Researchers studying the health of large populations are often up against some difficult challenges involved with lack of resources for gathering data, time lags in reporting data, and lack of reporting by patients.
However, an innovative and revealing approach is the passive collection and assessment of digital data such as that generated when the general public searches online for health information.
One study, published recently in the American Journal of Preventive Medicine, employed this approach to investigate seasonal patterns in online searches for mental health information.
The study involved monitoring all mental health queries submitted to Google in the United States and Australia between 2006 and 2010, subdivided according to the following search terms: ADHD (attention deficit-hyperactivity disorder), anxiety, bipolar, depression, anorexia or bulimia, OCD (obsessive-compulsive disorder), schizophrenia, and suicide. The average (mean) search volume seen during winter was then compared with that seen during summer.
The study found that overall mental health searches followed seasonal patterns, peaking in the winter and falling to their lowest during the summer, with a 14% difference in volume between seasons in the United States and an 11% difference in Australia. Similar patterns were also seen for all subcategories, ranging from a 7% seasonal difference for anxiety to a 37% seasonal difference for eating disorder queries in the United States.
The authors note that they investigated and rejected several possible nonclinical factors that may have provided causality for such seasonality, including media trends and academic interest.
They also point out that these patterns of seasonality would have gone unnoticed under traditional methods of surveillance.
Similar studies using search engine traffic and social media content have revealed more about other conditions. For example, an analysis of Google searches on ‘migraine’ revealed that more traffic occurs during weekdays than on weekends and holidays, with Mondays showing higher occurrences than Friday.