Making Sense of Novel Real World Data to Improve the Development of Patient Relevant Endpoints: A New Framework

As part of achieving drug approvals and patient access, momentum is shifting from traditional methods of collecting and analyzing clinical trial and patient outcomes data to a hybrid approach that involves assessing novel real world data (RWD) for post-market monitoring. The hybrid model combines the results data collected during structured trials with that of out-of-clinic patient and caregiver data from the real world. Real world data helps drug companies hear from a wider breadth of patients and gauge the concerns patients, caregivers, and providers are voicing online about their disease state or medication. As patients usage of social media increases, gathering this information in a digestible and approachable manner will help organizations assess disease burden. This shift is already leading to a change in operating models for pharmaceutical organizations and Clinical Research Organizations (CRO).  

In June 2018, the United States Food and Drug Administration (FDA) encouraged the use of social media to amplify the patient’s perspective emphasizing  the value of considering  the patient experience.  The FDA produced its guidance document in 2020 - Patient-Focused Drug Development: Collecting Comprehensive and Representative Input.

The FDA also has its own programs for and perspectives on collecting information about adverse events from social media, signaling the increased reliance on utilizing specialized social media in Pharmacovigilance. The FDA continues to explore the value of social media mining for safety signal detection and patient experiences. The FDA outlined its potential direction in FDA Perspectives on Social Media for Postmarket Safety Monitoring.

The FDA in examining collection methods for increasing focus on patient experience outlined the following:

• Interviews 
• Focus groups 
• Facilitated discussions at meetings 
• Observational methods 
• Documents (e.g., medical charts) 
• Survey instruments 
• Audiovisual materials 
• Digital health technologies 
• Social Media and Verified Patient Communities 

The advances in Artificial Intelligence (AI) and Natural Language Processing (NLP) have made it possible to curate and interpret large amounts of unstructured data. What many organizations, including the FDA struggle with, is the ability to collect and make meaningful use of such large amounts of unstructured data, organize and categorize it in a standardized way that allows analysis and decision making.

The FDA states in  FDA Perspectives on Social Media for Postmarket Safety Monitoring that the regulatory body has limitations:

• “Many posts do not describe symptoms using the medical terminology needed to explicitly confirm certain conditions or diagnosis. “
• Operational social media surveillance is burdened by curation of large volumes of data.”

Real Life Sciences has experience across therapeutic areas (Alzheimer's Disease, Oncology, Rheumatoid Arthritis, Psoriatic Arthritis, Multiple Sclerosis, Depression, Anxiety and Diabetes)  in novel Real World Data collection and analysis using publicly available social media and verified patient communities. Using the Real Life Sciences patient analytics platform, a combination of natural language processing and data mining approaches were put in place to aggregate, consolidate and structure real world patient reported data from social media and drug safety databases. Pulling together all of this unstructured social media data and structuring it in order to produce outputs has uncovered three overarching parameters:

1. Natural Language Processing is a fundamental requirement to curate, categorize and structure large amounts of data collected from specialized forums and websites related to various disease areas
2. In order to operationalize the data there must be a mechanism to apply standardized terminology to large amounts of unstructured data.  The standardized narratives must then be applied to a framework that can organize patient experiences for analysis
3. Scalable data curation and quality control processes that ensure resulting outputs minimize the noise of RWD

Below is an excerpt from a recent Real Life Sciences’ study “Alzheimer’s Disease: Developing Quantifiable Patient and Caregiver Insights from Self-Reported Specialized Social Media Data” which illustrates how collecting, standardizing and categorizing unstructured patient and caregiver narratives may effectively work:

“Reports were classified against a series of standard medical taxonomies such as the WHO’s International Classification of Functioning, Disability and Health (WHO-ICF) and Medical Dictionary for Regulatory Activities Terminology (MedDRA), and further into the following categorizations:

Social, Physical, Emotional, Cognitive, and Role Activity (SPEC-R)”

Real Life Sciences’ Real World Data collection and analysis solution in combination with its proprietary SPEC-R framework is referenced in “Developing an integrated strategy for evidence generation” in the Journal of Comparative Effectiveness Research.

The power of utilizing social media and verified patient communities lies in the ability to connect the data and make it meaningful. These previously disconnected and unstructured free text sources of valuable information can now identify key insights and disease patterns to conduct analysis that uncovers strategic and actionable findings about patient burden patterns. Further analytics, such as those provided by Real Life Sciences, can lead to decisions resulting in new research, new labels, missed revenue opportunities, stronger instrumentation tools, patient centric research models and more.

Market Access teams, R&D leadership, Medical Affairs, Regulatory Affairs and other teams within the pharmaceutical organization and CRO’s can leverage Natural Language Processing to be proactive and inclusive around patient and caregiver perspectives during the drug development life cycle. Coupled with an organizational framework such as SPEC-R, these organizations can harness the true utility of RWD on social media and verified patient community data to identify patient disease burden including Quality of Life (QOL) measures.