Decision support and speed improvements for adverse event handling in the pharmaceutical industry
Can we speed up the adverse event handling process and make it more consistent using AI?
Our client is a large player in the pharmaceutical industry, with a substantial department focusing on animal health products. These products are sold all over the world, with every region having its own strict regulations when it comes to handling adverse event registrations. Also, our client wants to offer its customers a quick, seamless and consistent experience when adverse events emerge, currently employing dozens of people around to globe to make this happen. To cope with an increased number of incoming client contacts (due to more product sales), we jointly explored the application of AI-based decision support.
With our client’s domain experts, we mapped the adverse event handling process and found various opportunities for automatization and decision support. Next, we gathered relevant data, both historical cases and medical protocols, to train and validate our models. The majority of the data used consisted of text.
“To make our models usable for the client’s adverse event handlers, we created a user interface around the models that runs in parallel with the current administrative systems.”
As soon as an adverse event gets reported, the case handler can paste the case description into the new software for analysis. The symptoms that a customer describes are automatically extracted from the text and encoded using a standard medical encoding. Then a number of automated risk assessments is performed, e.g.: how serious is this case? How likely are these symptoms to occur with this product within a given timespan?
Next, answer templates are automatically generated for the case handler to use in her/his reaction to the client and relevant authorities. The basis for all these predictive models is in a combination of historic cases and medical documentation of the products and clinical protocols.
Our client strives to handle adverse events as quickly and professionally as possible, worldwide. By aiding the case handlers in every step of the handling process with an AI-based decision support tool, the process can be shortened significantly. Not only that, but we can even make sure the most severe adverse events automatically get routed to the right expert.
The two main benefits of implementing this solution are faster and better handling of adverse events for customers, and better compliance with local regulations worldwide, now and in the future.