"How to quite smoking" and "how to lose weight," are the kind of terms a person with specific health concerns might type into Google, but according to new research, they could be indicative of a more general, and more serious, medical condition. By analyzing data from Google Trends, scientists in the UK claim that they can more effectively monitor symptoms and therefore the emergence of Type 2 diabetes, something that could help public health officials tackle the disease.

Type 2 diabetes is a lifestyle disease that can be brought on by a mix of factors such as diet, exercise and family history of diabetes. But diagnosing it is far from straightforward, typically involving things like tests of blood, urine and glucose, along with physical exams. Surveillance models meanwhile, track things like age, gender and lifestyle to identify those at risk of developing it, but these are far from foolproof.

So scientists have been looking for more effective ways to pick up the precursors to the disease. More and more, people are using Google as a way of checking on their health. According to researchers at the University of Warwick, 21.8 percent of people in Britain chose to self-diagnose illnesses using the internet rather than a doctor in 2015.

So the researchers find themselves with a growing pool of data they say can be leveraged as a tool to identify the emergence of type 2 diabetes in certain areas. They did this by tapping into data from Google Trends, looking specifically for key words entered into search engines and posted on social media in Central London relating to symptoms of the disease.

They then placed this data alongside two of the UK's principal surveillance models for diabetes, looking for fluctuations in the search terms that correlated with the prevalence of disease in certain areas. The results, the team says, demonstrate that Google Trends can indeed offer real-time insights into how prevalent the disease is in particular areas, and more generally, provide a useful source of data on the health of populations.

"Self-diagnosing behaviors online could be effectively leveraged for real-time health monitoring tools, with the biggest potential to be anticipated for chronic and non-communicable diseases," says the University of Warwick's Nataliya Tkachenko, who led the study. "Unlike quickly spreading diseases (eg. flues), such slowly developing conditions are largely dependent on the personal and community lifestyles, the factors, which are currently unaccounted for in the screening models. Human online behaviors could help to bridge the gap between real-world human health landscape and synthetic, predominantly bio-centric monitoring tools."

The research was published in the journal Scientific Reports.