AI-assisted learning: ChatGPT for anamnesis in women's health education

Clinical Snapshot

Key Findings

• P Value: p < 0.0001

• Effect Size: Not quantified

• Primary Outcome: Significant improvement in anamnesis performance after ChatGPT-assisted learning

• Nnt Or Sensitivity: Not applicable - educational intervention study

• Confidence Interval: Not reported

Clinical Application

Requires faculty development and institutional support for responsible implementation Relevant to Australian medical schools but would need adaptation to local curriculum standards and RACGP/RANZCOG training requirements Medical students in women's health education, particularly in obstetrics and gynaecology training

Abstract

UNLABELLED: Artificial intelligence (AI) tools such as ChatGPT are rapidly being adopted in medical education, yet evidence regarding their impact on learners' history-taking (anamnesis) skills remains limited. This gap is particularly relevant in women's health education, where high-quality anamnesis is essential for addressing sensitive, context-dependent concerns and for ensuring the safety of two interrelated patients: mother and fetus. METHODS: This study comprised two phases. First, semi-structured virtual interviews were administered to students and faculty to assess their familiarity with ChatGPT and access to digital technologies. Second, an in-person classroom intervention was implemented within the Women's Health discipline, integrating ChatGPT into history-taking instruction. Students' anamnesis performance was evaluated before and after ChatGPT-assisted learning. Additionally, a novel instructional activity was introduced in which ChatGPT generated cases classified as plausible, intermediate, or physiologically impossible, serving as prompts for guided discussion focused on diagnostic pitfalls, critical appraisal, and clinical reasoning. RESULTS: In the virtual phase, faculty (n = 62) and students (n = 115) completed the questionnaires. The classroom intervention involved 40 medical students participating as part of routine Women's Health teaching activities, with faculty oversight to ensure pedagogical alignment and responsible AI use. A pronounced discrepancy emerged between students' familiarity with ChatGPT (99.1%) and lower levels of familiarity among faculty. Anamnesis performance improved significantly after ChatGPT-assisted learning (p < 0.0001). Students reported increased confidence and perceived enhancement of history-taking quality, and all participants viewed ChatGPT as a valuable educational resource. The AI-generated cases effectively stimulated discussion and reinforced clinical reasoning processes. Findings contributed to the development of a practical guideline for the responsible integration of ChatGPT into Women's Health education. CONCLUSIONS: ChatGPT can enhance anamnesis and promote critical reasoning through structured interaction with AI-generated cases. However, its effective and responsible use depends on targeted faculty development to ensure alignment with pedagogical goals, appropriate oversight, and consistent integration into curriculum.

References

1. 1.Espírito Santo, D. S., Lott Bezerra, T., Viana Lopes Gomes, M. C., Augusto de Souza Fonseca, P., & Joviano-Santos, J. V. (2026). AI-assisted learning: ChatGPT for anamnesis in women's health education. Medical Education Online, Article 2655562. https://doi.org/10.1080/10872981.2026.2655562