Large language models for post-discharge follow-up in erectile dysfunction care: a narrative review

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Urology

Background:

• Erectile dysfunction (ED) is prevalent and linked to cardiometabolic and psychological issues.
• Effective ED treatment requires robust post-discharge follow-up, often challenged by stigma and fragmented care.
• Large language models (LLMs) offer potential solutions to these follow-up barriers.

Purpose of the Study:

• To review the potential role, risks, and implementation of LLM-assisted follow-up in ED care.
• To explore how LLMs can address communication barriers and enhance patient monitoring in ED.
• To identify necessary governance and oversight for safe LLM integration.

Main Methods:

• A targeted literature search was conducted in PubMed, Embase, and Google Scholar.
• Relevant evidence on ED follow-up, conversational AI, and safety governance was synthesized thematically.
• Included 280 sources, with 10 ED-specific and 270 from related chronic disease settings.

Main Results:

• LLMs can facilitate symptom check-ins, reinforce education, and improve outcome collection.
• Potential benefits are highest for structured, low-risk tasks like adherence support.
• Key risks include inaccurate recommendations, privacy breaches, and inadequate handling of sensitive issues.

Conclusions:

• LLM-assisted follow-up can be a valuable adjunct in ED care with clear boundaries and human oversight.
• A nurse-supervised, trigger-based model can balance feasibility and safety.
• Future research should focus on patient acceptability, safety, and implementation feasibility.