Best practices for the collection and analysis of patient experience data from social media for patient-focused drug development

Patient experience data derived from social media captures the unsolicited conversations of patients and helps in understanding their subjective experiences with disease and treatments. By comparison, many other real-world datasets, such as electronic health records, have the drawback that they solely capture the perspective of health care practitioners. Regulators such as the FDA or EMA have recognized the potential of social media as a source of patient experience data that can inform patient-focused drug development. While social media has limitations, such as the reliance on patient or caregiver self-reporting, it allows us to understand the subjective perception and context of patients, how they experience their condition, its progression, existing treatments and how they manage these, which unmet needs they have, and how the disease affects their daily lives and activities. All this is crucial information that can inform drug development initiatives, and help substantiate relevant outcomes measured, both in clinical trials as well as in post-marketing evidence generation activities. This paper proposes best practices for Social Media Listening (SML) for the purpose of Real World Evidence generation along the following dimensions: purposes and objectives of a SML study, data collection, and data analysis. To illustrate how these best practices can be adopted, we showcase their application in a case study, aiming to unveil the key symptoms and comorbidities that diabetes type II patients face and how these affect their quality of life across an observation period of 24 months. We believe the proposed best practices will contribute to provide a rigorous methodological ground for the use of social media in generating patient experience data that can inform patient-focused drug development and could be accepted in regulatory processes.