Background/Purpose:

Little is understood about the determinants of symptom expression within individuals with fibromyalgia syndrome (FMS). While FMS sufferers often report environmental influences, including weather events, on their symptom severity, a consistent effect of specific weather conditions on FMS symptoms has yet to be demonstrated. Twitter is a popular internet-based social media platform that enables users to express their thoughts, feelings and details of their daily lives using short text-based messages (tweets) in the public domain. We used computerized content analysis of tweets to investigate the subjective experience of FMS in a large, widely distributed population and any association  with coincident environmental factors.

Methods:

We performed an automated search of Twitter between January 2008 and November 2014 using the hashtags #fibromyalgia, #fibro and #spoonie as keywords. Sentiment analysis, a computerized linguistic method that uses natural language processing and text analytics to identify subjective information in written source materials, was performed using a Streamcrab Python library incorporating the Stanford CoreNLP libraries to quantify the affective content of each included tweet. The classification model was trained using two sets of pre-labelled negative and positive tweets, then used to automatically compute negative and positive sentiment scores between 0 and 100 for each tweet. The sum of the two assigned scores for each tweet is 100. A higher negative sentiment score implies a more severe pain experience. Date, time and location data for each individual tweet were used to identify corresponding climate data (temperature, humidity, wind speed, “feels like”, heat index, wind chill, and dew point) via World Weather Online. The association between negative sentiment scores (indicative of greater pain) and environmental variables was measured using Pearson correlation.

Results:

The search returned 140,432 English language tweets for which location data were available. Examples of tweets with their negative and positive sentiment analysis scores are shown in Table 1.  There was a low positive correlation between humidity and negative sentiment scores which was significant at the 0.05 level (r=0.009, p=0.001). There was no significant association between the other environmental variables and negative sentiment scores.

Conclusion: Twitter users who tweet about fibromyalgia are slightly more likely to include content that suggests a higher pain burden as atmospheric humidity increases. Other local weather features, including temperature and wind speed, are not associated with changes in pain sentiment expressed via Twitter. Computerized content analysis is a novel and potentially powerful method for exploring relationships between environmental variables and the subjective experience in rheumatic and other diseases.

Table 1

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/investigation-of-environmental-associations-of-fibromyalgia-pain-using-twitter-content-analysis/