CancerClarity app: Enhancing cancer data visualization with AI-generated narratives

Abstract

Background: Community cancer centers face challenges in accessing cancer data and communicating health information to patients and community members due to limited tools and resources. The CancerClarity app, recognized at the 2023 Catchment Area Data Conference Hackathon, addresses this need by integrating data visualization with Artificial intelligence (AI)-driven narrative generation. Converting quantitative cancer statistics to narrative descriptions using large language models (LLMs) may help cancer centers communicate complex cancer data more effectively to diverse stakeholders.Methods: The CancerClarity app employs LLM prompting within the R Shiny web framework, sourcing data from Cancer InFocus. It offers users an interactive exploration of cancer incidence, mortality, and health determinants across U.S. counties.Results: The CancerClarity app integrates LLM via its application programming interface (API) for real-time, linguistically tailored narratives, making cancer data accessible to a broad audience. The app offers cancer centers a cost-effective solution to swiftly identify their catchment areas and assess the cancer burden within the populations they serve.Discussion: By enhancing public health decision-making through AI-driven narratives, the app underscores the critical role of effective communication in public health. Future enhancements include the integration of Retrieval Augmented Generation (RAG) for improved AI responses and evidence-based public health guidance.

Keywords

• Artificial intelligence
• large language models
• low-graph literacy
• cancer-related communication
• cancer statistics