Innovation in Pharmacovigilance: Use of Machine Learning
Pharmacovigilance (PV) is the pharmacological study relating to the collection, detection, assessment, monitoring, and prevention of adverse drug reactions (ADRs), i.e., side effects of medicinal products. Adverse drug reactions are one of the major causes of morbidity and mortality in the world. Therefore, for testing and marketing a drug in most of the countries, ADR data need to be collected by the license holder (usually a pharmaceutical company), through PV studies and must be submitted to the local drug regulatory authority.
In PV industry, most of the operations, e.g., detecting and reporting ADRs, assessment of its seriousness and causal relationship with the suspected drug, coding of adverse events in technical terms, preparing safety reports of individual cases, depending upon the human interventions which consume a lot of time. So, it is highly required to develop the new methods to facilitate the automatic detection of ADRs, especially those which are arising due to drug-drug interactions.
For such situation, an effort by regulatory authorities is the discussion related to the machine learning (ML) technology which could support PV because this technology requires no translator to translate the codes of ADRs used under PV system and it can learn by its own without human intervention. Machine learning has an extremely important role in PV, some of which are highlighted below:
- It can be used for the identification of ADRs that were previously considered unexpected from the drug.
- It can provide guidance in the product’s labeling as how to minimize the risk of using the drug in a given patient population.
- In drug development, it can enhance the stages of the drug discovery process (chemical structure of the drug, identification of the effect of the drug in both preclinical and clinical trials, etc.) by analyzing and interpreting the biomedical data from research experiments to predict the ADRs of the drug.
The AI-led PV system can automatically extract and code ADR data from multiple sources to reduce case processing effort. By adopting such a knowledge-driven idea, there is an evident rise in the journals, media, and articles that some pharmaceutical companies tie-up with IT companies and launch AI-based solutions for drug safety.
- Genpact: A professional services company has developed Pharmacovigilance Artificial Intelligence (PVAI) tool which can extract and code ADRs from source documents. This fully-integrated tool can eliminate manual workflow and can save significant time and resources. Above all, it can continuously build predictive insights as more ADRs goes through it over time.
- Bayer: A multinational pharmaceutical company, Bayer, in collaboration with Genpact, has developed advanced AI and ML technologies to increase the efficiency of PV operating model and case processing while maintaining high quality and compliance standards.
Besides helping to extract and code ADRs, ML can provide inputs to progress the processing of Individual Case Study Reports (ICSRs), which involves entering details of the cases related to ADRs, drugs, patient history, etc. As it is a content-rich report, adding information in it is one of the biggest challenges. The benefits of using AI tools in ICSRs are:
- It can reduce the time of processing cases through automation.
- It can improve the quality and accuracy of reports.
- Using this application, AI can easily place information related to ADRs in case reports.
As AI and ML are in developing stage under PV, awareness about it is a bit less, due to which issues are solved at a slower pace. This can be managed by a strong collaboration of IT firms and pharmaceutical companies, through which pharmaceutical and medical devices companies can pursue global growth, achieve cost reduction, improve regulatory compliance and above all, benefit the human race.