Social media for pharma – an expert’s view
Even with risks and regulations, pharma needs to engage with patients and learn more about them
By Marilynn Larkin Posted on 2 December 2014
Patients increasingly are turning to social media to learn more about their diseases and treatments because much of the information they receive from traditional sources is "difficult to understand and interpret." Still, pharmaceutical companies are hesitant to become more involved in social media, which puts them at a disadvantage.
That was the message of Dr. Sherri Matis-Mitchell, Text and Data Analytics Commercial Lead for Life Sciences Solutions at Elsevier. Speaking at the Medicine 2.0 World Congress on Social Media and Mobile Apps in Malaga, Spain, in October, she said studies suggest that nearly nine out of 10 adults have difficulty using the health information available in healthcare facilities, retail outlets and conventional media, and more than half have trouble learning about their medical condition due to poor understanding of written information. However, pharma companies are not fully responding to this need – and their own need to better understand their customers. She explained:
While pharmaceutical companies are using social media to provide product information and promotional materials, they should also be using it to better understand patients' needs and experiences, and to provide additional education, particularly to those with chronic illnesses. Social media conversations also can inform R&D and pharmacovigilance efforts.
In a nutshell, there's more to social media than sharing photos of cats and lost dogs.
Risk versus benefits of using social media
As any corporation knows, social media has its challenges, but these are usually outweighed by the benefits.
"Although pharma perceives some risk in using social media, it's actually riskier not to use it," Dr. Matis-Mitchell said. Specifically, social media offers drug companies opportunities to connect with patients and understand what they are saying about:
- The industry — e.g., costs of medicine
- Their therapy — e.g., tolerance, adverse events, side effects
- Their disease or condition — which is an opportunity for R&D to identify unmet needs
- A particular company — e.g., its service and reputation
- Competitors — so this can be a form of competitive intelligence
Sherri A. Matis-Mitchell, PhD
Dr. Matis-Mitchell is a senior informatics professional with more than 25 years of experience in informatics, most recently at AstraZeneca. She has a wide breadth of expertise in text analytics, knowledge engineering, data science, natural language processing, genomics and preclinical, translational and clinical Informatics.She received her PhD in molecular biology from the University of Pittsburgh School of Medicine and did her post-doctoral work in informatics at Oak Ridge National Laboratory, where she worked on feature and gene prediction as part of the Human Genome Project. She has numerous articles in peer-reviewed publications and is an industry-recognized speaker on the use text mining of social media and developing analytics to support R&D decision making.
At Elsevier, she serves as Text and Data Analytics Commercial Lead for Life Sciences Solutions.
Social media insights are especially underutilized by pharma R&D in areas such as rare diseases, she said. "Advocacy groups are using social media to get the word out about rare genetic diseases, for example, which affect a small number of people, but still represent an opportunity to help."
Social media can assist in recruitment and protocol development for clinical trials, she added — although a comment in Nature Medicine Drug Reviews pointed out that doing so also opens the door to information-sharing during a trial, which can lead to challenges in terms of blinding, safety and privacy. Therefore, companies and organizations conducting clinical trials "may wish to provide social media guidelines to participants to manage risk," she suggested.
Pharmaceutical industry use of social media seems to be increasing, particularly among large and mid-sized companies, who are engaging with consumers primarily on Facebook, Twitter and YouTube, Dr. Matis-Mitchell pointed out. (See breakout box: "Social Media Engagement: Top 10 Pharma Companies.") However, she stressed that "there is plenty of room for growth."
Problems with social media as "big data"
As a form of big data, social media presents many of the problems inherent in that type of "high-volume, high-velocity and high-variety information asset," Dr. Matis-Mitchell explained.
As an information source, it's noisy and there's a lot of garbage. Therefore, the veracity of social media posts is suspect. Some people say things that are untrue to make mischief or meet their personal agendas. It's difficult to verify individuals who are posting because of privacy walls, so there's a lot of anonymity associated with social media.
Dr. Matis-Mitchell proposed a fifth "v" to the big data issues of volume, velocity, variety and veracity — namely, vocabulary. "Because social media users have their own vocabularies, slang, abbreviations and so forth, we need to build or adapt traditional semantic-based methodology to analyze the content. In addition, trying to make sense of data from multiple social media sources to uncover trends and actionable information takes time, tools and resources.
Regulatory considerations for pharma on social media
Regulatory issues are also a concern. "If someone posts about a side effect and names a specific drug, that is a potentially reportable event, even if there is no apparent causal relationship between the side effect and the drug," Dr. Matis-Mitchell said. "Because such reports are difficult to verify, a lot of manpower can potentially be wasted on false signals."
That is likely to change both in the United States and in Europe, as regulatory agencies address the need for clear, formal guidelines on how to use and respond to social media.
For now, the US Food and Drug Administration (FDA) acknowledged in an announcement of its Patient-Focused Drug Development Initiative, that "patients who live with a disease have a direct stake in the outcome of FDA's decisions and are in a unique position to contribute to the understanding of their disease," including via social media, according to a recent post on Medicity News. The FDA's latest guidance to the industry is the result of an ongoing "collaboration between the FDA, industry and patient groups to identify best practices for engaging patients and to conduct pilot programs that solicit patient input via social media," Matis-Mitchell said.
The European Medicines Agency (EMA) – the European counterpart to the FDA – has its own policyon adverse event reporting, which includes requiring "Marketing Authorisation Holders" to regularly screen Internet or digital media for potential reports of suspected adverse events and assess whether such information qualifies for reporting to the EMA. "A lot of the details are being worked out by both regulatory bodies," Dr. Matis-Mitchell observed. "Meanwhile, multinational companies need to follow local regulations, even if it means sharing different materials in different regions."
