Enhancing Primary-to-Specialist Care Transitions with an LLM Chatbot: Results from a Randomized Controlled Trial

The study on the PreA healthcare platform, approved by various ethical committees, aimed to enhance consultation efficiency and patient-centered care via a patient-facing chatbot and clinician interface. Informed consent from participants was obtained, adhering to ethical guidelines. The co-designed architecture facilitated data collection through voice/text from patients, generating structured referral reports for clinicians. The project involved two refinement cycles, engaging various stakeholders, including 120 patients, healthcare workers, and physicians, to address disparities in healthcare delivery. A randomized controlled trial (RCT) with over 2,000 participants evaluated the effectiveness of the PreA system across three groups: PreA-only, PreA with human support, and No-PreA. Key outcomes focused on consultation duration, care coordination, and communication ease. Statistical analyses examined patient outcomes and clinical notes to assess the impact of the intervention. The methodology aligns with WHO guidelines for ethical AI implementation in healthcare, ensuring comprehensive healthcare delivery and system utility. The trial is registered as ChiCTR2400094159.

Source link