Background Hand and upper limb trauma constitutes a major proportion of emergency presentations and is frequently associated with significant functional disability, socioeconomic burden, and prolonged rehabilitation. Over the years, several injury scoring systems have been developed, including the Strickland Digital Scoring System, Hand Injury Severity Score (HISS), Modified Hand Injury Severity Score (MHISS), Duncan’s classification, Tulipan’s classification, Pichitchai’s infection risk score, and the Mangled Upper Extremity Score (MUES). While these systems provide structured frameworks for assessing injury severity, they evaluate injuries in isolation and are limited by inter-observer variability and lack of holistic assessment. With advances in electronic health records (EHR) and artificial intelligence (AI), there is an opportunity to integrate multiple scoring systems into a unified, data-driven platform capable of providing comprehensive prognostication and guiding clinical decision-making. Aims and Objectives The aim of this study is to integrate existing hand injury scoring systems into an AI-based algorithm within an EHR platform, in order to improve prognostication and outcome prediction in hand and upper limb trauma. The objectives are: 1. To develop an electronic health record for Acute hand and upper limb injury 2. To identify dominant parameters that influence the score from multiple scoring system using feature engineering from machine learning 3. To develop a comprehensive score using AI based algorithms from the identified dominant parameters Methodology This study follows a prospective design with a five-year retrospective component. Retrospective data from 1112 patients with acute hand and upper limb trauma and prospective data from approximately 120 patients presenting to Kasturba Hospital, Manipal, are included. Data are collected using the standardized acute hand injury sheet and digitized into an institutional EHR with secure cloud storage. Patient identifiers are anonymized before integration with the computational platform. Existing scoring systems are embedded within the EHR, and AI models are developed using Google Colab with TensorFlow and Keras. Regression analysis and iterative training are used to identify dominant prognostic variables. The retrospective dataset is used for training and testing, while the prospective cohort serves as validation. Statistical analysis is performed using SPSS version 23.0, with Chi-square tests for categorical variables and t-tests or ANOVA for continuous variables, with significance set at p < 0.05. Discussion/ROL The literature highlights the progressive development of hand injury scoring methods but also their limitations. Strickland’s early scoring system provided thresholds for digital salvage but did not account for microsurgical advances. Campbell and Kay’s HISS correlated with return-to-work outcomes but excluded vascular and forearm injuries. Urso and colleagues refined this with MHISS, which predicted return-to-work more effectively but still lacked full prognostic capacity. Tulipan and Atthakomol introduced classification systems for open fractures that addressed infection risk, while Savetsky’s MUES attempted to quantify outcomes in mangled extremities. Despite these contributions, no single system comprehensively captures all injury domains. Artificial intelligence offers a solution by integrating diverse parameters into a reproducible predictive framework. Preliminary analyses indicate that an AI-based comprehensive score improves predictive accuracy over traditional systems, reduces subjectivity, and provides faster decision support in emergency settings. Beyond clinical applications, the platform standardizes documentation, facilitates data sharing, and offers potential for multicentric research and commercialization. |