is exceeded).

Figure 5:

Average weekly N. meningitidis cases in Niger’s districts produced by our model and observed in the data for the complete strain replacement scenario. See Figure S10 in Appendix for the fit of the model under the scenario of no strain-replacement.

Figure 6:

Expected percentage reduction in annual meningococcal cases over a 28-year simulation period for the vaccination strategies described in Table 3 compared to the Base strategy. Bars represent the 95% prediction intervals.

Figure 7:

Expected number of vaccines used per year (over a 28-year simulation period). A) Complete strain replacement scenario. B) No strain replacement scenario. We note that preventive campaigns are implemented only once at the beginning of the projection period. Error bars represent 95% projection intervals (error bars that are shorter than the width of symbols are not shown). PMP: polyvalent meningococcal polysaccharide; PMC: polyvalent meningococcal conjugate.

Figure 8:

Economic evaluation of vaccination strategies described in Table 3 for the complete strain replacement scenario (A, C) and the no strain replacement scenario (B, D). The price of PMP and PMC vaccines are $4 per dose (see Appendix for sensitivity analysis to the vaccine prices). In figures C and D, the expected gain in net monetary benefit (NMB) of a strategy is calculated with respect to the Base strategy. The dashed line in these figures represents the cost-effectiveness threshold of three per capita gross domestic product of Niger which is estimated to be 1,077 USD in 2015 [45].

Figure 9:

Economic evaluation of vaccination strategies described in Table 3 for the complete strain replacement scenario (A, C) and the no strain replacement scenario (B, D) for the scenario where only 50% of meningococcal cases are reported and correctly categorized. The price of PMP and PMC vaccines are $4 per dose. In figures C and D, the expected gain in net monetary benefit (NMB) of a strategy is calculated with respect to the Base strategy. The dashed line in these figures represents the cost-effectiveness threshold of three per capita gross domestic product of Niger which is estimated to be 1,077 USD in 2015 [45].