Artificial Intelligence can improve patient safety

Patient safety is the most critical aspect in the drug lifecycle. Early detection of any reaction and its causality assessment is therefore one of the most important activities in the pharmaceutical industry. Digitalization has brought speed, but significant human work is still required. Medical causality necessitates understanding medical and clinical vocabulary, having medical knowledge, and being able to derive relationships and correlations. Each of these requires human cognitive capabilities.

The challenge increases with new diseases and an avalanche of new drug discoveries. Monitoring and processing the huge number of drug reactions across the globe requires a large number of skilled medical professionals. However, even with the assumed availability of experts, the process is slow. Fortunately, with current advancements in Artificial Intelligence (AI), digital clones of medical professionals and pharmacists, AI digital PV case processors, can solve these problems and scale operations.

AI digital clones can mimic some pharmacovigilance functions

The AI digital PV case processor can add value in the pharmacovigilance automation function, reducing the workload for the human processing. The capabilities of a digital PV case processor include:

• Capture Individual Case Study Report (ICSR) inputs from multiple source types – voice calls, emails, chat, websites, social, Alexa, etc.
• Automate case processing based on cognitive decision making and medical knowledge
• Integrate with multiple dictionaries and references like PubMed, WhoDD, UMLS
• Provide enterprise best practices including SOPs and herd knowledge

At the core of the digital PV case processor is an AI and Machine Learning (ML) based inference engine. The engine can understand the pharmacovigilance context of an entity in essay or paragraph format, infer its meaning, and determine its relationship with other entities.

The workflow engine orchestrates information and data flow to ensure streamlined case management and monitoring. The intelligent modules are loosely coupled and designed for microservices architecture and can be hosted on cloud, on-premise, or in hybrid forms. The engine can also integrate with external systems in multiple ways and understand E2B(R3).

Inside the brain of the digital case processor lies a multi-module AI pipeline that works to accurately identify the entities expected in a human case processor.

Figure 1: How the AI digital PV case processor works

Introducing the AI digital PV case processor simplifies the ICSR process with significantly reduced human touch and processing times. What seemed to be a lengthy and complex set of 30-40 tasks can be limited to only two or three tasks. In the transformed ICSR processing, most tasks are performed by the AI digital PV case processor, leaving cognitive review to humans. 

Figure 2: AI digital PV case processor workload reduction

The world needs safer drugs and treatments. While multiple simulations with digital technologies and high-power computing are currently being evaluated, human trials over a long period of time give the best picture. However, with the growing number of diseases and the recent COVID-19 pandemic, long testing and trial periods are not always possible. Pharmaceutical companies are forced to release medicines faster with calculated risks. These risks can be significantly mitigated by ensuring faster detection, better reporting, and streamlined remedial action. For this to happen, AI-enabled PV automation platforms (like a digital PV case processor) should be embraced by pharmaceutical companies worldwide.