The regulatory drive to increase the amount of data and improve the quality of pharmacovigilance (PhV) reporting systems is key to maintaining the databases at the forefront of signal detection.
In a recent report by a subgroup of the European Medicines Agency (EMA), considerations and recommendations are made to regulatory authorities on achieving the best performance of signal detection systems.
Pharmacovigilance reporting systems provide critical data from which the majority of signals of adverse drug reactions (ADRs) originate. Sources of information such as healthcare professionals (HCP), patients, legal representatives, literature articles and clinical trials produce a multidimensional and extensive dataset.
The European Union (EU) PhV legislation has stressed the role of spontaneous reporting by HCP’s or patients, either directly or through the marketing authorization holder. These systems are competent at detecting new and existing risks of medicines, despite the issue of under-reporting. Along with under-reporting, missing information and reporting of duplicates are limitations of PhV reporting systems as well as misclassification, information bias, increase in databases size and variability of safety concerns.
Increasing the use of electronic reporting or mobile devices can encourage spontaneous reporting as much as possible. Increasing access to gateways such as electronic health records (EHR), social media reports and PhV databases could allow combination different data sources in order to achieve the best performance of the system for signal detection.
Furthermore, by expanding the dimensions of the data collected, such as features and variables, it may be possible to move beyond using the data for risk identification only. However, the potential increased availability and variety of data will inevitably bring new challenges associated with evaluating increasingly large datasets and gaining useful insights to guide and inform decision making.
Hence, regulatory authorities should consider:
- Investing in methods to integrate PhV and other real-world data with non-clinical data and methods to validate data integration;
- Exploring the use of new analytical tools, such as forecasting and machine learning, that leverage increased dimensions of data (spatial-temporal, other variables in case reports, meta-data);
- Fostering reproducible research as a way to ensure the network is made aware and can learn and execute new analytical approaches;
- Exploring how to implement automation and natural language processing to improve efficiency, data management quality and free expert reviewer time.
Regulatory authorities have an opportunity to facilitate data integration and harness its potential. Increasing efficiency and capacity should be the two main drivers behind improving data analysis of ADRs.
Click here to view the full report.