Twitter and Big Data Could Predict Emergency Room Rush Hours

Researchers at the University of Arizona are saying there may be a potential life-saving value to people tweeting about their health problems.

Research is being published in Journal of Biomedical and Health Informatics. The study analyzed asthma-related emergency room visits in the Dallas-Forth Worth area between October and December 2013. 


Tweeting about Your Health Problems Could Help Emergency Rooms

Scientists analyzed big data from Twitter and the connection between tweets about asthma and asthma-related emergency room visits. They scanned  Twitter streams for 19 keywords and phrases

Over a half billion tweets were analyzed. 1.3 million of the tweets contained asthma-related keywords with “asthma attack” OR “inhaler” being the top keywords used in tweets. 

They then compared the tweets to air quality data from Environmental Protection Agency and emergency room visit data from Children’s Medical Center of Dallas.

There seemed to be a correlation. Researchers noticed people were tweeting a lot about their asthma symptoms.

The data combines:

  • Twitter posts
  • Air quality in particular area
  • Hospital data
    • Data combination creates model that may predict emergency room trends
  • This new model is first to look at chronic illnesses such as asthma
  • Researchers say this prediction model can anticipate emergency room visits with 70% precision

Existing surveillance models:

Published from U.S. Centers for Disease Control and Prevention, surveillance models are usually not able to determine in that far advance how many visits emergency rooms will get. 

Google Flu Tracker integrated some social media and search engine data to monitor the spread of contagion, but has been discredited by some critics. They say it may overestimate actual number of flu cases because not everyone who searches for info about it actually has the flu.

Critics of this model question:  How often do people tweet about chronic illnesses?