MEDICAL MATTERS:For all the research and evidence, nobody really knows where the flu comes from, writes Dr. Muris Houston
"I had a little bird; its name was Enza,
I opened the window and in-flu-Enza"
Skipping rhyme 1918
Will this be the winter of a flu pandemic? Statistically, we are overdue a "bad" flu epidemic, whatever about a pandemic on the scale of the 1918 Spanish flu.
Infectious disease epidemiologists are a little like those involved in predicting earthquakes or volcanic eruptions: they can say a catastrophic event is statistically likely, but really have no method of predicting when exactly it might happen.
Step forward the search engine Google. It has developed a web tool called Google Flu Trends which, it is hoped, could give us about 10 days' advance warning of a flu outbreak.
At present, the Health Protection Surveillance Centre (HPSC) taps into a network of 52 GP practices around the State that regularly sends data on the number of patients who consult with an influenza-like illness.
The doctors also send a combined throat and nose swab from a selection of their patients to the National Virus Reference Laboratory, where these are tested for the influenza virus. And regional departments of public health report on influenza activity in their area on a weekly basis.
An analysis of last winter's data, published in Epi-Insight, the HPSC disease surveillance report, shows influenza activity was generally low and peaked moderately early in the flu season. Influenza A of the H1N1 subtype was predominant early in the season, followed by Influenza B later on.
So how might Google add to this already comprehensive system? Its unique selling point is a promise to alert public health specialists before people even bring their symptoms to the doctor.
Google Flu Trends is based on the idea that people who are feeling sick turn to the web for information. By tracking searches for phrases like "flu symptoms" or "muscle aches", the web tool builds up a picture of regional queries that could be used to predict a disease outbreak. Early trials of the system suggest it might be able to predict flu outbreaks at least a week before they appear on the US Centre for Disease Control's (CDC) radar.
The idea is an example of predictive forecasting, based on the immense amount of data already collected by search engines. For the flu trends project, Google's engineers came up with a sample of key phrases related to influenza, such as thermometer and chest congestion.
They then analysed five years of information from the Google database and superimposed this data on to the CDC's flu reports for the same five years. Google says it found an almost perfect fit between its predictive data and the actual CDC data for influenza-like illness.
A separate study by epidemiologists from the University of Iona, College of Medicine, is believed to show a similar link between data collected by Yahoo and the incidence of influenza.
In theory, the concept could also be applied to other disease trends. The sheer volume of information available to the major search engines is considered especially powerful because keywords and phrases typed into search engines represent people's immediate intentions.
Last year, Microsoft researchers found that almost 2 per cent of all web searches were health-related.
And researchers from University College Cork have just announced a project to tap into what they call "the wisdom of crowds". Dr Dylan Evans, lecturer in behavioural studies at the School of Medicine will study so-called "prediction markets" to forecast key indications of public health.
The idea is that the collective knowledge of many people, each with a different perspective, will be more accurate than that of individual experts. Ultimately, it is hoped to predict the number of people who will catch the flu here every year.
Each of these initiatives is most interesting. But perhaps we shouldn't hold our breath. Despite all the scientific progress made since 1918, we still do not know where Enza came from, where it went, or what guise it may fly in again.
• Dr Houston is pleased to hear from readers at mhouston@ irishtimes.com but regrets he is unable to reply to individual medical queries