The growing focus on specialty drugs, including more personalised and targeted treatments in oncology and for rare disease, including new therapeutics such as CAR T-cell therapies, strong drug safety/pharmacovigilance (PV) oversight is imperative – not least linked to new mechanisms of action. Many regulators are striving to accelerate access to important new drugs too, making effective real-world monitoring even more critical than it ever has been.
For drug companies, the practical PV challenge is real and growing. General economic/market pressure to be more cost-efficient in routine processes must be balanced with a continued commitment to quality and safety. It follows that effective, attentive, proactive tracking of adverse events right across a medication’s lifecycle demands process innovation.
Medical literature monitoring – the systematic review of published reports of adverse events – is already a key target for process transformation. Globally, this practice is relatively straightforward – drawing on well-known, indexed journals via large databases – rending it ripe for transformation through technology-enabled automation.
The screening of local literature, by contrast (critical for identifying adverse events relevant to specific countries or populations), remains inefficient and error-prone.
It involves the manual tracking of region-specific sources, including non-indexed journals and local publications. This resource-intensive activity is often outsourced yet, without efficient and reliable processes, costs can run high and returns can seem low.
The scale of the local monitoring challenge
To date, the protracted nature of local literature monitoring has been an accepted frustration. Considered a necessary evil, in-country literature monitoring is the only way to ensure that adverse drug reactions and safety signals published in local journals are identified and reported. These sources often contain critical safety data that may not appear in global databases.
Comprehensive local literature monitoring is generally expected if not mandated by health authorities as part of PV reporting. It completes the global picture, and is a way of identifying population-specific adverse events – important trends which otherwise might be missed. Local sources including journals, web sites or print sources provide crucial insights at the country level. If gaps in routine monitoring are discovered during inspections or audits, this could have implications for ongoing licensing and sales, as well as brand reputation.
Strategically, local insights enable better decision-making, allowing pharmaceutical companies as well as healthcare systems to respond proactively to emerging safety concerns, and adapt labelling or guidance at a country level. By extension, discrete local findings can also serve as an early warning system, and provide insights for ongoing product development.
Despite its recognised importance, local literature monitoring remains deeply inefficient when the activity is carried out in the traditional manual way. The screening process often necessitates the commitment dedicated staff at an affiliate or regional level, potentially with local language capabilities and journal access.
In addition, despite the large volume of data being reviewed, local monitoring typically yields only limited safety information. Errors and omissions are common.
Because of its intense administration burden, local literature monitoring is usually outsourced to clinical research organisations, to alleviate the pressure on in-house PV scientists. Usually, the numerous designated sources are listed in multiple, long, and unwieldy Excel spreadsheets, with assigned teams expected to monitor several thousand different web sites on a weekly or monthly basis. Each row in each spreadsheet contains a web link which needs to be searched, the results then captured and populated in the same Excel file. At the end of each cycle, all of the disparate Excel files are aggregated and distributed as needed.
The problem is compounded by huge variances in the literature format, literature access issues, and language barriers. Local sources often lack consistency in format, indexing, and language, making it difficult to implement a simple unified process, meanwhile many local journals require paid subscriptions or may be only available in print.
On top of this, regulatory reporting timelines vary by country, something else that has had to be tracked manually to ensure respective adherence. In the EU and Australia, MAHs are expected to conduct a literature review at least once a week. Within Europe, individual countries have differing expectations. Some EU competent authorities have a required list of local sources to be reviewed; others have recommended lists; and others don’t (yet) specify which sources should be monitored.
The data/tech difference: how capabilities are aligning to enable transformation
It is for all of these reasons that the pharma industry and its service provider community are looking to next generations of automation technology for an answer, both to improved cost-efficiency in local literature monitoring and to increased accuracy and reliability.
Processes powered by advanced technology, including large language models (LLMs), are proving instrumental in structuring data now, for instance – enabling “normalisation”, unification, centralised management, and governance, as pharma transitions to “data-first” ways of working. In addition to improved data structuring, advanced “crawling” techniques are transforming automated browsing, “scraping”, and indexing of content from target web sites and publications. AI then adds a layer on top of that, making it possible to search all of that content and very quickly and identify safety events, in one fell swoop, via a single unified database.
Not every pharma organisation will need these advanced facilities. For a small company that wants to search just three countries, with perhaps up to 10 journals in each, manual searches are unlikely to be onerous. But once that burden multiplies, to become hundreds or thousands of sources which need to be reviewed weekly or monthly, the workload soon becomes untenable without the help of robust automation to ensure cost-efficiency, speed, and accuracy – as well as instant reporting.
Patient safety can never be sacrificed in the interests of process efficiency, though. In the life sciences, there will always be an important role for human oversight and process governance. Technology-assisted human ingestion is another option, where companies are more hesitant about immediate technology reliance.
Yet persisting without automation is hard to justify now. Most pharma companies have come to recognise that, unless they buy into technology-enabled process innovation, they will struggle to keep pace with the competition, and with soaring operational costs in a challenging market. Even regulators accept this, and are becoming increasingly open to automation as a means of improving compliance and patient outcomes.
Certainly, PV workloads are not diminishing and PV scientists are already overstretched and keen to have more time to perform actual safety assessments – instead of searching web sites, performing manual data entry tasks, and trying to decipher messy spreadsheets. With easy, rapid access to the data they need, they could skip to the more fulfilling and higher-value activities that form the core of their role. Automation further enables PV scientists to better manage the ever-growing workload associated with the increase in publications year on year.
The bigger picture: earlier insights into emerging trends
Finally, but crucially, this isn’t about adopting technology for its own sake because of a performance promise. Companies looking for an advantage through automation in local literature monitoring must do so in the context of this being a regulated activity. That is, process innovation should be introduced in a balanced way that will not compromise delivery timelines, future audits, or other fundamental requirements.
Looking ahead, as companies do transition away from laborious manual processes in all aspects of their literature monitoring, the strategic potential lies in the new, richer, data-driven insights they will gain about safety trends. Ultimately, this is an opportunity for companies to better understand the safety trends around their drugs, and at a more discrete level. The value of having earlier visibility into trends in different patient populations, in different countries and regions, will soon become apparent.
As well as informing ongoing drug discovery and development, these insights could also inform the respective healthcare system, and patient journey, in a given region, with wider societal benefits.
This will be thanks to the facility to hone searches, and more efficiently identify true safety events –by sifting the highest available volume of information to confidently arrive at the smallest number of safety events for review; trusting that no vital detail has passed unnoticed.
