Abstract P195: Incorporating Physiological Signals to Blood Loss Prediction Based on Discrete Wavelet Transformation
Introduction: While analysis of ECG features may be useful for detecting acute changes in circulating blood volume it may be insufficient for accurate early detection of the hemorrhage, in particular when the detection is made based on the most recent 1–2 minutes of the recorded ECG. Thus, incorporating other physiological signals such as thoracic impedance (an indicator of blood volume) may improve the accuracy and reliability of the early detection of hemorrhage. Our objective in this study is to compare the use of short readings of multiple physiological signals when estimating the severity of hypovolemia instead of using only a single ECG signal.
Methods: Discrete wavelet transform (DWT) is applied to analyze the combined set of continuous ECG, arterial blood pressure (ABP) and thoracic impedance (TI) signals recorded from human subjects undergoing lower body negative pressure (LBNP) as a hemorrhage mimetic. The LBNP protocol consisted of a 5-min rest period (0 mm Hg) followed by 5 min of chamber decompression of the lower body to −15, −30, −45, and −60 mm Hg and additional increments of −10 mm Hg every 5 min until cardiovascular collapse. Because there is a wide range of LBNP levels resulting in collapse, volume loss was divided into 2 classes: severe (collapse stage together with preceding LBNP stage), and nonsevere as the remaining LBNP stages. This challenges the algorithm to be more individually specific. Two min of each 5 min stage of LBNP recorded signals (500 Hz) were processed. DWT features verified as significant features using ANOVA (p value <0.05) were examined. Each reading is then classified using SVM with leave-one-out cross-validation.
Results: A 276 sample set is used for classification. Classification accuracy with ECG only was 90% improving to 91% when both ECG and ABP were used. When ECG and TI were used accuracy increased to 92.3%. Use of ECG, ABP, and TI reduced accuracy to 91.3%. The results indicated that ABP adds no predictive accuracy. The best combination considering both complexity of the system and accuracy was the use of ECG and TI. The specificity and sensitivity of this combination were 91.1% and 92.2%, respectively.
Conclusion: ECG and TI are low level easily obtained signals that may assist in the detection and classification of hemorrhage.