Abstract
Illness severity is a major determinant of neonatal mortality. In neonatal intensive care, severity-of-illness scales, primarily relied on subjective judgment by experienced physicians, are used to assess the mortality risk prediction and therapeutic intensity. Because cardiovascular morbidity and other disease state may contribute to hemodynamic abnormalities, the illness severity may reflect on the state of hemodynamic abnormalities. Several diagnostic approaches, such as laser Doppler flowmetry and orthogonal polarization spectral, have been developed to evaluate the local information of blood flow or vascular distribution on skin surface, but they may not efficiently evaluate the systemic circulation. In this study, an intelligent systemic circulation monitoring system was proposed to non-invasively evaluate the blood perfusion state and further estimate the illness severity. In the proposed system, a wireless sensing device was designed to real-time monitor the changes of hemoglobin parameters under the altered external pressure, and several indexes were defined from the hemodynamic response to present the blood perfusion state. The relationship between the blood perfusion state and the illness severity was also investigated in this study. Moreover, the technique of neural network was also applied in the classification of illness severity. The experimental results indicate the differences between many defined blood perfusion indexes in neonates with different illness severity were exactly significant and the mild and severe illness severity group could also be efficiently recognized by the neural network technique. Therefore, the proposed system may contain the potential of future clinical applications in the evaluation of illness severity or other diseases.
Original language | English |
---|---|
Pages (from-to) | 127468-127478 |
Number of pages | 11 |
Journal | IEEE Access |
Volume | 11 |
DOIs | |
State | Published - 2023 |
Keywords
- blood perfusion
- hemodynamic abnormalities
- Illness severity
- neonatal intensive care
- neural network