Skip to main content
Discover how data is having a big impact on predicting illness
Every year a new device or app is created that changes how we communicate, bank or even find a partner, yet improving medical treatments is often a long process.
The good news is that healthcare is starting to close the gap thanks to data, really big data! How? Having lots of medical data about a person, or groups of people, makes it possible to build accurate medical models and make useful predictions about their health.
For years, companies have been using data from our online activities to predict the movies we might like or products we want to buy. Valid concerns about data security have slowed the advance of big data in medical science. Now, cautiously, those methods are being used to make us healthier due to increased availability of data.
But, where is this data coming from and how do we get more of it? The average Briton visits a doctor just 5 times a year making it difficult for doctors to collect large amounts of data. The explosion in wearable health devices is one route. A smart device is constantly tracking vital health indicators and as they evolve from recreational to research grade, the data will become more valuable. We will eventually all have a complete health profile stored in our devices.
Even in these early stages the benefits of big data are being seen. Hitachi, in collaboration with University of Utah Health, announced a new technology that analyses electronic medical records to predict the probability of different medications treating patients with type-2 diabetes mellitus. The high level accuracy of the technology in predicting the effect has been confirmed by comparing the results to past medical records.
As well has having lots of data on one person, having data from lots of people is also very beneficial for our understanding of disease. One of the earliest examples of this was Google Flu Trends, a project set up by the tech giant to map cold and flu outbreaks. The data, gleaned from users searching for cold and flu symptoms, remedies and doctors’ appointments within a geographical area, was then used to map future infections. This idea has been developed further by a team from Northeastern University who developed a new model to predict the spread of flu in real time using Twitter.
Other examples include tracking the sales of over-the-counter cold and flu remedies in US pharmacies, allowing public health officials to predict spikes in hospital admissions for respiratory issues, which come on average 2 weeks following the initial purchase.
Although the healthcare industry must remain wary of patient data breaches, the potential for understanding more about our health will ultimately mean more people being healthier in society.