COREQ+LLM: Developing Reporting Guidelines for the Use of Large Language Models in Qualitative Research

Introduction: Qualitative research is crucial for capturing the complex human experiences in the health and social sciences. Since 2007, the COREQ checklist has served as an important framework for transparent reporting. The increasing use of large language models in qualitative research brings both opportunities and challenges. These models can support analysis and interpretation, but they also raise concerns about variability in results, potential biases, and limited reproducibility. Such issues are not sufficiently addressed by current reporting standards.

Methods: The COREQ+LLM project aims to extend the COREQ checklist by developing a framework specifically designed for studies that utilize large language models. The development process follows the recommendations of the EQUATOR Network and involves several steps. It begins with mapping existing applications of large language models in qualitative research and continues with a Delphi study involving 141 international experts in artificial intelligence and qualitative research. The Delphi process is designed to reach consensus on the most relevant reporting items. Early results from the Delphi study already indicate a high level of agreement among participants.

Results: The emerging guideline is organized around key domains. These include the research team and reflexivity, with attention to the role and qualifications of operators, the study design and theoretical framework, focusing on the role of large language models in analysis and human interaction with these systems, as well as the description of model access and parameters. Further domains address prompting strategies and their refinement, management of research context across sessions, procedures for analysis, validation, bias mitigation and ethical aspects.

Conclusions: COREQ+LLM will provide the first dedicated framework for documenting the use of large language models in qualitative research. By fostering methodological transparency and ethical accountability, the extension will support researchers in reporting their work, enable reviewers and editors to assess rigor, and strengthen trust in AI-assisted qualitative inquiry.