Dr Helen Keen  

Rheumatoid arthritis is a common, chronic incurable inflammatory disease. Good therapies have existed for several decades in the form of synthetic disease modifying anti rheumatic drugs (sDMARDS). More recently biologic drugs (bDMARDS) have been developed, but they are associated with risk and enormous cost ($23,000 per patient per year). While bDMARDs have been a great advance for RA patients who have not responded well to sDMARDs, or who experience severe side-effects, bDMARDs are not necessarily better than sDMARDs for most patients. Worldwide trends have seen an increase in the use of bDMARDs, and a reduction in the use of sDMARDS, resulting in spiraling healthcare costs. Patient perceptions almost certainly play a role in this phenomenon.

This project aimed to use software to explore online social media platforms to see if online discussions demonstrate recurring themes relating to drug therapy of RA that may have positive or negative connotations (sentiment analysis). This will provide insight into community perceptions of drug therapy in RA.

We asked a third party company to extract data from social media websites (such as twitter, reddit, facebook) which contained comments regarding the commonly used DMARDS. The company applied algorithms to determine whether the sentiment for each piece of information posted on social media websites were positive, or negative, and the general topic discussed (side effects, cost etc).

We found that in the setting of social media, comments regarding the bDMARDS were generally more positive than comments regarding sDMARDS; the main reason people commented negatively about a drug was because it wasn’t working well.

This finding is important, as online discussions indicate that the RA community are in general more negative about the conventional, cheaper drugs, and generally indicate they don’t work as well as newer drugs. This almost certainly has the potential to impact how effective patients perceive their drugs to be. Clinicians should be aware of this when working with patients to choose drugs to achieve optimal benefit from DMARDs.

This research hasn’t yet benefited people with musculoskeletal disease, but we are planning to publish this work, and undertake further work in the area. In particular, we are looking in more detail at the themes that people with rheumatoid are posting on social media. This is unprompted information from patients (rather than traditional qualitative research which can be biased by predetermined themes) and may uncover issues of importance to people with rheumatoid arthritis. We also wish to explore the language used by people with rheumatoid arthritis in the social media setting.