Who is saying what? Sounds simple doesn’t it? When it comes to health and illness, a huge amount is shared online by a variety of individuals and organizations, through social media and other online content. For pharmaceutical teams looking to inform their communications, content development and more, understanding who is saying what can make a huge difference. Are treating physicians unsure of the implications of the latest research? Are patients having a difficult time getting access to care? Are there topics emerging related to care givers that are not addressed sufficiently in the guides?
And then there is noise. For every valuable post there may be tens, hundreds or more that are of little value, making analysis time consuming. At IQVIA, we help our customers cut through the noise of social media data to surface usable posts that are classified based on the richness of the content and the perspective or type of author that created them. We also classify posts based on sentiment, and not just a generic scope, but sentiment in the context of relevant topics.
For example, we recently trained a machine learning model to know the difference between a patient writing about their own experiences and a physician discussing their professional points of view. Understanding the key needs, concerns, and experiences of patients and other stakeholders is crucial to developing effective communications and ultimately improving patient outcomes.
Working with one pharma customer, an enhanced data feed was created using a combination of social listening tools, machine translation, and AI such as natural language processing and machine learning. The process we developed allowed us to extract insights from millions of social media posts across multiple languages and geographies covering several therapy areas. The resulting enhanced data feed was ingested into a data warehouse and fed into interactive visuals that surfaced trends and highlighted key topics. The algorithms used by IQVIA were able to accurately tag and categorize posts, removing irrelevant and noisy content. This allowed the brand teams to quickly identify core topics and adjacent topics for targeted marketing and communication activities. The models used by IQVIA were flexible, allowing new techniques to be deployed in a modular fashion, balancing the need for results today without losing out on the potential of tomorrow's technology.
The work completed by IQVIA had a significant impact on the overall program, providing brand teams with valuable insights to move from insight to action more efficiently and effectively. Importantly, IQVIA's deliverables helped democratize the social listening process across various stakeholders across the client's organization.
Patient perspective modeling was an essential aspect of the work, making the data easier to filter, understand and work with. Context and richness are key. Real world sources such as social listening platforms contain the valuable information needed, but it is the process of reducing this data to focused, relevant information from which conclusions can be drawn and actions taken that is essential to unlocking that value.
For more information about our work, contact us at
nlp@iqvia.com.
