Internet-based surveillance of Influenza-like-illness in the UK during the 2009 H1N1 influenza pandemic

Abstract

Background: Internet-based surveillance systems to monitor influenza-like illness (ILI) have advantages over traditional (physician-based) reporting systems, as they can potentially monitor a wider range of cases (i.e. including those that do not seek care). However, the requirement for participants to have internet access and to actively participate calls into question the representativeness of the data. Such systems have been in place in a number of European countries over the last few years, and in July 2009 this was extended to the UK. Here we present results of this survey with the aim of assessing the reliability of the data, and to evaluate methods to correct for possible biases.

Methods: Internet-based monitoring of ILI was launched near the peak of the first wave of the UK H1N1v influenza pandemic. We compared the recorded ILI incidence with physician-recorded incidence and an estimate of the true number of cases over the course of the epidemic. We also compared overall attack rates. The effect of using different ILI definitions and alternative denominator assumptions on incidence estimates was explored.

Results: The crude incidence measured by the internet-based system appears to be influenced by individuals who participated only once in the survey and who appeared more likely to be ill. This distorted the overall incidence trend. Concentrating on individuals who reported more than once results in a time series of ILI incidence that matches the trend of case estimates reasonably closely, with a correlation of 0.713 (P-value: 0.0001, 95% CI: 0.435, 0.867). Indeed, the internet-based system appears to give a better estimate of the relative height of the two waves of the UK pandemic than the physician-recorded incidence. The overall attack rate is, however, higher than other estimates, at about 16% when compared with a model-based estimate of 6%.

Conclusion: Internet-based monitoring of ILI can capture the trends in case numbers if appropriate weighting is used to correct for differential response. The overall level of incidence is, however, difficult to measure. Internet-based systems may be a useful adjunct to existing ILI surveillance systems as they capture cases that do not necessarily contact health care. However, further research is required before they can be used to accurately assess the absolute level of incidence in the community.

Figure 1

Comparison of the flusurvey population in different age and gender categories with the UK population [22].

Figure 2

Proportion of individuals in an influenza risk group, comparing the population of England with English Flusurvey participants [22,23].

Figure 3

Comparison of weekly ILI incidence as calculated by the RCGP with the weekly HPA case estimates.

Figure 4

Total number of reports per participant in the crude (complete) dataset.

Figure 5

Time series of the proportion of participants reporting ILI each week (using the GIS definition of ILI). Three different flusurvey incidence curves are plotted: one that uses the crude (complete) dataset, one using the censored dataset (ignoring all participants' who participated only once) and one using the censored dataset and reweighting the population to account for demographic unrepresentativeness, compared with the cases estimated by the HPA [21]. All curves are plotted using a 3-week moving average; plots without this moving average can be found in the Supplementary material.

Figure 6

Comparison of the weekly symptoms questionnaires that were returned with the number of ILI reports for the same week, in the crude (complete) dataset.

Figure 7

Comparison of ILI incidence according to different definitions of ILI with the HPA's estimated ILI cases [17,21,27,28]. As in Fig 3, the denominator used is those participants who completed the symptoms questionnaire on the week in question. All curves are plotted using a 3-week moving average; plots without this moving average can be found in the Supplementary material.

Figure 8

The estimated attack rates using 3 different definitions of ILI, broken down into age groups, along with estimated attack rates by the HPA [23]. The attack rates for the HPA estimated cases have been calculated based on Baguelin et al. [23] method.

Figure 9

Incidence according to different denominators compared with the case estimates of the HPA [21]. All curves are plotted using a 3-week moving average; plots without this moving average can be found in the Supplementary material.