Estimating the Direct Impact of Influenza Vaccination and Vaccine-Averted Illnesses Using an Extended Static Compartmental Model

A previously developed static compartment model informed influenza program decisions by estimating the direct impact of influenza vaccination. To improve the estimation of the burden averted by vaccination, we extended this model to consider multiple influenza seasons, allow natural and vaccine-induced immunity to wane over time allow influenza vaccine effectiveness (VE) to differ by hospitalized versus non-hospitalized influenza-associated illness, and measure deaths averted. For model demonstration, we created a synthetic Northern Hemisphere-like and year-round influenza activity dataset spanning four seasons among older adults (>65 years) and assumed VE against symptomatic illness of 37% (17%-53%) and VE against hospitalization of 45% (28%-59%). Under Northern Hemisphere-like activity and an existing vaccination program with an average annual vaccine coverage of 33%, vaccination averted 15.1% (95% UI: 12.2%-16.3%) of influenza-associated illnesses across four seasons; the annual prevented fraction for hospitalized deaths ranged from 14.4% (95% UI: 11.7%-16.8%) to 18.5% (95% UI: 14.9%-21.5%). Under a hypothetical year-round vaccination program with an annual coverage of 20%, vaccination averted 4.9% (95% UI: 3.8%-5.4%) of influenza-associated illnesses across four seasons; the annual prevented fraction for hospitalized deaths ranged from 5.5% (95% UI: 4.3%-6.5%) and 5.8% (95% UI: 4.8%-6.6%). These model extensions allow estimation of vaccine-averted burden across multiple seasons and diverse vaccination program settings, providing information to support influenza vaccination decision-making.