Beyond Binary Detection: A Multi-Dimensional Taxonomy of Cancer Misinformation on Reddit

2026-07-14

AI Summary

This study addresses the limitations of existing cancer misinformation research, which often relies on binary labels, narrow definitions, and limited datasets, thereby failing to capture the complexity of misinformation on social media. To overcome these constraints, the authors propose the first seven-dimensional, multi-granular classification framework tailored for social media content, encompassing dimensions such as presence, information type, risk level, stance, and topic. Leveraging expert annotations and a few-shot prompting strategy with large language models (LLMs), the framework enables automated fine-grained classification across four major cancer-related Reddit communities. The analysis reveals that approximately 6% of discussions contain misinformation, with significant variation across communities and topics. Furthermore, few-shot prompting substantially enhances classification performance and uncovers three prevalent narrative patterns in cancer-related misinformation.

Abstract

Cancer-related discussions on social media provide an important space for information exchange and peer support, but also facilitate the spread of misinformation that may influence prevention, screening, and treatment decisions. Existing research on cancer misinformation often relies on narrow definitions, small-scale datasets, or binary labeling frameworks. We introduce a multi-dimensional taxonomy for characterizing cancer misinformation in Reddit discussions of breast, lung, colon, and prostate cancer. The taxonomy captures seven dimensions, including misinformation presence, information type, risk level, stance, and topical focus. Using expert-annotated data, we evaluate multiple large language models (LLMs) for scalable misinformation annotation and analyze cancer misinformation across Reddit communities. Our results show that cancer-related misinformation constitutes approximately 6\% of Reddit cancer discussions, with substantial variation across communities and misinformation topics. Few-shot prompting substantially improves classification performance, particularly for nuanced taxonomy dimensions. We additionally identify recurring misinformation narratives centered on unsupported treatments, distrust of conventional medicine, and misleading claims about diagnosis and screening. Our taxonomy, dataset, and findings provide a foundation for multi-dimensional modeling of online cancer misinformation.

Problem

• cancer misinformation
• social media
• multi-dimensional taxonomy
• Reddit
• misinformation detection

Innovation

• multi-dimensional taxonomy
• cancer misinformation
• large language models
• few-shot prompting
• social media analysis