Presentation Details
| Enhancing Preoperative Evaluation Through AI-Enabled Patient Summaries and Dynamic Risk Stratification Using Health Information Exchange Data Phillip Lau1, 2. 1Pacific Valley Medical Group, Calabasas, CA, USA.2Huntington Health, affiliate of Cedars-Sinai, Pasadena, CA, USA |
Abstract
BACKGROUND: Accurate risk assessment and efficient data synthesis is central to perioperative safety and surgical readiness. Clinicians often navigate fragmented information across electronic health records (EHRs) and external systems, leading to incomplete risk profiles, redundant review, and workflow delays. To address these challenges, an artificial intelligence (AI) – enabled platform was developed to automatically generate structured patient summaries and provide dynamic perioperative risk stratification using Health Information Exchange (HIE) data.
PURPOSE: Existing preoperative workflows were analyzed to identify pain points in data access, manual chart review, and variability in risk assessment. The platform extracts structured and unstructured data, including medical history, comorbidities, medications, laboratory values, and prior procedures, from both institutional EHRs and regional HIE feeds. Natural language processing and predictive modeling were used to create concise patient summaries and calculate individualized risk scores for perioperative complications (e.g. cardiovascular, pulmonary, or functional risks). The platform was evaluated through pilot cases and structured clinician feedback to assess accuracy, clinical relevance, and usability.
RESULTS: Automated patient summaries accurately captured key medical information and active problems, presenting them in a standardized, clinician-friendly format. Early clinician feedback revealed opportunities to expand the scope of extracted data to include additional comorbidities and clinical parameters influencing perioperative readiness. Reviewers noted that while the AI effectively synthesized documentation, a broader condition set would allow nurses to rely on the summaries and comprehensive rather than selective risk review. Dynamic risk outputs aligned with perioperative guidelines and were consistently judged to improve efficiency and situational awareness during preoperative assessment.
CONCLUSIONS: Leveraging AI and HIE data to automate patient summaries and risk stratification represents a scalable innovation in perioperative care. By consolidating disparate health information and presenting actionable risk insights, this approach supports more efficient, consistent, and informed preoperative evaluations, laying the groundwork for future integration of automated testing recommendations and readiness tools.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
PURPOSE: Existing preoperative workflows were analyzed to identify pain points in data access, manual chart review, and variability in risk assessment. The platform extracts structured and unstructured data, including medical history, comorbidities, medications, laboratory values, and prior procedures, from both institutional EHRs and regional HIE feeds. Natural language processing and predictive modeling were used to create concise patient summaries and calculate individualized risk scores for perioperative complications (e.g. cardiovascular, pulmonary, or functional risks). The platform was evaluated through pilot cases and structured clinician feedback to assess accuracy, clinical relevance, and usability.
RESULTS: Automated patient summaries accurately captured key medical information and active problems, presenting them in a standardized, clinician-friendly format. Early clinician feedback revealed opportunities to expand the scope of extracted data to include additional comorbidities and clinical parameters influencing perioperative readiness. Reviewers noted that while the AI effectively synthesized documentation, a broader condition set would allow nurses to rely on the summaries and comprehensive rather than selective risk review. Dynamic risk outputs aligned with perioperative guidelines and were consistently judged to improve efficiency and situational awareness during preoperative assessment.
CONCLUSIONS: Leveraging AI and HIE data to automate patient summaries and risk stratification represents a scalable innovation in perioperative care. By consolidating disparate health information and presenting actionable risk insights, this approach supports more efficient, consistent, and informed preoperative evaluations, laying the groundwork for future integration of automated testing recommendations and readiness tools.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
Content Locked. Log into a registered attendee account to access this presentation.