To save time spent on manual reporting, pharmaceutical companies are turning to Natural Language Processing (NLP) which holds the combined potential for automating processes and enabling a deeper understanding of previously untapped data.
The amount of available health data has exploded in recent years – increasing by 878 percent since 2016 and climbing. This rise in information volume and complexity has made analysis and reporting of safety and compliance data a near-impossible undertaking without the assistance of technology. Manual reporting is no longer necessary or productive - pharmaceutical companies can dedicate this time to higher-value initiatives which will drive greater innovation and value.
Natural Language Processing (NLP) is one tool pharmaceutical companies are starting to leverage in pursuit of digital transformation. NLP holds the combined potential for automating processes and enabling a deeper understanding of previously untapped data. NLP can help pharmaceutical companies maintain compliance in a way that simultaneously boosts productivity and reorients the compliance functions toward an understanding of their products that would have otherwise been impossible.
Currently, more than 60 percent of time and resources spent on pharmacovigilance activities are dedicated to administrative tasks, including data extraction and quality checking. This estimate does not account for time spent on staff training - which must happen as regularly as reporting requirements change. Put simply, humans can no longer achieve compliance using manual processing alone. Trying to operate in this way amounts to massive hours spent on internal tasks that ultimately may not capture everything – certainly not without many mistakes resulting from human error.
Pharmaceutical professionals should consider NLP to address these productivity issues. In one example, CSL Behring, a global biopharma company developing rare disease drugs, was able to double its accuracy of adverse event auto-coding to Medical Dictionary for Regulatory Activities (MedDRA) from 30 percent to 60 percent with one-to-one verbatim text matching alone. (NLP technology therefore cut manual coding down from 70 percent to only 30 percent.) Freeing up resources from manual reporting burdens will reduce the overall hours spent on routine pharmacovigilance tasks, while reducing costly errors. Most importantly, this will enable companies to reinvest in activities that do drive true business value.
Leveraging NLP to mine and convert unstructured text-based documents into a readable format for computers is not a new concept for the pharma industry. It has already been used extensively in various aspects of clinical development, such as trial participant identification and mining medical literature for important insights in drug development. While still in the more nascent stages of adoption for applications in the compliance space, the potential for tapping into rich safety insights is incontrovertibly strong with 80 percent of health care data currently residing in unstructured formats.
For adverse event reporting, this unstructured data is especially important because it encompasses repositories for patient-reported outcomes like discussion boards and online forums, doctors’ notes captured in electronic health records, social media platforms, notes from medical call centers, and more. Given that patients often use different terms and phrases to explain the adverse events they may have experienced (e.g., “sleeplessness” as compared with “unable to fall asleep”), NLP becomes a necessary asset to consider context and the intricacies of natural language throughout massive data sets.
A newfound understanding of natural language-based data will provide a competitive edge for biopharmaceutical companies. It can help them deepen their understanding of their current offerings and may enable them to identify entirely new opportunities for clinical development. Through the same channels that detect adverse events, companies may discover new indications for which their products are proving effective. Expanding the depth and breadth of knowledge-especially with direct, real-time insights from patients - puts businesses in the driver’s seat to solve existing problems and proactively uncover emerging opportunities.