Pre-consultation gives healthcare providersa history of present illness (HPI) prior to a patient's visit,streamlining the visit and promoting shared decision making.Compared to a digital questionnaire,LLM-powered AI agents have proven successful inproviding a more natural interface for pre-consultation.But LLM-based approaches struggle to ask productive follow-up questions andrequire complex prompts to guide the consultation.While effective automated prompting strategies exist for medicalquestion-answering LLMs, the task of question generation for pre-consultationis lacking effective strategies.In this study, we develop a methodology for evaluating existing approaches to medical pre-consultation,using prior datasets of HPIs and patient-doctor dialogue.We propose a novel approach of converting abundant clinical note datainto question generation demonstrations and then retrieving relevantdemonstrations for in-context learning.We find this approach to question generation for pre-consultationachieves a higher recall of facts in ground truth consultationscompared against competitive baselines in prior literature across a range of simultated patient personalities.