Comparison of static and dynamic models of maternal immunization to prevent infant pertussis in Brazil

•

Dynamic transmission models of infectious disease capture the herd immunity effects of vaccination.

•

We compared dynamic and static models of maternal acellular pertussis (aP) immunization built with Brazilian data.

•

At infant vaccine coverage < 90–95%, both models estimate that maternal immunization is cost-effective.

•

Only the dynamic model shows that maternal immunization is not cost-effective at infant coverage > 90–95%.

•

The background effect of routine infant vaccination is critical to the cost-effectiveness of maternal aP immunization.

Background

This paper compares cost-effectiveness results from two models of maternal immunization to prevent pertussis in infants in Brazil, one static, one dynamic, to explore when static models are adequate for public health decisions and when the extra effort required by dynamic models is worthwhile.

Methods

We defined two scenarios to explore key differences between static and dynamic models, herd immunity and time horizon. Scenario 1 evaluates the incremental cost/DALY of maternal acellular pertussis (aP) immunization as routine infant vaccination coverage ranges from low/moderate up to, and above, the threshold at which herd immunity begins to eliminate pertussis. Scenario 2 compares cost-effectiveness estimates over the models’ different time horizons. Maternal vaccine prices of $9.55/dose (base case) and $1/dose were evaluated.

Results

The dynamic model shows that maternal immunization could be cost-saving as well as life-saving at low levels of infant vaccination coverage. When infant coverage reaches the threshold range (90–95%), it is expensive: the dynamic model estimates that maternal immunization costs $2 million/DALY at infant coverage > 95% and maternal vaccine price of $9.55/dose; at $1/dose, cost/DALY is $200,000. By contrast, the static model estimates costs/DALY only modestly higher at high than at low infant coverage. When the models’ estimates over their different time horizons are compared at infant coverage < 90–95%, their projections fall in the same range.

Conclusions

Static models may serve to explore an intervention’s cost-effectiveness against infectious disease: the direction and principal drivers of change were the same in both models. When, however, an intervention too small to have significant herd immunity effects itself, such as maternal aP immunization, takes place against a background of vaccination in the rest of the population, a dynamic model is crucial to accurate estimates of cost-effectiveness. This finding is particularly important in the context of widely varying routine infant vaccination rates globally.

Clinical Trial registry

Clinical Trial registry name and registration number: Not applicable.

Cost-effectiveness of maternal GBS immunization in low-income sub-Saharan Africa

BACKGROUND

A maternal group B streptococcal (GBS) vaccine could prevent neonatal sepsis and meningitis. Its cost-effectiveness in low-income sub-Saharan Africa, a high burden region, is unknown.

METHODS

We used a decision tree model, with Markov nodes to project infants' lifetimes, to compare maternal immunization delivered through routine antenatal care with no immunization. 37 countries were clustered on the basis of economic and health resources and past public health performance. Vaccine efficacy for covered serotypes was ranged from 50% to 90%. The model projected EOGBS (early-onset) and LOGBS (late-onset) cases and deaths, disability-adjusted life years (DALYs), healthcare costs (2014 US$), and cost-effectiveness for a representative country in each of the four clusters: Guinea-Bissau, Uganda, Nigeria, and Ghana. Maximum vaccination costs/dose were estimated to meet two cost-effectiveness benchmarks, 0.5 GDP and GDP per capita/DALY, for ranges of disease incidence (reported and adjusted for under-reporting) and vaccine efficacy.

RESULTS

At coverage equal to the proportion of pregnant women with≥4 antenatal visits (ANC4) and serotype-specific vaccine efficacy of 70%, maternal GBS immunization would prevent one-third of GBS cases and deaths in Uganda and Nigeria, where ANC4 is 50%, 42-43% in Guinea-Bissau (ANC4=65%), and 55-57% in Ghana (ANC4=87%). At a vaccination cost of $7/dose, maternal immunization would cost $320-$350/DALY averted in Guinea-Bissau, Nigeria, and Ghana, less than half these countries' GDP per capita. In Uganda, which has the lowest case fatality ratios, the cost would be $573/DALY. If the vaccine prevents a small proportion of stillbirths, it would be even more cost-effective. Vaccination cost/dose, disease incidence, and case fatality were key drivers of cost/DALY in sensitivity analyses.

CONCLUSION

Maternal GBS immunization could be a cost-effective intervention in low-income sub-Saharan Africa, with cost-effectiveness ratios similar to other recently introduced vaccines. The vaccination cost at which introduction is cost-effective depends on disease incidence and vaccine efficacy. Clinical Trial registry name and registration number: Not applicable.

What Pertussis Mortality Rates Make Maternal Acellular Pertussis Immunization Cost-Effective in Low- and Middle-Income Countries? A Decision Analysis

BACKGROUND

 Despite longstanding infant vaccination programs in low- and middle-income countries (LMICs), pertussis continues to cause deaths in the youngest infants. A maternal monovalent acellular pertussis (aP) vaccine, in development, could prevent many of these deaths. We estimated infant pertussis mortality rates at which maternal vaccination would be a cost-effective use of public health resources in LMICs.

METHODS

 We developed a decision model to evaluate the cost-effectiveness of maternal aP immunization plus routine infant vaccination vs routine infant vaccination alone in Bangladesh, Nigeria, and Brazil. For a range of maternal aP vaccine prices, one-way sensitivity analyses identified the infant pertussis mortality rates required to make maternal immunization cost-effective by alternative benchmarks ($100, 0.5 gross domestic product [GDP] per capita, and GDP per capita per disability-adjusted life-year [DALY]). Probabilistic sensitivity analysis provided uncertainty intervals for these mortality rates.

RESULTS

 Infant pertussis mortality rates necessary to make maternal aP immunization cost-effective exceed the rates suggested by current evidence except at low vaccine prices and/or cost-effectiveness benchmarks at the high end of those considered in this report. For example, at a vaccine price of $0.50/dose, pertussis mortality would need to be 0.051 per 1000 infants in Bangladesh, and 0.018 per 1000 in Nigeria, to cost 0.5 per capita GDP per DALY. In Brazil, a middle-income country, at a vaccine price of $4/dose, infant pertussis mortality would need to be 0.043 per 1000 to cost 0.5 per capita GDP per DALY.

CONCLUSIONS

 For commonly used cost-effectiveness benchmarks, maternal aP immunization would be cost-effective in many LMICs only if the vaccine were offered at less than $1-$2/dose.