Every year more than one million adverse events relating to different drug reactions are reported to the FDA, resulting in the common criticism of slow detection of safety problems. However, the FDA has now approached the possibility of mining search engine data to detect potential adverse drug reactions. By analyzing people’s Google searches, in which one in 20 Google searches are related to health, and common search terms that are associated with drug names, adverse interactions can be revealed. Search data were also more likely to reveal new reactions that occurred much later during treatment, thus overlooked in clinical research.
Another potential application of search engine data is the monitoring of internet searches that may help alert someone to go visit a doctor. Research showed that sequential or grouped internet searches of subtle symptoms can be used to predict an accurate diagnosis.
Obvious limitations to these applications include privacy and noisiness of data. However, as research continues to advance and machine learning becomes more sophisticated, these applications may become more realistic.
Written by: Samantha Chan
Gaglani, Shiv. How ‘Googling It’ Can Be Used for Health Care. CardioSourceWorldNews 2016, 5 (11):35.