A Low Cost Multi Parameter Monitoring Framework for Pressure Ulcer Prevention in Resource Limited Healthcare

Abstract

Pressure ulcers (PUs) continue to represent a significant clinical and economic burden on healthcare systems worldwide. They are associated with prolonged hospitalization, increased treatment costs, reduced quality of life, and a higher risk of morbidity among patients with limited mobility. The challenge is particularly pronounced in resource-constrained healthcare environments, where shortages of trained personnel, limited access to specialized equipment, and inadequate monitoring capabilities often impede timely prevention and intervention. Although numerous technological solutions have been proposed for pressure ulcer prevention, many existing systems rely primarily on posture detection or expensive sensing platforms, limiting their applicability in low-resource settings. This study investigates the feasibility of a low-cost, multi-parameter monitoring framework that can provide continuous assessment of pressure ulcer risk through real-time physiological and environmental sensing. To establish the research foundation, a systematic review of intelligent sensor-based pressure ulcer prevention technologies was conducted to identify prevailing trends, methodological limitations, and unmet clinical needs. Insights from the review informed the design of a proof-of-concept monitoring framework utilizing affordable and widely available hardware components, including an ESP32 microcontroller, force-sensing pressure sensors, and temperature-humidity sensors. The proposed framework captures pressure distribution, skin microclimate conditions, and moisture-related indicators, integrating these variables into a Multi-Parameter Risk Index for continuous bedside monitoring. To evaluate different levels of analytical sophistication, three risk-detection approaches were implemented: a conventional pressure-threshold model, a multi-parameter sensor fusion algorithm, and a machine-learning-based anomaly detection model employing the Isolation Forest algorithm. System performance was assessed through simulation and virtual embedded-system testing, while maintaining an estimated deployment cost of approximately USD 20–45 per bed. The findings demonstrate that incorporating multiple physiological indicators substantially improves pressure ulcer risk detection compared with pressure monitoring alone. The baseline pressure-only model achieved an estimated sensitivity of 76%, whereas the multi-parameter fusion approach increased sensitivity to 84%. The machine-learning-based anomaly detection model produced the strongest performance, achieving an estimated sensitivity of 89%, corresponding to a 13% improvement over the baseline approach. In addition to enhanced detection capability, the framework maintained real-time operational performance with inference latency below 150 ms and functioned entirely offline, eliminating dependence on cloud-based infrastructure and supporting deployment in settings with limited connectivity. Overall, the study provides evidence that affordable multimodal sensing combined with lightweight artificial intelligence techniques can support more effective and accessible pressure ulcer prevention. By addressing both technological and economic barriers, the proposed framework contributes toward the development of practical monitoring solutions for underserved healthcare environments. Future research will focus on clinical validation with real patient populations, refinement of patient-specific risk modeling, incorporation of additional biomechanical factors such as shear force, and the development of predictive algorithms trained on longitudinal clinical datasets to improve early risk identification and preventive decision-making

Keywords

Other

Citation

References

Reference metadata is pending and must be finalized before DOI deposit.