Over the past two years Twitter data has been used to conduct real time research into public health issues across the world. This ground-breaking form of ‘digital epidemiology’ has proven successful in both predicting and monitoring outbreaks such as flu and measles, as well as offering an alert system for crisis situations.
Twitter data could also be used to track daily public health at community or city-level and inform the pharmaceutical industry of a wide range of public health issues in real-time.
Blurrt have tracked Twitter discussion around a range of illnesses including eczema, the common cold, the flu virus and migraines for three weeks in November. Through their advanced analytics platform Blurrt have extracted some key finding which are presented below.
Key Findings
-
This analysis suggests females tend to tweet more about illnesses and ailments than males, with females contributing to 60% of the discussion on average.
-
Those suffering from illnesses are more likely to tweet in the early stages of getting an illness than the latter stages. This suggests Twitter data could be used to track and monitor early outbreaks of illnesses before they become an epidemic.
-
Eczema generated the most emotive reaction on Twitter with over 40% of tweets expressing anger. This demonstrates how frustrating this particular condition is for sufferers compared with other illnesses.
-
Sufferers often coin their own terms to describe illnesses and symptoms. Twitter data around illnesses could be used by pharmaceutical companies to inform product descriptions or create more relevant and engaging advertising campaigns.
Analysis period – 5 November to 24 November 2015
Flu
There were 186,910 tweets posted discussing the flu virus during the analysis period of which 126,473 were tweets and 60,437 were retweets.
The biggest spike in discussion around ‘flu’ on Twitter happened on November 12, when 15,268 tweets were posted worldwide. Females tweeted more about flu than males with 55% of tweets coming from females and 45% coming from males.
The most commonly mentioned topic within the data was the preventative flu injection, which was described as ‘flu shot’, ‘flu vaccine’ and ‘flu jab’ depending on the location of those tweeting. ‘Flu shot’ was the preferred term in the US whereas ‘flu jab’ was more commonly used in the UK.
People were more likely to discuss the early stages of flu on Twitter than the letter stages, with terms such as ‘getting the flu’ and ‘coming down with the flu’. These early warning signals mean Twitter could provide the first signs of an outbreak or epidemic before it takes hold within a community.
Common Cold
There were 60,973 tweets posted during the analysis period discussing the common cold, of which 41,064 were tweets and 19,909 were retweets. Again more females contributed to the discussion than males with 57% of tweets coming from females and 43% from males.
The top trending topic in discussion around the common cold was ‘the sniffles’, which was used to replace the term cold and to describe the main system. The words ‘throat’ and ‘face’ were also mentioned frequently as the areas most affected by the virus.
Eczema
There were 17,786 tweets posted discussing eczema during the analysis period, of which 11,359 were tweets and 6,427 were retweets. As with other illnesses discussion was fairly even with a spike on the 13 November when an article about dietary impacts on eczema was retweeted over 550 times.
Again more females tweeted about eczema than males with 62% coming from females and just 38% from males.
Unlike the other illnesses monitored throughout November, tweets discussing eczema were particularly emotive with 64% of sentiment bearing tweets expressing a negative sentiment. In order to understand this reaction further we can look at the emotions expressed.
Anger was by far the most common emotion present with just over 40% of tweets expressing anger in some way. Here are some common examples of tweet expressing anger and frustration:
‘Hate’ was the word most commonly used to express anger as well as ‘annoying’, ‘irritated’ and ‘mad’.
Migraines
There were 90,520 tweets posted mentioning migraines during the analysis period of which 66,258 were tweets and 24,262 were retweets.
More females tweeted about migraines than males during the analysis period, with 64% of tweets coming from females and 36% coming from males.
The terms ‘such pain’, ‘ache’, ‘excruciating’ and ‘ear piercing’ were the top four terms most commonly used to describe the pain associated with migraines. The term ‘distinct headache’ was also a top trending term within the data, which was often used to describe the symptom of a migraine.
To sum up
These snapshots into discussion around illnesses on Twitter tell us a great deal about the type of information being shared, how frequently they’re mentioned and how people are describing the symptoms they are experiencing. This information has a multitude of uses. e.g. creating more accurate and engaging product descriptions/ advertising campaigns, understanding when demand for pharmaceutical products will be highest, and reacting to outbreaks before they happen.