606 Artificial Intelligence-powered Mobile Tool for Burn Injury Evaluation for First Responders
Bibliographic record
Abstract
Abstract Introduction Accurate assessment and early interventions in the field are crucial to the prognosis of burn injuries. Studies have indicated that up to 35% of burn patients are inappropriately transferred to hospitals. In some pediatric burns, reports have suggested the total burn surface area (TBSA) has been overestimated as much as 44%. First responders play a pivotal role in the assessment and management of burn injuries, especially in remote areas. An intuitive mobile application that incorporates the standardized practices of Advanced Burn Life Support (ABLS) with integrated artificial intelligence (AI) to assist in burn size, depth, and management is needed. Methods The mobile application was designed with an experienced team of burn specialists, physicians, and software engineers to identify the gaps in first responder burn care and to standardize methods for initial burn assessment. Previously assessed burn photos were characterized based on their depth into split partial thickness, deep partial thickness, and full thickness burns. Laser doppler imaging taken at the time of clinical assessment confirmed the burn depth. These images were used to build a convolutional neural network from to predict burn depth and boundaries. Results An AI-integrated mobile application was developed encompassing a primary survey with management solutions with the fundamentals based on ABLS. The application ensures the collection of pertinent information for burns, such as the patient's weight for calculating fluid resuscitation and provides recommendations based on burn depth and surface area. Photos taken using the mobile device can be analyzed by the AI in real time to aid in burn assessment. The accuracy of the prototype AI model can in distinguish severity assessment with an F1 accuracy score of 78% with a receiver operating characteristic of 85%. Furthermore, the model has a 92% accuracy for determining the boundaries of the burn. Conclusions A smartphone burn application provides an integrated way to improve the efficiency and accuracy of first responders to assess, manage, and triage burns before reaching the hospital. Through the application, management recommendations are tailored to the extent of the injury and provides a detailed report to secondary healthcare providers. Applicability of Research to Practice A state-of-the-art burn application can be a tool that can be used by first responders to capture essential data while ensuring the accuracy of the initial assessment. The integration of artificial intelligence will help personalize and streamline the management pathway improving communication between field and hospital care providers and improving patient care.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.021 | 0.006 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".