Rule-based natural language processing to extract clinical trial and research study enrollment history from...

Area of Science:

• Health Informatics
• Clinical Research
• Natural Language Processing

Background:

• Tracking clinical trial participation across facilities and sponsors is challenging.
• A systematic approach to monitoring patient enrollment in clinical trials is needed.
• Electronic health records (EHRs) contain valuable clinical trial information.

Purpose of the Study:

• To develop and validate natural language processing (NLP) methods for extracting clinical trial enrollment data from EHRs.
• To create a novel data resource for analyzing and tracking trial enrollment at a population level.
• To assess the feasibility of using NLP to capture trial enrollment nationwide.

Main Methods:

• Utilized natural language processing (NLP) techniques to extract study enrollment status, consent date, and study title from clinical notes in the Veterans Affairs EHR.
• Developed and tested a classifier to identify clinical trial participants within a large healthcare system.
• Evaluated the precision and recall of the NLP method for key enrollment data points.

Main Results:

• The NLP method achieved high test-set precision for enrollment status (0.94), consent date (0.97), and study title (0.87).
• Recall for these metrics was acceptably high (0.76, 0.70, and 0.84, respectively).
• A single-center classifier accurately identified 88.8% of trial participants across 12 distinct trials.

Conclusions:

• Natural language processing (NLP) is a feasible method for capturing clinical trial enrollment data from EHRs.
• This NLP algorithm provides a novel data resource for population-level analysis of clinical trial participation.
• The developed methods can enhance the systematic tracking of clinical trial enrollment across a nationwide healthcare system.