How the right tools can help
Despite the problems and regulatory challenges, participating in social media is worth the effort, particularly as part of pharmacovigilance, Dr. Matis-Mitchell stressed. "Twitter simplifies the process by encouraging sharing and analysis," she noted. To facilitate analysis, Twitter has released an API (advanced programming interface) that makes tweet streams available to developers who are building analytics tools to analyze the anonymized content.
"Whereas many of the blogs where patients share information are locked up behind passwords, and terms and conditions that don't allow use of that information, Twitter content can be used for the benefit of patients," she explained. "For example, if patients are tweeting and retweeting about a drug making them sick, and if that's really true, being able to detect that signal could be useful."
Social media aggregators such as Crimson Hexagon, which Dr. Matis-Mitchell used when she worked at a pharmaceutical company, can mine not only Twitter but also public forums, blogs, Facebook and other social networks to uncover patient sentiments and opinions, she said.
Social Media Engagement: Top 10 Pharma Companies
Of the 50 largest drugmakers worldwide, only half are involved in social media at all, and just 10 use all three of the largest social media sites — Facebook, Twitter and YouTube — according to a recent report, Engaging patients through social media, by the IMS Institute for Healthcare Informatics. Following are the top 10 ranked pharmaceutical companies, according to the IMS Health Social Media Engagement Index, and each company's score. The Index reflects the current overall usage of social media by pharmaceutical companies on healthcare-related topics.
She described a pilot program in which a pharmaceutical company asked the question, "What is the overall impact of side effects for patients taking a therapeutic cancer drug Z, which has a similar mechanism of action to our own compound X?" By monitoring and analyzing many sources of online comments on the topic, the aggregator was able to provide insights into relevant conversations of lung cancer therapies, track opinions over time and also identify opinion influencers.
From the pharmacovigilance perspective, treating social media input as a "text analytics problem" enables companies to use existing text mining software from Elsevier and others to at least become aware of what people are saying about their products, short of considering any potentially reportable events, Dr. Matis-Mitchell said. "The algorithms may not be sophisticated enough yet to weed out incidents in which false signals are created with malicious intent, but the approach does identify trends."
The text analytics approach also can support decision-making in early-stage drug discovery, Dr. Matis-Mitchell noted. At Elsevier, to help a client decide whether to move forward with a certain compound, text mining technology was used to mine tweet streams, blog postings and other comments for relevant information. This content was treated as though these social media comments were text documents — i.e., input from the literature. This type of analysis requires a "deep understanding of colloquial vocabularies and how they relate to technical and medical vocabularies," she explained.
What will it take to induce pharma to enthusiastically embrace the potential of social media? Beyond the ability to make sense of social media input, more clarification is needed on the regulatory side and on the ethical implications of social media mining, Dr. Matis-Mitchell said. However, patients are driving the use of social media and increasingly demanding that their healthcare providers and medication manufacturers participate, so very soon, companies may have no choice but to "embrace social media in a bigger way by not just pushing information out but also by improving their ability to listen to what patients say."
Patients also are advocating for greater sharing of health information among providers using electronic health records. But the healthcare system needs to go beyond simply sharing, Dr. Matis-Mitchell said. "Rather than just allowing an electronic health record to be sent from one specialist to another, we'd want to run analytics across that record to ensure the key points are clear and summarized for busy healthcare providers."
The ability to highlight and extract actionable data in a patient's electronic record "could help avert errors and streamline care," she said.
Dr. Matis-Mitchell recalled how she was treated incorrectly due to a coding error in her own electronic health record. "I switched my insurance, and when the new company reviewed my record, they coded me as having hyperthyroidism. My mother has hyperthyroidism, and my record likely stated, 'patient's mother is hyperthyroid; patient has no history of hyperthyroidism.' Clearly, the review process identified the word 'hyperthyroid' in my record but didn't read further or look at the context. That's something text analytics can do —give us context, as well as the relevance of the occurrence of a specific term in a sentence. Had that approach been taken, the outcome could have been very different."
"Social media is here to stay and pharma should be usefully — and responsibly — engaging in it to get in touch with patients," Dr. Matis-Mitchell concluded. "When companies engage, and have the right tools and analytics capabilities, they can make better medicines."
Being absent from the conversation could have a long-term (negative) effect on our relationship with patients and providers. Mistakes are inevitable. Even social media 'experts' are not immune to them. But over time, technology will help us gain valuable insight into better treatments that impact and improve the quality of lives.
"Engaging Patients thru social media," IMS Institute white paper (January 2014)
"Industry Usage of Social and Digital Media Communities in Clinical Research," Tufts Center for the Study of Drug Development white paper (June 2014)
"Engage with research participants about social media,"Craig H. Lipset, Nature Medicine (March 2014)
Elsevier Connect Contributor
Marilynn Larkin is an award-winning science writer and editor who develops content for medical, scientific and consumer audiences. She was a contributing editor to The Lancet and its affiliated medical journals for more than 10 years and a regular contributor to the New York Academy of Sciences' publications and Reuters Health's professional newsfeed. She also launched and served as editor for of Caring for the Ages, an official publication of the American Medical Directors Association. Larkin's articles also have appeared in Consumer Reports,Vogue, Woman's Day and many other consumer publications, and she is the author of five consumer health books.