Eurosurveillance - Volume 25, Issue 21, 28 May 2020

Despite the early development of Google Flu Trends in 2009, standards for digital epidemiology methods have not been established and research from European countries is scarce.

In this article, we study the use of web search queries to monitor influenza-like illness (ILI) rates in the Netherlands in real time.

In this retrospective analysis, we simulated the weekly use of a prediction model for estimating the then-current ILI incidence across the 2017/18 influenza season solely based on Google search query data. We used weekly ILI data as reported to The European Surveillance System (TESSY)  each week, and we removed the then-last 4 weeks from our dataset. We then fitted a prediction model based on the then-most-recent search query data from Google Trends to fill the 4-week gap (‘Nowcasting’). Lasso regression, in combination with cross-validation, was applied to select predictors and to fit the 52 models, one for each week of the season.

The models provided accurate predictions with a mean and maximum absolute error of 1.40 (95% confidence interval: 1.09–1.75) and 6.36 per 10,000 population. The onset, peak and end of the epidemic were predicted with an error of 1, 3 and 2 weeks, respectively. The number of search terms retained as predictors ranged from three to five, with one keyword, ‘griep’ (‘flu’), having the most weight in all models.

This study demonstrates the feasibility of accurate, real-time ILI incidence predictions in the Netherlands using Google search query data.

Understanding influenza seasonality is necessary for determining policies for influenza control.

We characterised transmissibility during seasonal influenza epidemics, including one influenza pandemic, in Spain during the 21th century by using the moving epidemic method (MEM) to calculate intensity levels and estimate differences across seasons and age groups.

We applied the MEM to Spanish Influenza Sentinel Surveillance System data from influenza seasons 2001/02 to 2017/18. A modified version of Goldstein’s proxy was used as an epidemiological-virological parameter. We calculated the average starting week and peak, the length of the epidemic period and the length from the starting week to the peak of the epidemic, by age group and according to seasonal virus circulation.

Individuals under 15 years of age presented higher transmissibility, especially in the 2009 influenza A(H1N1) pandemic. Seasons with dominance/co-dominance of influenza A(H3N2) virus presented high intensities in older adults. The 2004/05 influenza season showed the highest influenza-intensity level for all age groups. In 12 seasons, the epidemic started between week 50 and week 3. Epidemics started earlier in individuals under 15 years of age (−1.8 weeks; 95% confidence interval (CI):−2.8 to −0.7) than in those over 64 years when influenza B virus circulated as dominant/co-dominant. The average time from start to peak was 4.3 weeks (95% CI: 3.6–5.0) and the average epidemic length was 8.7 weeks (95% CI: 7.9–9.6).

These findings provide evidence for intensity differences across seasons and age groups, and can be used guide public health actions to diminish influenza-related morbidity and mortality.