There were some advances that proposed some tools to track epidemic trends using cell phones or social (unstructured) data. Which are the most promising advances on this area? What needs to be improved?
That wouldn't be scientifically established, unless we know the previous prevalence of the said disorder and now we need to confirm increase in number of cases. To label Epidemic we want to see an increasing trend.
Most published work in this has been on predicting influenza using either Internet searches (Google Flu Trends, which has now stopped, see http://j.mp/1vWuOVp) or using Twitter (see Paul et al, PMC4234396). For me, it's not clear whether epidemic trends of respiratory or gastro-intestinal infections can be reliably traced from social media behaviour or content. More promising would be if smart phone could directly detect coughing/sneezing or measure the number of times one visits the toilet.
The role of internet searches and social media posts have more to do with 'early detection' of epidemics before slower case reports trickle in from providers and get sorted by surveillance officers. This may serve as an early warning providing some lead time for response and may even help inform active surveillance on specific disease symptoms look out for.
I appreciate your feedback. I also believe it might serve as a "detector" rather than a "predictor".
- Badges
- Science topic