Thoughts on Sentiment Analysis in Health Care

March 6, 2012

There is increasing interest in Sentiment Analysis of social media. The brief messaging systems like Twitter and Facebook allow for picking out words and phrases using algorithms to find positive or negative sentiment. There are an increasing number of tools to do sentiment analysis and market research firms willing to do it. Using a simple tool like TwitterSentiment can get quick results but are they meaningful for healthcare? For instance, looking at “Mayo Clinic”, the results are rated as negative but examining the tweets show a mix of statements and some rated negative (78%) have to do with waiting for someone in surgery or concern about illness rather than care or quality. But using their twitter handle @MayoClinic shows 90% positive. Similar results are seen with Cleveland Clinic and Hopkins Medicine. Perhaps tools with a license fee are more sophisticated and show clearer results. An market research firms may take more time to tweet algorithms to sort out positives and negatives.

But hospital reputation is not the only use of sentiment analysis in health care social media. What are sentiments about diseases and conditions? In a brief check, cancer was more positive than cancer. But some comments are mixed, “thanks love! My roommate got diagnosed with cancer” is rated positive. My conclusion, sentiment analysis in healthcare requires fine tuning – what is good news, what is bad, what is neutral? The temporary pain of treatment can be a real downer but the results a real high. I will be interested to see more written on this in the future. Some research has begun, such as, this article on cancer survivorship.

Share this:
  • Facebook
  • Twitter
  • Delicious
  • Digg
  • Tumblr
  • StumbleUpon
  • Pinterest
  • Add to favorites
  • Email
  • RSS

Leave a Reply