A novel skin temperature estimation system for predicting pressure injury occurrence based on continuous body sensor...

Area of Science:

• Biomedical Engineering
• Medical Device Technology
• Clinical Informatics

Background:

• Pressure injuries pose a significant risk to immobile older patients.
• Accurate and efficient prediction of pressure injury development is essential for prevention.
• Current risk assessment tools may be invasive or restrictive.

Purpose of the Study:

• To develop a non-invasive, non-constrained method for estimating skin temperature changes.
• To utilize temperature sensors placed under bedsheets for objective risk assessment.
• To detect early signs of ischemia and inflammation indicative of pressure injury risk.

Main Methods:

• Conducted a thermal skin simulation study using a body surface and bed environment model.
• Performed a descriptive correlational study with healthy participants (N=17).
• Attached temperature sensors to the sacral skin of participants lying supine on a mattress.

Main Results:

• Machine learning models accurately estimated skin temperature changes under sheets using time-shifted data (R² up to 0.9967).
• An extra-tree regressor model estimated absolute skin temperature in healthy individuals with high accuracy (R² = 0.8145).
• The method demonstrated feasibility for estimating temperature variations indicative of pressure injury risk.

Conclusions:

• Combining interface pressure and temperature sensors enables accurate skin temperature change estimation.
• This research contributes to developing a novel skin temperature measurement method for clinical settings.
• The proposed tool offers objective, non-invasive, and continuous monitoring for pressure injury prevention.