Social media may highlight research needs, clinical care gaps in lupus
Click Here to Manage Email Alerts
Patient-generated data gathered via social media surveillance may highlight unmet research needs and clinical care gaps for patients with lupus, including underrepresented groups, according to data published in Arthritis Care and Research.
“Researchers have been studying disease communities on Twitter and other social media platforms for a few years,” Katja Reuter, PhD, study co-author and research assistant professor in the department of public health and preventive medicine at SUNY Upstate Medical University, in Syracuse, New York, told Healio. “We wanted to explore the potential to identify lupus patients from diverse groups on Twitter. The fact that there is an active group of patients on Twitter allows us to think about ways to reach them, such as providing information resources or learning from their perspectives.”
To assess the feasibility of using Twitter to identify diverse patients with lupus and examine their perspectives on symptoms and medications, Reuter and colleagues conducted a retrospective user and content analysis of public Twitter posts from U.S.-based accounts between Sept. 1, 2017, and Oct. 31, 2018. Posts discussing lupus symptoms or medications were included in the analysis. After filtering out duplicate content and posts from automated and commercial accounts, the researchers accounted for 8,446 tweets from 179 accounts belonging to patients with lupus.
“In this study, we focused on Twitter and demonstrated that patients with lupus can be identified successfully on this platform,” Swamy Venuturupalli, MD, FACR, founder of Attune Health, attending physician at Cedars-Sinai Medical Center and associate clinical professor of medicine at UCLA, told Healio. “Patients with lupus who are active on Twitter often discuss their symptoms and medications and perspectives.
“This group includes underrepresented groups including women and people of color,” he added. “Additional research involving other social media platforms can provide rich data on patient perspectives and inform clinical care and research in rheumatology.”
According to the researchers, 85.47% of the analyzed accounts belonged to female patients, and those patients created 89.83% of the posts. Among the 179 included accounts, 34.64% belonged to people of color, who also accounted for 33.07% of analyzed posts, the researchers wrote.
Regarding the types of symptoms described on the platform, researchers found that posts from female patients contained significantly more references to an inability to perform daily tasks (P = .0025), skin sensitivity (P = .0167), depression (P = .0613) and arthritis (P = .0057). Meanwhile, posts from male users contained significantly more references to fatigue expression (P < .0001).
Certain symptoms were discussed by a significantly larger margin by white patients, including general pain (P < .0001), flares (P = .0223), arthritis (P < .0001), skin sensitivity (P = .035) and rash (P =.0379). Meanwhile, patients from underrepresented groups more commonly discussed inflammation (P = .0339), self-consciousness (P = .0006), alopecia (P < .0001), visual disturbances (P = .0013), confusion (P = .0149) and difficulty concentrating (P = .0268), the authors wrote.
“Only a few studies have looked at lupus patients on social media,” Reuter said. “Today, the medical community does not take advantage of patient-generated health data from social media. We want to raise awareness and spark more research in this area with this study. The main idea is using this data to determine research needs and potential gaps in clinical care.”
Perspective
Back to Top
John J. Cush, MD
There is a real and potentially substantial value to research on social media “signals.” Sure, there are potential biases like diagnosis veracity, selective reporting and reporting that may be reactive to the discussion rather than voluntary. Nevertheless, social media has been shown to reliably predict the future in everything from drug safety and drug withdrawals to elections. Dr. Milton Packer once said that registry data is “not proof” but can be valuable in “hypothesis generation.” The same can be said of insights from social media.
This real world, social media lupus cohort was more representative of patients with lupus than that seen in clinical trials — being more than 85% female and 35% persons of color (POC). The analysis challenge lies in homogenizing your cohort from a very large heterogenesis population (eg, Twitter). When scrubbed and verified, this analysis provided insights from 179 SLE patients. They found subgroup differences between men and women (fatigue, arthritis and depression) and between whites and POC (arthritis, confusion and alopecia). This is an important beginning to identifying key medical signals from social media. I do believe that social media surveillance of select patient groups, like lupus, can lead to novel hypothesis generation about patient concerns and trends. However, this will require large samples and a data methodology — as in this report — to make assumptions reasonable predictive.
Collapse
Read more about
Click Here to Manage Email Alerts