Automated Auditing of Hospital Discharge Summaries for Care Transitions
Summary: arXiv:2604.05435v1 Announce Type: new
Abstract: Incomplete or inconsistent discharge documentation is a primary driver of care fragmentation and avoidable readmissions. Despite its critical role in patient safety, auditing discharge summaries relies heavily on manual review and is difficult to scale. We propose an automated framework for large-scale auditing of discharge summaries using locally deployed Large Language Models (LLMs). Our approach operationalizes core transition-of-care requirements such as follow-up instructions, medication history and changes, patient information and clinical course, etc. into a structured validation checklist of questions based on the DISCHARGED framework. Using adult inpatient summaries from the MIMIC-IV database, we utilize a privacy-preserving LLM to identify the presence, absence, or ambiguity of key documentation elements. This work demonstrates the feasibility of scalable, automated clinical auditing and provides a foundation for systematic quality improvement in electronic health record documentation.
Introduction
Hospital discharge summaries are critical documents that ensure continuity of care for patients transitioning from inpatient to outpatient settings. However, issues such as incomplete information or inconsistent documentation often lead to care fragmentation, which can result in avoidable readmissions. This highlights the importance of effective auditing processes for discharge summaries.
The Need for Automation
Traditionally, auditing discharge summaries has been a manual process, heavily reliant on healthcare professionals to review each document for completeness and accuracy. This method is not only labor-intensive but also challenging to scale, especially in large healthcare systems. As healthcare data continues to grow, there is an urgent need for innovative solutions that can enhance the efficiency and effectiveness of this auditing process.
Proposed Framework
To address the limitations of manual auditing, we propose an automated framework that leverages Large Language Models (LLMs) to conduct large-scale audits of discharge summaries. Our framework operationalizes core transition-of-care requirements into a structured validation checklist based on the DISCHARGED framework, which includes:
- Follow-up instructions
- Medication history and changes
- Patient information
- Clinical course summaries
By utilizing adult inpatient summaries from the MIMIC-IV database, our approach employs a privacy-preserving LLM to assess the presence, absence, or ambiguity of these key documentation elements.
Benefits of Automated Auditing
The implementation of an automated auditing framework offers several advantages:
- Scalability: The use of LLMs allows for auditing at a scale not feasible with manual reviews, enabling healthcare institutions to ensure quality across a larger volume of discharge summaries.
- Consistency: Automated systems can provide uniform assessments that reduce variability in the auditing process, enhancing reliability.
- Efficiency: Reducing the time and labor associated with manual audits allows healthcare professionals to focus on patient care rather than administrative tasks.
- Quality Improvement: By systematically identifying documentation deficiencies, healthcare providers can implement targeted interventions to improve electronic health record documentation.
Conclusion
The transition of care is a critical phase in the patient journey, and effective discharge summaries play a vital role in ensuring safety and continuity. By employing an automated auditing framework using LLMs, healthcare institutions can enhance their auditing processes, leading to improved quality in documentation and ultimately better patient outcomes. This work not only demonstrates the feasibility of scalable clinical auditing but also lays the groundwork for ongoing quality improvement initiatives in healthcare.