Social Media Sheds Light on Patient Experience With Infliximab

Jackie Syrop

Social media posts by patients with Crohn’s disease are a useful data set to understand patient perspectives on treatment with infliximab (Remicade), according to a new study.

Social media posts by patients with Crohn’s disease (CD) are a useful data set to understand patient perspectives on treatment with infliximab (Remicade), according to a new study by Marco Roccetti, PhD, and colleagues, published in JMIR Public Health Surveillance.

The researchers used a sentiment analysis of Facebook and Twitter posts made by patients with CD to understand their perceptions of infliximab treatment because it is thought that these posts are more frank and sincere than direct patient—doctor interactions. The researchers said that their study suggests that data mining represents a good starting point to provide a completely automated approach for analysis of such data. In addition, their results substantially agreed with large-scale retrospective trials in the medical literature, in which the percentage of patients who experience negative reaction to infliximab is between 20% and 40%.

The researchers borrowed techniques from the new fields of infoveillance and infodemiology that combine different technologies and expertise developed in computer science, econometrics, ethnographic research, and psychometrics to investigate the well-defined community of patients affected by CD. The investigators chose CD to study in part because the disease is often diagnosed in young people (ages 15 to 30), who typically spend a lot of time on the Internet. Infliximab is the pharmaceutical treatment most often cited online, in both positive and negative terms, by patients with CD.

The feedback of real patients is more easily found on Facebook than on Twitter. A 3-valued Likert scale was used to assess the sentiment value of a patient toward infliximab. The researchers compared the ethnographic analysis performed by a computer science researcher and senior gastroenterologists, and analyzed the concordance of such assessments using the square-weighted Cohen’s kappa coefficient method.

Over 26,000 tweets and almost 56,000 Facebook posts made between October 2011 and August 2015 contained the keyword “Crohn.” Concentrating on Facebook, the researchers selected terms that appeared at least 50 times. The terms “adalimumab,” “azathioprine,” and “infliximab” triggered the longest and most vibrant discussions among people posting, the researchers found, with “infliximab” the most sentiment-related term.

Using ethnographic approaches, the investigators performed an expert review of the threads of 261 posts that contained the keyword “infliximab.” All posts containing the word “infliximab” or alternative trade names such as “Remicade” were read by 2 different groups of experts to either confirm or deny the positive or negative evaluations assigned to these posts by the software system used. The ethnographic analysis of these posts gave similar results, with an inter-rater agreement between the computer science and gastroenterology experts equaling 87.3% (228 of 261), a substantial agreement according to the square-weighted Cohen’s kappa coefficient method (w2K = 0.6470). “This result indicates that the evaluation of the feeling that was communicated by a post was independent of the scientific background of the reader, although the senior gastroenterologist tended to classify as neutral a slightly larger share of posts, as not deemed relevant from a clinical point of view,” the investigators conclude.

The classification performed by the software system provides a different outcome than those given by the computer science expert and the senior gastroenterologist. “The number of posts classified as neutral increases, as the sentiment analysis algorithm was unable to determine with a precision similar to a human being the underlying meaning of a piece of text,” they conclude. “Nonetheless, the proportion between positive and negative posts remains comparable, showing that the algorithmic tool could be useful to determine the existence of situations where positive and negative remarks concerning infliximab were made.”