Natural Language Processing in Clinical Research Recruitment: A Scoping Review Enriched with Stakeholder Insights

Abstract

We conducted a scoping review to characterize natural language processing (NLP) applications in clinical trials recruitment and conducted semistructured interviews to obtain stakeholders’ perspectives on these technologies, with a focus on ethical considerations. The scoping review focused on English-language original articles published from January 2021 to June 2024, sourced from Ovid Medline. Data extracted included the characteristics of NLP systems, their evaluations, and ethical considerations regarding patient autonomy and equity. Additionally, semistructured interviews with experts from various specialties were conducted, and the data were analyzed using thematic analysis. Most of the 47 eligible articles focused on NLP models for electronic health records screening. The literature predominantly emphasized the models’ accuracy and efficiency, while ethical considerations received little attention. Interview findings underscored the need for more ethical reflection and real-world implementation analysis, revealing differing opinions on anonymization, consent, and the impact of NLP tools on fair opportunities. NLP applications for participant recruitment in clinical research are in early stages, with a gap between ethical discourse and reporting in current literature. Practical guidelines are needed for implementing and reporting ethical aspects throughout the lifecycle of NLP applications, along with empirical research to assess their ethical impact.