Literature is regarded as a significant source of information in the monitoring of a medicinal product, and its safety profile and benefit-risk balance. Literature surveillance can be a key tool for the detection of new or emerging safety signals/issues.
During clinical development it is mandated that drug developers perform literature surveillance in accordance with and in support of other surveillance activities, such as ongoing safety evaluation or annual safety reporting (i.e. DSURs). This obligation only intensifies when the drug developer submits a new drug application (NDA) or marketing authorization application (MAA), and receives approval for marketing authorization, since it then becomes a requirement to perform a systematic literature review no less frequently than once a week. It is also expected that NDA and marketing authorization holders have procedures in place to monitor local publications in countries where the medicinal product has been authorized to be marketed.
Given the amount of available literature, and the number of potential sources (i.e. Embase, Medline, Reactions Weekly, Biosis), it is optimistic to expect to find every piece of literature that might be published. A realistic, pragmatic and objective approach should be taken, coupled with the development of appropriate targeted search strategies, to ensure that the recall (i.e. number of results) is as precise (i.e. number of results that are considered relevant) as possible. Does anybody really want to review 10,000 results only to find 1 that is relevant for the medicinal product?
Search strategies should be designed in such a way to meet the objective of the search. Is the objective to look for general safety information or individual case safety reports (ICSRs), to identify “missing” information about the medicinal product, or to monitor a particular type of safety signal? It is critical to build a robust and targeted strategy, which can evolve as more about the medicinal product is learnt and the objective of the search changes. In the clinical development setting, exposure to a medicinal product is within a controlled setting and can be limited to a certain patient population. When the medicinal product moves to market the exposure may widen to other types of patient population and may result in new or emerging safety information being observed.
As a way of putting search strategy design into perspective, below is a very simple example of how a single search term (or tag) can alter the amount of recall (in this example the number of results decreased by 97%):
On Friday 24 Aug 2018, a PubMed search for Ibuprofen produced 13,468 results.
On Friday 24 Aug 2018, a PubMed search for Ibuprofen AND headache produced 310 results.
Despite the very best efforts, and the most sophisticated of strategies, the amount of literature recalled can still be overwhelming particularly for those companies with hundreds of medicinal products in clinical development and/or on the market. As with most things, technology and automation can help.
Robust platforms for literature surveillance, such as DistillerSR from Evidence Partners, can help reduce the burden and increase efficiencies for literature search and review. Technology allows for standardized and automated workflows that maximize literature operations, audit trail as well as offering a single standalone platform/repository for the storage of such surveillance data. When needing to search multiple data sources, technology can perform de-duplication (i.e. quarantine duplicate results) as well as allowing for keyword highlighting to expedite literature reviews and shorten the timeframe in finding the relevant information. Implementation of such robust platforms can allow for seamless oversight and coordination of literature surveillance activities, ensuring inspection readiness as well as enabling the ongoing monitoring of the performance of the process and search strategy through the generation of on-demand reports.