When developing content for new drug labels, life sciences companies’ regulatory teams face challenges in gathering intelligence to better understand competitive factors, market dynamics, and effective approval strategies. To overcome these barriers, organizations are increasingly looking to technology such as natural language processing (NLP) to reduce the manual, repetitive steps of acquiring regulatory intelligence.
Drug labels hold value for life sciences companies because they are rich, extensive documents that can be used across labeling, regulatory, safety, and medical affairs. Typically, a drug label refers to any information provided with prescription drugs, as requested by regulators such as the U.S. Food and Drug Administration (FDA) or European Medicines Agency (EMA). The main label contains data that is generally in a semi-structured format, with sections covering: description of the drug, clinical pharmacology, indications (uses for the drug), contraindications (who should not take the drug), warnings and precautions, and adverse events (side effects).
Analyzing drug label information can provide a valuable source of competitive data for life sciences companies, enabling them to fulfill several use cases, including:
- Comparing labels for a class or group of drugs to assist with authoring
- Finding labels for products with similar attributes such as mechanisms of action or pharmacokinetics
- Extracting adverse reactions and normalizing them to improve comparative analysis
- Understanding patient sub-group exceptions for competitive or for label development purposes
- Extracting contra-indications to understand competitive coverage
Challenges in Working with Label Data
Despite the value that drug label data holds for life sciences companies, many struggle to efficiently access and extract it, resorting to traditional methods of search and chart review. Indeed, a host of challenges are associated with the comprehensive analysis of drug label information, which can be complex and inconsistent, such as:
- Manually intensive: When regulatory, medical, and safety teams are dependent on manual search and feature extraction, they must spend a significant amount of time searching for the right labels and label-based information.
- International sources: The focus for many labeling teams is on major sources such as FDA and EMA. However, some teams need additional English language sources, as well as sources from other countries in local languages.
- Multiple sources: Searching for specific terms and concepts across different websites can be complicated, manual, and time-consuming.
- Inconsistent, noisy data: Varied terminology around therapies, patients, adverse reactions, drug interactions, and many other elements of label content makes reliable analysis challenging
Life sciences companies are turning to NLP to alleviate these challenges. Drug label exploration using NLP solutions enable researchers to search for the right label and the right content from within the label, then extract that data for use in comparative analyses.
NLP techniques are particularly effective in enabling life sciences researchers to realize the full value of unstructured data. NLP automates text mining, helping machines “read” text by simulating the human ability to understand languages. This enables the analysis of unlimited amounts of text-based data without some of the limitations inherent to humans, such as fatigue and bias.
NLP holds several significant benefits in extracting essential drug label information, including:
- Users can save time and effort because text mining approaches facilitate more effective identification and extraction of the right label information.
- The user experience is simpler and more intuitive thanks to rich features such as NLP search, ontology navigation, and side-by-side comparison.
- The significant acceleration of time to insight and richness of results by taking advantage of pre-developed ontologies such as “drugs and diseases” which contain millions of synonyms.
- The ability to customize the solution to incorporate bespoke sources in a wide range of languages and formats.
Example: Top Pharma Company Using NLP to Explore Drug Label Data
Several teams within a top 10 pharma company—including global labeling, regulatory affairs, medical, and safety—needed to identify labels and label content from across sources and in multiple languages. Users had a clear need to search these sources in a more efficient way, generating less noisy results and enabling users to refine results and export them for further analysis.
A drug label exploration tool was deployed using FDA Drug Labels, EMA Drug Labels, as well as several European local databases. Users were able to build their own search queries and use pre-built queries to answer their most important questions and find the data they needed. They could also select specific labels for comparison through an interactive view and link through to original documents. Ultimately, such a solution can save the team time and make the process of developing new labels, updating existing labels, and securing regulatory approval faster and easier.
Life sciences companies understand the important role that drug label information plays in regulatory intelligence gathering, but often lack an efficient means of acquiring that information and organizing it in a manner suitable for analysis. By improving upon traditional methods of intelligence discovery, NLP helps regulatory teams save significant amounts of time and money compared with legacy approaches to manual search and feature extraction.