IEEE journal of translational engineering in health and medicine · 2026
Toward Real-Time Detection of Intravenous Infiltration Using a Low-Cost Strain-Sensing Patch: Validation in Animal Models and Pediatric Clinical Monitoring.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
El-Khalili T, Morgenstern S, Fackler J, Yang R
Аннотация
OBJECTIVE: To develop and evaluate a low-cost, skin-safe strain-sensing patch and real-time-capable algorithm for early detection of intravenous (IV) infiltration during continuous monitoring.
METHODS: A custom strain sensor composed of discrete piezoresistive elements was encased between two transparent medical dressings and mounted adjacent to the IV site. Data were collected during controlled IV infiltrations in juvenile pigs and during inpatient monitoring in pediatric patients without infiltrations. A strictly causal, event-based detection algorithm was implemented using features derived from the strain signal and its first difference. Performance was evaluated using event sensitivity, time-to-detection (TTD), and false-positive alerts per hour (FP/hr). Candidate thresholds were derived from the animal training-set evidence distribution. The operating point was selected using training-data sensitivity, detection time, and control false-alert burden, then fixed before held-out testing.
RESULTS: Across 16 animal recordings comprising 35 infiltration events, the algorithm detected 30 events (event sensitivity = 0.86) with a median TTD of 57 seconds at the selected operating threshold. Approximately 86.7% of detected events occurred within two minutes of infiltration onset. On the held-out recording subset, 8 of 9 infiltration events were detected, with a median TTD of 53.5 seconds. Evaluation on 117.8 hours of pediatric inpatient control data demonstrated a threshold-dependent false-positive alert burden, operating at approximately 1.2 alerts per hour at the selected operating threshold.
CONCLUSION: A strain-sensing patch integrated with a real-time-capable, event-based detection algorithm enabled early IV infiltration detection at low infused volumes while maintaining a false-positive alert burden that is acceptable for early-stage clinical evaluation.
METHODS: A custom strain sensor composed of discrete piezoresistive elements was encased between two transparent medical dressings and mounted adjacent to the IV site. Data were collected during controlled IV infiltrations in juvenile pigs and during inpatient monitoring in pediatric patients without infiltrations. A strictly causal, event-based detection algorithm was implemented using features derived from the strain signal and its first difference. Performance was evaluated using event sensitivity, time-to-detection (TTD), and false-positive alerts per hour (FP/hr). Candidate thresholds were derived from the animal training-set evidence distribution. The operating point was selected using training-data sensitivity, detection time, and control false-alert burden, then fixed before held-out testing.
RESULTS: Across 16 animal recordings comprising 35 infiltration events, the algorithm detected 30 events (event sensitivity = 0.86) with a median TTD of 57 seconds at the selected operating threshold. Approximately 86.7% of detected events occurred within two minutes of infiltration onset. On the held-out recording subset, 8 of 9 infiltration events were detected, with a median TTD of 53.5 seconds. Evaluation on 117.8 hours of pediatric inpatient control data demonstrated a threshold-dependent false-positive alert burden, operating at approximately 1.2 alerts per hour at the selected operating threshold.
CONCLUSION: A strain-sensing patch integrated with a real-time-capable, event-based detection algorithm enabled early IV infiltration detection at low infused volumes while maintaining a false-positive alert burden that is acceptable for early-stage clinical evaluation.