Monitoring the safety of medications after they have been released to the market, is amongst the greatest responsibilities that the Pharma industry has. Getting this area wrong can not only cause catastrophic reputational damage to both specific businesses and the sector in general, it directly impacts the lives of patients, contradicting the absolute first principle of medicine — “do no harm”. A 2015 report by the British Journal of Clinical Pharmacology found that 6.5% of UK hospital admissions were the result of adverse drug reactions (ADRs), at huge cost to the industry, and it’s long been well established that the vast majority of ADRs go unreported (somewhere around 92% according to this study). Traditionally, pharmacovigilance has relied on healthcare providers and patients to report ADRs through official channels. Social media, however, has opened up a new frontier of real-world data being shared online by patients every day. What’s more, it’s doing so in ways that are substantially more affordable than traditional pharmacovigilance research, which can be hugely resource-intensive.
What is Social Media Pharmacovigilance?
Social media pharmacovigilance involves mining social media platforms for mentions of adverse drug reactions (ADRs). Social data can provide a more comprehensive view compared to traditional reporting methods, coming directly from patients without being interpreted by healthcare professionals or facing the usual hurdles associated with reporting adverse events through official medical channels — time constraints during doctor visits or lack of awareness about reporting systems.
The potential benefits are significant. Effective social listening can detect unknown adverse reactions, identify increased frequency of known reactions, and better understand the real-world benefit-risk profile of their medications.
The Social Media Pharmacovigilance Process
The process of social media pharmacovigilance typically involves several key steps:
- ➔1) Extracting relevant data from social media platforms
- ➔2) Analysing the language used to identify potential ADRs
- ➔3) Removing any irrelevant personal information
- ➔4) Classifying the data into relevant categories using techniques like Named Entity Recognition and Topic Modeling
Named Entity Recognition sorts the raw data and classifies it into predefined categories like medication names, prescribed amounts, and known adverse effects. This is useful for tracking increases in the frequency of reported reactions. Topic Modeling is a more exploratory approach that uses automated keyword searches to surface potential ADRs when the specific reactions to look for are unknown.
Challenges and Ethical Considerations
Social media isn’t quite the miracle cure it can seem for pharmacovigilance. While it holds great promise, it also comes with challenges and ethical considerations that the industry must grapple with. A key challenge is developing best practices for safely collecting large amounts of social media data while protecting the privacy of users who may be unaware their information is being used for this purpose.
As the authors of the BJCP study point out, 'Just because someone has not kept their social media posts 'private' does not necessarily mean that they are happy for that information to be used in ways that they have not consented to.' They argue that using social media data for pharmacovigilance without informed consent could undermine public trust.
There are also technical challenges around accurately interpreting the informal language used on social media to describe ADRs. Automated listening tools need to be able to handle the variations and evolutions in how people talk about medications and side effects online.
The Future of Social Media Pharmacovigilance
Despite the challenges, the pharmaceutical industry increasingly recognises the importance of social media pharmacovigilance. The WEBRADR project, a public-private partnership funded by the Innovative Medicines Initiative, for example, is working to develop a mobile app to enable seamless adverse event reporting by patients directly through social media platforms.
AI is also playing a huge role, and that’s a field that is developing at speeds that are often frightening. It means sifting through the vast amounts of social media data to accurately detect signals of potential ADRs is getting easier and faster. However, human expertise will still be essential to provide context and turn the raw data into actionable insights. As we’ve found often at Buzz Radar, AI is always a tool to make work more efficient, and rarely a replacement for human specialism.
How Buzz Radar is Using AI for Social Media Pharmacovigilance
At Buzz Radar, we're making pharmacovigilance easier with our AI life science marketing assistant, ‘BRIANN’. It’s designed to automatically monitor all mentions and side effects across social media in multiple languages. By continuously scanning social media platforms, BRIANN can quickly identify potential adverse drug reactions and send alerts in line with guidelines.
BRIANN can process vast amounts of data at game-changing speeds. Manually sifting through social media posts to identify potential ADRs is incredibly time-consuming (and thus expensive) but BRIANN can analyse millions of posts in real-time. This allows pharmaceutical companies to identify potential safety issues much more quickly and take action to mitigate risks, and do so without employing large teams to search through endless data, or redirecting resources.
BRIANN's advanced natural language processing capabilities allow it to understand the context and sentiment of social media posts, even when they use informal language or slang terms to describe side effects. This enables us to capture a wider range of potential ADRs that might be missed by traditional keyword-based approaches. Plus, BRIANN can operate in multiple languages, ensuring that potential safety signals are not missed due to language barriers.
Social media pharmacovigilance is still an emerging field, but it’s one that’s developing fast … and one that is likely to become one of the most important tools Pharma has to monitor ADRs, creating a more proactive, vigilant and substantially cost-effective approach to pharmacovigilance.
It’s an area of our work with pharma and life sciences that we’re extremely proud of and excited about. Drop us a line to find out how our expertise can help make vital pharmacovigilance easy, efficient and affordable.
Published on 2024-06-04 12:18:02