Background: Annual influenza infection can cause severe disease, especially among adults ≥65 years of age. Vaccination is an effective prevention strategy recommended annually by the US Advisory Committee on Immunization Practices. While the recommendation is to get vaccinated by October, the optimal vaccination timing may be later if vaccine effectiveness wanes over time.

Methods: We conducted a population-based cohort study within Kaiser Permanente Northern California (KPNC) members aged ≥65 years during the 2023-2024 influenza season to identify optimal timing of vaccination between Sept-Nov before influenza begins to circulate. Using polymerase chain reaction (PCR)-confirmed influenza infection as the outcome, we estimated the risk and relative risk (RR) of infection by April 30, 2024 if everyone in the study population was vaccinated on week 37 (early Sept) or week 46 (mid-Nov). We used longitudinal targeted maximum likelihood estimation (LTMLE) to model time-varying vaccination, censoring, and the outcome, adjusting for baseline demographics and time-varying comorbidities weekly.

Results: Among 736,194 members in the study population, there were 3,389 PCR-confirmed influenza cases (0.46%). Based on preliminary LTMLE analyses which estimated the counterfactual cumulative influenza risk, assuming the full cohort was vaccinated on a particular week, the risk was 0.46% at week 37 and 0.33% at week 46 (Figure 1). When compared with being unvaccinated, the RR of being vaccinated at week 37 was 0.78 (95% CI: 0.62, 0.97) and 0.55 (95% CI: 0.40, 0.74) at week 46. When compared with being vaccinated at week 37, the RR of having influenza if vaccinated at week 46 was 0.70 (95% CI: 0.49, 1.01).

Conclusion: The risk of influenza infection was relatively lower if individuals were vaccinated in November versus in September. The lower risk of influenza infection associated with being vaccinated in November suggests that being vaccinated later during the season may be more protective against influenza.