What’s been deciphered so far?

Before launching the Flu project across the country we wanted to test our data collection process.  One school kindly volunteered their illness absence data for the whole of last year, which is plotted in the graph.

Here’s a quick explanation of what the graph shows:

  • X axis: Week
  • Primary Y axis (Left):  Shows national flu levels based on data from the Royal College of General Practitioners (RCGP).  It measures the number of patients showing Influenza Like Illness per 100,000 people and is currently one of the best available indicators for flu activity (you can read more about it here).
  •  Secondary Y axis (Right): Shows the the number of half days missed due to illness per 100 students at the school

When I first plotted this graph I got very excited. It shows exactly what we are hoping to see!   The school absence data (red) peaks three weeks before the national flu data (blue).   This means school absence data could act as an early indicator for a flu outbreak, giving more time to make essential preparations.

However, before we get too carried away, it’s important to remember the biggest limitation of this analysis; we’ve only used data for one school. This means that the red line you see is not very accurate.  The accuracy of the data will improve with each school involved in the project.

The other issue is that it’s much easier to spot the peak when looking at the data in one go. When we run the project live, we will only have the data up until that point in time.  This makes the analysis more interesting but also more difficult to carry out. One of the things that will help us spot a peak will be more school absence data for the academic year so far. If you have time we would be extremely grateful if you could submit as much of your school illness data as possible. I know that’s a bit tedious but it will mean that we have a better idea of what makes a real peak if we have a baseline to compare to everything to.


Posted on by Dr Rob | Leave a comment

Leave a Comment