Because pharma is one the world’s most regulated industries — with continuously evolving frameworks, guidelines and reporting requirements — drug developers must keep pace with regulatory changes to avoid compliance issues. But too often, the traditional document-centric approach for solving regulatory problems taken by pharma companies creates barriers in terms of efficiency, collaboration and compliance.
Many of the issues involved with pharma’s traditional approach to maintaining compliance stem from the heavy manual and repetitive workflows that are involved. For example, tasks such as developing regulatory submissions, keeping labels up to date, understanding guidelines, and maintaining compliance with constantly shifting regulations require regulatory teams to access and analyze critical data scattered across vast amounts of documents.
Pharma companies expend a significant investment on these activities in the form of time, money and effort that adds up on the cost side of the ledger but does little to enhance revenue. As a result, regulatory teams are looking for digital transformations that will help them evolve from the traditional document driven-approach to more modern, data-driven methods.
Through digital transformation, innovative technologies and systems can help regulatory teams discover essential data within regulatory documents, then extract and standardize this information for use in downstream processes like reporting, labeling, master data management or structured content authoring.
To accomplish these objectives, leading pharmaceutical teams are looking to artificial intelligence-based technologies to automate processes, improve efficiency and ensure accurate reporting. One of these notable technologies is natural language processing (NLP), which uses algorithms to ‘read’ through unstructured text, then transform it into structured data that is suitable for analysis and visualization.
NLP in pharma
NLP is a subset of AI that pertains to how computers and human language interact. In lay terms, think of it as computers gaining the ability to understand, interpret and generate natural language. With NLP, computers can process, analyze and derive meaning from human language from a wide range of text inputs, including emails, social media posts, reports and reviews, in a way that is similar to humans.
Those capabilities are beneficial to pharma companies’ regulatory teams in particular, because they frequently must comb through substantially large collections of documents to discover new information and answer specific research questions. NLP enables teams to identify facts, relationships and assertions that may otherwise remain buried among hard-to-scrutinize mounds of unstructured data within regulatory documents and other sources.
Once converted to a structured format, this information can be integrated into databases, data warehouses or business intelligence dashboards and is incredibly valuable for a wide range of use cases pertaining to descriptive, prescriptive or predictive analytics.
NLP is not a new technology, but it has become widely used in recent years in consumer products, such as
Siri, Alexa and Google’s voice search. More recently, the emergence of generative AI built on large language models (such as ChatGPT), has generated new interest in applications of NLP in a broad range of disciplines such as medical research, customer care, fraud detection and risk management.
As it relates to the pharma industry, NLP has become an indispensable tool aiding in drug discovery, development and commercialization, largely because it significantly outperforms previous methods of search and information discovery. For example, traditional methods of search merely point to the location of documents, tasking human researchers with the issue of potentially spending hours of time to read through large amounts of individual documents to pick out necessary data.
Finally, NLP is particularly well-suited for health care data discovery due to the prevalence of unstructured data in the industry, from electronic health records to medical images to social media posts. Overall, pharma companies have used NLP to boost the accuracy and efficiency of the drug development life cycle by unlocking key information from the many unstructured data sources that are relevant to the industry.
NLP and compliance: A winning combination
NLP delivers value to drug developers across a variety of purposes within regulatory affairs by speeding compliance, boosting labeling processes, standardizing regulatory data, mapping to master data management systems, and driving digital transformation.
NLP offers benefits in many areas, including three major regulatory disciplines:
- Labeling: Access to drug labels from prominent regulatory authorities is important to help labeling teams find reference information for disease and symptom terms, contraindications, adverse events and special populations.
- Intelligence: Access to the landscape of regulatory updates, with integrated data flows to consume textual documents, both internal (such as corrective and preventive actions) and external (such as regulatory guidelines and FDA letters) is essential for regulatory teams.
- Mapping: Compliance teams need a means of finding key data attributes from unstructured text documents and mapping that data to standards, such as Identification of Medicinal Products (IDMP), a set of international standards that define the rules that uniquely identify medical products.
NLP in the real world
The following use cases demonstrate some real-world examples of how life science companies are using NLP to support regulatory operations.
Internal and external risk management
A large pharma company’s product development team sought more efficient ways of understanding internal and external risk management information to optimize the formulations, commercial supply and post-market regulatory compliance of its products.
The team created a data lake to capture important internal and external feeds. External feeds included FDA warning letters, biological license application review reports, white papers and industry benchmark repositories, while internal feeds included deviations, corrective and preventative actions, risks and response to regulatory questions.
The initiative relied on NLP to structure and generate the data by extracting critical concepts, relationships and sentiments from the sea of information. User-friendly visualizations enabled team members to drill down and navigate the information, driving wider use of the technology among the team. These internal and external data flows were updated automatically to deliver scalable reporting of the regulatory landscape, featuring key risks and recommendations to act upon.
Semi-automated regulatory intelligence tracking
Regulatory and compliance teams often employ manual approaches to monitor regulatory authorities, such as requiring team members to check relevant regulatory websites and subscribe to emails to remain informed of changes to guidelines.
While the process does generate needed intelligence to uncover concerns, deadlines, events and regulatory decisions for compounds of interest, it is a costly and resource-intensive approach.
To overcome the limitations of manual processes, one leading agrochemical company used NLP to develop a workflow to semi-automate information acquisition and summaries. The company integrated NLP with large language model technology to create a regulatory intelligence assistant, which provided team members with user-friendly question-and-answer access to updated regulatory information, safety alerts and risk categorization for compounds of interest.
Access to drug labels for more effective authoring
A leading drug developer used NLP to more efficiently explore and classify drug label data, helping its global regulatory affairs team overcome challenges associated with identifying and accessing label content from diverse sources in multiple languages.
To accomplish this goal, the company used an NLP-powered labeling intelligence hub, which synthesized drug label information across key sources such as the FDA and EMA. The hub enabled users to compare specific labels through an interactive view and access digitalized and original documents directly. The tool helped the team streamline processes associated with developing new labels and updating existing ones, expediting regulatory approval.
An adaptive approach
Regulations are continually shifting and evolving so it’s important for pharma companies to adapt their approaches to assist with regulatory review and compliance.
In the past, the industry relied on traditional search methods to uncover essential regulatory data, but with new advances in AI-based technologies such as NLP, there are options that can replace these inefficient, laborious and error-prone methods. With NLP, internal and external data can be transformed into high-value, actionable insights, enabling regulatory teams to unlock important supporting evidence to rapidly address today’s critical business issues.
