Digital Transformation of EMT Daily Reporting: A Preliminary Study on Image Analysis for Automated Data Collection and Report Generation from Paper Medical Records

Introduction: Emergency Medical Teams (EMTs) must compile daily reports of 91 Minimum Data Set (MDS) items for each operational site during disaster responses. This requirement demands extensive data aggregation into the standardized EMT MDS data format, creating a significant administrative burden in resource-constrained settings. While paper-based medical records remain essential in austere environments, manual compilation consumes critical time and resources. This preliminary research explores the feasibility of an automated approach using image analysis technology. Methods: The exploratory development investigated three technical components: 1) Structured paper form design optimized for MDS parameter extraction through strategic placement of checkboxes and data fields; 2) Experimental implementation of Python-based image processing system with OCR capability and CNN-based checkbox recognition; 3) Data conversion pipeline generating CSV files compliant with EMT MDS eDATA format. Initial testing utilized simulated medical records to assess technical feasibility. Results: Preliminary testing demonstrated the technical viability of automated data extraction and EMT MDS eDATA format generation. A convolutional neural network (CNN) model, trained with 1,504 data samples for checkbox recognition, achieved high accuracy metrics (accuracy, precision, recall, and F1-score all 0.98) in validation testing. The automated system reduced processing time from 20 to 10 seconds per medical record compared to manual data entry. The prototype system successfully generated standardized CSV outputs while reducing operator fatigue through automated data compilation. Conclusion: This preliminary investigation demonstrates the successful bridging of paper-based medical documentation and digital reporting requirements in disaster response settings. The achieved 98% accuracy in data extraction and 50% reduction in processing time establishes a proof-of-concept for automated EMT daily reporting. Further development will focus on algorithm optimization for enhanced accuracy and speed, mobile application development for improved field usability, and systematic validation through multi-user testing under simulated operational conditions. This solution provides an efficient pathway for both mobility-focused EMTs and electronic health record backup strategies.