Can routinely collected electronic medical record (EMR) data support hospital resource allocation? A retrospective analysis of 332,711 presentations to a public quaternary teaching hospital in South Australia (2020-2025)

Background Hospitals face increasing strain from rising clinical complexity and demand. Traditional resource allocation approaches often lack the granularity and timeliness needed for responsive planning. This study evaluates whether routinely collected electronic medical record (EMR) data can be used to classify hospital inpatients into resource-based groups to support real-time planning and hospital-wide operational management. Methods A retrospective analysis was conducted on 332,711 inpatient admissions to a quaternary public hospital in South Australia between January 2020 and January 2025. Patients were classified into one of four flow streams within 72 hours of admission using a resource-based classification framework developed through a modified Delphi process and validated by clinical review. Summary statistics were used to assess differences in resource use across streams and to evaluate classification stability. Data quality limitations and documentation variability were also assessed. Results Flow streams demonstrated distinct differences in length of stay, diagnostic testing, consultations, and allied health input. The model showed strong initial stability, with fewer than 5% of patients changing streams during admission. Key data quality issues included inconsistent consultation documentation, underuse of structured fields, and retrospective overwriting of demographic information, affecting visibility of resource use. Despite these limitations, flow stream classification effectively differentiated patients by resource intensity and care complexity, offering a practical framework to support real-time hospital operations, complementing diagnosis-based groupings. Conclusion A structured classification model using routinely collected EMR data can differentiate inpatient resource needs. Flow stream stratification offers a complementary approach to traditional coding-based systems and may help identify operational bottlenecks. With improved documentation and system integration, this approach could enhance hospital responsiveness, resource planning, and overall system performance.