Transmission patterns during the 2009 influenza A(H1N1) pandemicArchived

ECDC comment

The authors of this study from 12 European Union member states used statistical modelling and a novel European approach for combining data to explore the potential causes that might explain difference in transmission dynamics observed during the early stages of the 2009 influenza A(H1N1) pandemic.  

Different transmission patterns in the early stages of the influenza A(H1N1)v pandemic: A comparative analysis of 12 European countries

Flasche S, Hens N, Boëlle P-Y, et al. Epidemics 3(2): 125-133; 2011

Data were gathered by sending out a model and data-gathering instrument to 12 member states.(1) The authors considered six potential causes or hypotheses that could have regulated the transmission of the 2009 pandemic at that time across Europe and tested these against available data from 12 European countries. In order to do so, they analysed data on the time of symptom onset of laboratory confirmed cases over a period from April 2009 to October 2009, and data on the control measures that were implemented, school holidays and the weather.

They investigated six hypothetical factors which could have potentially regulated the transmission:

  1. Extinction by chance, i.e. that the epidemiological parameters were similar across countries but the epidemic took off by chance in the United Kingdom (UK) only, but not in other countries. A major generalised epidemic of influenza A(H1N1) 2009 occurred in the UK in June and July 2009, which only declined once schools closed for the summer holidays at the end of July. No such major epidemic was reported elsewhere in Europe until the autumn, when the UK had its second wave of infection. The authors argue that the observed epidemiology suggests a much higher transmissibility (measured by the average number of secondary cases per case) of influenza A(H1N1) 2009 in England than elsewhere in Europe during the summer.
  2. Relative susceptibility of adults and the elderly compared to children.
  3. Difference in the age distribution of imported cases.
  4. Differences in contact patterns could have favoured the spread in age groups which mix with the more susceptible age groups.
  5. Difference in implemented control measures.
  6. Late school closure relative to the time of the introduction in the UK.
  7. Weather patterns (pressure, relative and absolute humidity, temperature and wind speed) might have played an influential role.

The authors found that no single factor was able to explain the differences and an additive mixed model was used to model the country-specific weekly estimates of the effective reproductive number using the extinction probability, school holidays and weather patterns as explanatory variables. The authors performed a regression analysis and the effective reproduction number was calculated from the laboratory confirmed cases by an extended version of a method previously described (2) which uses the serial interval to assign a most likely source case for each case that was not imported and therefore estimates the average number of secondary cases caused by an individual (effective reproduction number). The authors note that the average extinction probability, its trend and the trend in absolute humidity were found to be significantly negatively correlated with the effective reproduction number (although they could only explain about 3% of the variability in the model). In addition, by comparing the initial epidemiology of influenza A (H1N1) 2009 across different European countries, their analysis was able to uncover a possible role for the timing of importations (extinction probability), mixing patterns including in schools and absolute humidity as underlying factors.

The authors point out that these results tentatively suggest that a possible explanation for the difference in epidemiology was that the UK had a relatively large number of early importations, transmission dynamics which were favouring the spread in school-aged children (in England the Schools remain open later into the summer than in any other country) and a relatively low level of absolute humidity. Moreover, they stress that the current understanding of the transmission of influenza, in terms of distribution of susceptibility within the population and observed contact patterns assuming equal effectiveness of control measures, does not fully account for the observed differences in the initial epidemiology of influenza A(H1N1) 2009 in Europe.

ECDC Comment (10th May 2011):What is a novel feature here is that the authors used an especially innovative approach to data gathering and perhaps as a consequence they were unusually successful in gathering  information.  The normal approach to this kind of multi-country modelling work is to ask countries to send data centrally where they are fitted into a model.  Here the approach was the other way round with the initiator of the project sending out a model which was especially user-friendly and allowed national representatives to analyse their own data as well as sending it centrally. Normally the send your data to us is not successful at any speed. This approach of sending out the model was especially successful in rapidly earning the confidence of researchers and authorities in 12 countries. 


  1. Flasche S, Hens N, Boëlle P-Y, et al. Different transmission patterns in the early stages of the influenza A(H1N1)v pandemic: A comparative analysis of 12 European countries Epidemics 3(2): 125-133; 2011 supplementary data at doi:10.1016/j.epidemi.2011.03.005.
  2. Wallinga, J., Teunis, P., 2004. Different epidemic curves for severe acute respiratory syndrome reveal similar impacts of control measures. Am. J. Epidemiol. 160, 509–516.