5 ways social media can make drug development more patient-centered

By Leah Sherwood, The Science Advisory Board assistant editor

September 14, 2021 -- Insights gleaned from analyzing social media can facilitate patient-centered drug development and help spur innovation in drug discovery, according to a paper published online recently in Drug Discovery Today.

Incorporating patient perspectives into the drug discovery and development process has become increasingly important for both pharmaceutical companies and regulatory authorities. The U.S. Food and Drug Administration (FDA) and other regulatory bodies have called for more patient-focused drug development efforts and are starting to demand patient-perceived benefits be included in the drug approval process.

Using big data techniques to analyze social media content -- a process called social media mining (SMM) -- may provide a viable complement to traditional drug development approaches, which often fail to incorporate the perceptions of patients in the drug development process, according to the paper, which was published on September 1, 2021.

"With the increase in both social media data and [the] number of patients sharing disease trajectory-related information, SMM-based research endeavors are expected to increase in the future and will become a key enabler of patient-centered drug development," wrote the authors, led by Jonathan Koss of the University of Witten/Herdecke in Germany.

Data obtained from social media can unearth new insights that could guide patient-centered drug development by identifying unmet needs and opportunities for innovation. In addition, online conversations on health-related topics can help identify side effects or inconveniences, among other patient perspectives. For example, social media posts complaining about the large size of a pill may tip off drug developers to patients' discomfort or decreased adherence to a treatment plan.

The 'what' and 'how' of social media mining

The paper's authors define social media mining as "the extraction and analysis of data gathered from online forums, blogs, and social-media platforms to gain knowledge concerning specific communities, as well as their members' perceptions and needs."

Extracting, filtering, and harnessing this potentially valuable data on social media is nearly impossible to do manually due to the sheer volume of data and the fact that it comes in an unstructured format. However, recent advances have enabled automated and cost-effective processing of vast amounts of unstructured text, allowing SMM to be applied to a wide range of disciplines and applications.

In the healthcare arena, researchers have used SMM among patients discussing their personal experiences with a disease on social media. For example, SMM has been used to monitor outbreaks of influenza, to characterize symptoms of the disease, and geographical dynamics.

Similarly, SMM can help pharmaceutical companies focus their drug development efforts on patient needs and gain valuable information at all stages of the development process, according to the paper.

SMM can be performed in the following five stages:

  1. Resource identification. Which social media platform is the best? Facebook and Instagram may have a large amount of data, but both platforms allow advertisements, which reduces the quality of the data. Smaller health-related forums that do not allow advertisements, however, may have higher quality data, albeit less of it.
  2. Data extraction. There are two automated techniques used to extract relevant data from social media sites: focused crawling and web scraping. Typically, third-party tools are used to extract data, although some social media services -- notably Twitter -- offer application programming interfaces that facilitate data extraction.
  3. Data preprocessing. In this stage, natural language processing techniques are used to reduce "noise" and to structure data so that it is more amenable to analysis. Certain stop words, such as "of" and "the," that provide little semantic contribution are removed during this stage.
  4. Analysis. The fourth stage typically relies on machine learning to extract useful insights from the preprocessed data. Unsupervised machine learning algorithms can detect patterns in the data without the need for predefined outcomes. However, if researchers need answers to specific questions with predefined outcomes, supervised machine learning techniques can be used.
  5. Evaluation. The final stage of the SMM pipeline is the point at which domain experts evaluate and verify the results of the analysis. This step often concludes with the production of charts and graphs that summarize the results of the data, which decision-makers can then use.

5 ways that SMM supports patient-centered innovations

The paper's authors envision five use case scenarios in which SMM could play a role in the drug development pipeline:

  1. It can identify and prioritize patients' unmet needs. Leveraging SMM, for example, can help companies identify patients' concerns about certain drug side effects, such as hair loss or sexual dysfunction, that are deeply important to them.
  2. It is a method for characterizing target populations. To improve the odds of success in developing new drug therapies, it can be necessary to locate a target population based on its phenotype, for example. This targeting allows pharmaceutical companies to recruit individuals who are likely to show benefit during clinical trials. Social media represents one method -- digital phenotyping -- that allows companies to identify subpopulations associated with the specific needs, conditions, and demographic characteristics they are seeking.
  3. SMM can help repurpose drugs. Companies can leverage SMM to develop new applications (e.g., smoking cessation) for drugs prescribed for existing indications (e.g., depression).
  4. It can help recruit patients. As mentioned earlier, SMM can be leveraged to recruit patients with specific health conditions and demographic characteristics that make them good candidates for clinical trials.
  5. SMM can help detect early adverse events. Using early signals from SMM, companies and regulatory authorities can identify adverse drug events ahead of established reporting systems (the FDA adverse event reporting system).

The authors noted that social media mining can be more effective in some circumstances than others. For instance, some illnesses cause minimal patient distress and, therefore, are not often mentioned on social media, and some patients are less likely to have access to social media platforms.

"Overall, SMM can contribute toward aligning the drug development processes of pharmaceutical companies with patient needs, and can assist in allowing these development processes to respond to the changing business environment," the authors concluded.

Do you have a unique perspective on your research related to drug development? Contact the editor today to learn more.


Funding for AI drug development skyrockets in 2021
The market for artificial intelligence (AI) in drug development and clinical trials has seen renewed interest in recent months, with total funding in...
AI drug development startups raised $2.1B in 1st half of 2021
The market for artificial intelligence (AI) in drug development and discovery has been red-hot in recent years. The potential impact that AI can offer...
Royalties challenge risk and reward in AI drug development
As artificial intelligence (AI) becomes increasingly important for drug development, how should AI firms -- many of which are small startups -- structure...
How is the pharma-AI partnership landscape changing?
Artificial intelligence (AI) is taking on a growing role in drug discovery and development, with some 250 firms in a market space that's worth over $9 billion....
Repurpose.AI, Scripps collaborate on COVID-19 drug discovery
Repurpose.AI will collaborate with Scripps Research to develop COVID-19 therapeutics with a focus on drug repurposing.

Copyright © 2021 scienceboard.net

Science Advisory Board on LinkedIn
Science Advisory Board on Facebook
Science Advisory Board on Twitter