Cedars-Sinai and UCLA Investigators Collect and Analyze Huge Trove of Tweets and Social Media Posts to Deepen Understanding of Major Public Health Problem


Los Angeles – Nov. 18, 2015 ─ Harnessing the power of social media, medical researchers have sifted through more than two billion tweets and online posts to study the harmful side effects of narcotics medication taken for chronic pain.

The research team, led by the Cedars-Sinai Center for Outcomes Research and Education, reviewed a vast collection of patient entries on Twitter and social media forums.

The unfiltered sentiments posted on these sites revealed details not often captured by physicians or traditional clinical research about the gastrointestinal side effects of narcotics medication. In online messages, for example, some patients described experiencing severe constipation that was even more debilitating than their underlying illnesses.

The researchers believe the study is the first of its kind to analyze social media data related to gastrointestinal side effects from narcotics.

“Social media can be used as a huge epidemiological database, a treasure-trove of insights from patients about their illness experiences, their treatments, and their attitudes and beliefs about health and disease,” said Brennan Spiegel, MD, MSHS, director of Cedars-Sinai Health Services Research and director of the Center for Outcomes Research and Education.

The study appears online in the Journal of Opioid Management. It was conducted in collaboration with Corey Arnold, PhD, from the UCLA Medical Imaging and Informatics Group, and funded by a grant from Takeda Pharmaceuticals International.

Investigators collected 2.5 million tweets and 217,000 posts from health-related social networking websites, using keywords such as “pain meds,” “bloating” and “nausea.” They also searched for the names of several pain-controlling narcotics, including hydrocodone, morphine and oxycodone.

To broaden their search, they turned to a social media data service called Treato, which collects and indexes patient and caregiver content from more than 3,000 health-related websites. Treato reviewed 1.8 billion posts in its repository.

The investigators narrowed their search to 3,303 relevant tweets and posts that mentioned gastrointestinal side effects from narcotics medication.

“Social media provides the opportunity to make observations about patient populations outside of a clinical or research context,” Arnold said. “We were able to filter key signals from the noise in these massive datasets, allowing us to more efficiently distill important patient perspectives that can inform clinical care.”

In the online messages, patients chronicled numerous gastrointestinal issues from narcotics, including nausea, vomiting and constipation, the most frequently cited issue.

The online data revealed that many people who use narcotics for pain relief often are not warned by their doctors of the potential side effects. Some patients turn to over-the-counter remedies or other solutions of questionable value without first consulting physicians.

The study had some limitations. Investigators were unable to determine whether any individual contributed data more than once by posting on multiple social media accounts. Also, because the demographics of the writers were unknown, it also was impossible to link the results to any specific group.

Still, the Cedars-Sinai and UCLA researchers said the rigor of sifting through social media allowed them to generalize their findings to a broader population and showed them how healthcare providers can improve patient care.

“These types of insights provide a blueprint for how to do better,” Spiegel said. “By informing doctors and prescribers about these results, we can hopefully improve the communication and shared decision-making between doctor and patient around pain medications.”