How satisfied are patients with their hospital? Twitter can tell

October 18 01:00 2015 by ushadmin

Boston Children’s Hospital and the Dana-Farber Cancer Institute have made the top 10 on a new scoreboard for US medical centers with the best patient care – hardly a first for these two world-renowned institutions.

The source of this information is unusual, though. The rankings are derived not from official surveys or outcomes but from the sentiment of the masses, based on a real-time watch of what patients and their families are saying about the institutions on Twitter.

“It’s a rough metric of how happy or mad are you at the hospital,” said Jared Hawkins, a researcher at Harvard Medical School and Boston Children’s who was among the group that built the rankings.

On CrowdClinical.com, patients-to-be can search for a hospital and the website will reveal a stream of recent tweets to that health facility, along with a graph that represents the overall mood online.

Like irate diners or passengers stranded at airports, many patients and their families now address their complaints and thanks directly to the social accounts of hospitals on Facebook and Twitter.

“That dialogue is certainly happening,” said Hawkins, who found that about half the hospitals in the US had their own Twitter account through which they communicated with patients. The website is a way to make those conversations more useful as information for the public.

CrowdClinical.com is part of a research project led by Hawkins and a group at Children’s that specializes in spotting patterns of information in public information that thousands of people are sharing on social media. Their goal, in part, is to provide more up-to-date data on the quality of patient care than is captured by official services like the CAHPS Hospital Survey, which collects information from patients. Those results usually become available more than a year after they’re collected.

Hawkins and team set out to investigate if tweets, viewed in aggregate, could be a more timely but still reliable indicator of patient care.

They analyzed 400,000 messages tweeted at national hospitals between October 2012 and through September 2013, selected out the ones that expressed feelings of anger or gratitude, and then constructed an algorithm that would tag them by the sentiment they expressed.

It turned out hospitals that people liked on Twitter were also doing better at not having patients come back within 30 days — one of the indicators of care quality, the group explained in a study published in the journal BMJ Quality and Safety this week.

Hawkins doesn’t see Twitter replacing surveys anytime soon. But it could be a complementary real-time tool, he said.