How can Sentiment Analysis assist in pharmacovigilance in controlling infectious diseases

Rodrigo Sampera, Senior Consultant – Insights Analyst at ATCS Inc writes about the role of pharmacovigilance in controlling infectious diseases (like COVID-19) and enhancing the pharma industry knowledge of patients’ reactions to therapies through sentiment analysis

In the era of customer-centric marketing, social media is a key tool that allows global companies to listen, monitor and engage with people and organisations’ commentary in real time. In the Pharma industry, this becomes even more important as the ecosystem is complex. In extreme cases, negative side effects of drug therapies or clinical trials can have a direct impact on people’s QOL and life expectancy. Pharma’s ecosystem involves a wide group of stakeholders such as patients, caregivers, healthcare providers, government officials, insurance providers, media organisations and beyond.

Sentiment can often be the initial metric leveraged to gain insights around consumer perception to address complex topics such as consumer’s need, product accessibility, product efficacy, side effects and misinformation. Sentiment analysis provides marketing and multi-functional teams within the organisation with directional indicators on testing hypotheses, validating consumer journeys throughout the marketing funnel, and identifying areas of opportunity to engage the customer throughout. At the fingertips of a marketer, communicator or scientist, sentiment analysis also becomes an important tool when reacting to time sensitive issues.

Sentiment analysis in social media allows to support organisations by proactively identifying events that could be either highly beneficial or highly damaging for a specific brand, therapeutic franchise or company’s reputation. Having a 24/7 social media vigilance mechanism allows organisations and PR departments to always be in the know about incidents or news that are gaining virality.

As large pharmaceuticals understand how to leverage social media data for directional listening, sentiment falls at the forefront of the process. Sentiment analysis plays the role of a primary triage point in PV. Leveraging AI models that identify sentiment allow PV monitoring for AE/PQC issues, which can be a proactive indicator of potential crisis arising. When conducting a broad sweep of social data, a pattern can be discerned through sentiment analysis and AI modelling to identify trends in negative conversation, with a deep dive into its specifics. In many scenarios, sentiment can become a diagnostic tool to indicate the tipping of a scale towards a specific conversation trend.

For example, in case of emerging diseases such as COVID-19, if one was to conduct a broad social analysis for mentions of respiratory illnesses and identified that there is a rising negative trend around a specific topic such as cough, one can take actions to listen in further, quantify the issue and understand details around the topic. In such a case, sentiment analysis and volume trends can become primary indicators of a potential crisis. From this point onward, AI models can be run to understand correlations between cough mentions and other symptoms (like fever), in order to identify emerging correlations and patterns of comorbidity. Social data combined with traditional data, can pose a strong combination of data sources towards eliciting a response during an emerging disease outbreak.

Sentiment tracking can also be helpful for brands, such trend can be adverse event monitoring. The diagnostic methodology remains the same in which brands conduct broad listening across consumer conversation to understand primary drivers – negative, neutral, positive, and then conduct deeper analysis to detect whether conversation is related to AE/PQC topics. Once a thorough analysis of conversations across media has been conducted, AEs may be logged, and specific preventative actions may be taken by the brand to curb AEs.

COVID-19entiment analysispharmacovigilanceRodrigo Sampera
Comments (0)
Add Comment