Nicole Baker who is the co-founder of biologit, an early-stage technology startup using artificial intelligence solutions for pharmacovigilance and clinical safety, took us through an interesting session on why AI is needed for monitoring safety of medicinal products at Rising 2020.
Baker who started as an immunologist soon realised that there is a lot of data to read through and that’s when she explored the use of artificial intelligence to bring about ease and efficiency in her pharmacovigilance work.
For the uninitiated, pharmacovigilance is the science and activities relating to detection, assessment, understanding and prevention of adverse effects or any other medicine-related problems.
Before a medicine is authorised for use, evidence of its safety and efficacy is limited to the results from clinical trials, but after the medicine goes into public use, there can still be cases of adverse drug reactions which can be reported by doctors, nurses, and even users themselves. Pharmacovigilance involves ensuring that the patient is safe and that the medicine is not causing adverse reactions.
This is where large volumes of data and information are collected. It goes into a central repository which can be then analysed to observe trends, complaints and other information. For instance, VigiBase is a WHO global dataset that has all the data on adverse drug reactions from across all the major countries.
How AI Comes Into The Picture
There are a lot of resources such as clinical data, medication, claims, molecular profiling, family history, mobile health, environmental, social media and more for pharmacovigilance experts to act on. Baker’s company biologit focuses on literature that has already been published to do its research.
“Drugs need to be monitored once a week. And most of the data conventionally is searched using search strings which can be a slow, laborious and costly process. It might take a lot of time to process thousands of pages and may sometimes result in missing information. AI helps us in increasing our efficiency and we have witnessed y-o-y growth by using AI in pharmacovigilance”, says Baker, who is also the managing director of QPPV, a pharmacovigilance and clinical safety consultancy service company.
She further highlighted the fact that literature is growing tremendously especially with Covid-19 and calls for automation in checking these papers for facts. The need to constantly monitor medicines, from development to post-market further makes it laborious.
With AI, biologit is trying to automate the detection of potential Adverse Events (AEs). It has resulted in a reduction of abstract screening cost by 80%, abstract annotation cost by 50%, removing the setup cost entirely, and more. Currently, in the alpha phase of development, they are in the process of operationalising and refining AI.
“We envision to keep patients safe by simplifying the detection and analysis of adverse effects from drug development to post-market,” said Baker while signing off.