Monitoring influenza vaccine effectiveness using the national influenza surveillance system

Background: Flu vaccine composition is reformulated on a yearly basis. As such, the vaccine effectiveness (VE) from previous seasons cannot be considered for subsequent years, and it is necessary to monitor the VE for each season. This study (MonitorEVA- monitoring vaccine effectiveness) intends to evaluate the feasibility of using the national influenza surveillance system (NISS) for monitoring the influenza VE. Material and methods: Data was collected within NISS during 2004 to 2014 seasons. We used a case-control design where laboratory confirmed incident influenza like illness (ILI) patients (cases) were compared to controls (ILI influenza negative). Eligible individuals consisted on all aged individuals that consult a general practitioner or emergency room with ILI symptoms with a swab collected within seven days of symptoms onset. VE was estimated as 1- odds ratio of being vaccinated in cases versus controls adjusted for age and month of onset by logistic regression. Sensitivity analyses were conducted to test possible effect of assumptions on vaccination status, ILI definition and timing of swabs (<3 days after onset). Results: During the 2004-2014 period, a total of 5302 ILI patients were collected but 798 ILI were excluded for not complying with inclusion criteria. After data restriction the sample size in both groups was higher than 148 individuals/ season; minimum sample size needed to detect a VE of at least 50% considering a level of significance of 5% and 80% power. Crude VE point estimates were under 45% in 2004/05, 2005/06, 2011/12 and 2013/14 season; between 50%-70% in 2006/07, 2008/09 and 2010/11 seasons, and above 70% in 2007/08 and 2012/13 season. From season 2006/07 to 2013/14, all crude VE estimates were statistically significant. After adjustment for age group and month of onset, the VE point estimates decreased and only 2008/09, 2012/13 and 2013/14 seasons were significant. Discussion and Conclusions: MonitorEVA was able to provide VE estimates for all seasons, including the pandemic, indicating if the VE was higher than 70% and less than 50%. When comparing with other observational studies, MonitorEVA estimates were comparable but less precise and VE estimates were in accordance with the antigenic match of the circulating virus/ vaccine strains. Given the sensitivity results, we propose a MonitorEVA based on: a) Vaccination status defined independently of number of days between vaccination and symptoms onset; b) use of all ILI data independent of the definition; c) stratification of VE according to time between onset and swab (< 3 and ≥3 days).