Health and economic benefits of early vaccination and nonpharmaceutical interventions for a human influenza A (H7N9) pandemic: a modeling study

Abstract

Background: Vaccination for the 2009 pandemic did not occur until late in the outbreak, which limited its benefits. Influenza A (H7N9) is causing increasing morbidity and mortality in China, and researchers have modified the A (H5N1) virus to transmit via aerosol, which again heightens concerns about pandemic influenza preparedness.

Objective: To determine how quickly vaccination should be completed to reduce infections, deaths, and health care costs in a pandemic with characteristics similar to influenza A (H7N9) and A (H5N1).

Design: Dynamic transmission model to estimate health and economic consequences of a severe influenza pandemic in a large metropolitan city.

Data sources: Literature and expert opinion.

Target population: Residents of a U.S. metropolitan city with characteristics similar to New York City.

Time horizon: Lifetime.

Perspective: Societal.

Intervention: Vaccination of 30% of the population at 4 or 6 months.

Outcome measures: Infections and deaths averted and cost-effectiveness.

Results of base-case analysis: In 12 months, 48 254 persons would die. Vaccinating at 9 months would avert 2365 of these deaths. Vaccinating at 6 months would save 5775 additional lives and $51 million at a city level. Accelerating delivery to 4 months would save an additional 5633 lives and $50 million.

Results of sensitivity analysis: If vaccination were delayed for 9 months, reducing contacts by 8% through nonpharmaceutical interventions would yield a similar reduction in infections and deaths as vaccination at 4 months.

Limitation: The model is not designed to evaluate programs targeting specific populations, such as children or persons with comorbid conditions.

Conclusion: Vaccination in an influenza A (H7N9) pandemic would need to be completed much faster than in 2009 to substantially reduce morbidity, mortality, and health care costs. Maximizing non-pharmaceutical interventions can substantially mitigate the pandemic until a matched vaccine becomes available.

Primary funding source: Agency for Healthcare Research and Quality, National Institutes of Health, and Department of Veterans Affairs.

Figure 1. Infections and deaths per day depending on timing of vaccination and case-fatality

At high case-fatality proportions, individuals begin to reduce social interactions in response to increasing mortality, reducing the number of infections per day, but lengthening the epidemic. As this reactive social distancing decreases mortality, individuals resume contacts, leading to sequential pandemic waves over time.. The lines in this figure demonstrate deaths per day (the areas under the curves represent cumulative deaths). Deaths per day are generally higher for vaccination at 6 months than those for vaccination at 4 months; however, with a case-fatality proportion of 10%, deaths per day for vaccination in 6 months drops below deaths per day for vaccination in 4 months after day 240. This occurs because by that point, a greater number of individuals in the 6-month vaccination category have been infected and developed immunity following infection, leading to fewer susceptible people and therefore fewer deaths. We see this only under the 10% case-fatality proportion, because as individuals reduce social interactions due to higher mortality, the epidemic lasts longer and there is still sustained influenza transmission at day 240 and beyond. The total cumulative deaths under the policy of vaccination in 6 months (represented by the total area under the curves) are always equal to or greater than those under the policy of vaccination in 4 months. The same reasoning explains daily death rates for vaccination at 9 months dropping below vaccination in 6 months later in the epidemic.

Figure 2. Infections and deaths per day with increasing non-pharmaceutical interventions

In case of delays to mass vaccination, public health officials could announce and implement the most restrictive NPIs (e.g., school closures or home quarantines) to mitigate the pandemic while awaiting vaccination. Increasing NPIs to 90% over pandemic days 30 to 60 would delay the first wave to 6 months; increasing to 90% over days 45 to 105 would delay the first wave to 9 months. The lines in this figure demonstrate deaths per day (the areas under the curves represent cumulative deaths). Deaths per day for vaccination in 6 months are generally higher than those for vaccination in 4 months; however, in the case of 90% reduction in contacts due to NPIs, the daily deaths for vaccination in 6 months decrease below the deaths per day for vaccination in 4 months later in the epidemic. This occurs because by that point, a greater number of individuals in the 6-month vaccination category have been infected and developed immunity following infection, leading to fewer susceptible people and therefore fewer deaths. The total cumulative deaths under the policy of vaccination in 6 months (represented by the total area under the curves) are always equal to or greater than those under the policy of vaccination in 4 months. The same reasoning explains daily death rates for vaccination at 9 months dropping below vaccination in 6 months later in the epidemic.