Can linked Inspire member content provide deeper medical insights into an individual’s chronic disease journey?
Project focus: Longitudinal linking of individual Inspire posts to glean richer medical insights
Pharmaceutical companies follow reports of adverse drug events as part of their post-marketing safety surveillance process. Unfortunately, underreported adverse events and data quality issues can make this process less accurate and informative.
Recent studies have evaluated how individual social media posts can contribute to safety surveillance. But social media limitations, such as limited character counts on Twitter and spur-of-the-moment posts on Facebook, may create bias. In addition, information shared in prior posts (such as the name of a drug) may not be repeated and get lost in subsequent posts within the same discussion thread.
The focus of this study was to evaluate whether linking Inspire member-generated content from the same individual in a longitudinal fashion would allow for a more holistic assessment of that person’s chronic disease journey.
Pharmaceutical and healthcare clients can leverage this type of analysis of real-world social media conversations to identify drug benefits and side effects that might get overlooked by traditional post-marketing safety surveillances. This information provides an opportunity for drug companies to quickly address potential issues. Physicians can also use this information to educate patients about a drug.
Inspire solution: Identify groups of Inspire member-generated content with the highest complexity scores to glean medical insights
Researchers obtained publicly available, deidentified Inspire member posts made from January 1, 2015 to November 1, 2015. The posts came from members of the Scleroderma Foundation Support Community and Arthritis Foundation Support Community on Inspire. Researchers assigned a unique identifier to each author.
Expert reviewers manually curated a random sample of 2,817 threads, containing 21,313 individual posts from 3,601 unique authors. To help systematically identify which collection of posts offered the best medical insights, researchers used an automated quality scoring algorithm to generate a complexity score. The score comprised 28 indicators for each author’s collection of longitudinal posts. Researchers evaluated the 16 groups of posts with the highest complexity scores for relevant medical insights and key attributes.
Benefit to the client: Richer medical insights through linked social media posts
Relevant medical findings from linked social media posts
medical history
disease burden
use of nonmedical treatments
lab results
treatment history
adverse events
adverse events
Nontraditional insights from linked Inspire member posts
Researchers also identified common nontraditional insights from individual authors on topics that discussed:
- environmental factors, such as exposure to pollutants and impact of the weather
- preventing adverse events
- impact of adverse events on attitude and quality of life
Study limitations
Researchers noted that the process of linking posts may introduce additional sources of bias. For instance, people with negative experiences may post more frequently than those with positive experiences.
There is also a need for further research to better characterize linked posts, including those from other social media channels. This information may help researchers better understand the strengths and limitations of this approach and how to optimally use it. Researchers also noted the importance of continuing to employ safeguards to protect the privacy of social media users.