Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference
-
Conf Proc IEEE Eng Med Biol Soc · Jan 2014
Prediction of mortality from respiratory distress among long-term mechanically ventilated patients.
With the advent of inexpensive storage, pervasive networking, and wireless devices, it is now possible to store a large proportion of the medical data that is collected in the intensive care unit (ICU). These data sets can be used as valuable resources for developing and validating predictive analytics. In this report, we focus on the problem of prediction of mortality from respiratory distress among long-term mechanically ventilated patients using data from the publicly-available MIMIC-II database. ⋯ We also find that variables related to respiration rate have more predictive accuracy than variables related to oxygenation status. Ultimately, we have developed a model which predicts mortality from respiratory distress in the ICU with a cross-validated area-under-the-curve (AUC) of approximately 0.74. Four methodologies are utilized for model dimensionality-reduction: univariate logistic regression, multivariate logistic regression, decision trees, and penalized logistic regression.
-
Conf Proc IEEE Eng Med Biol Soc · Jan 2014
Adaptive thresholding with inverted triangular area for real-time detection of the heart rate from photoplethysmogram traces on a smartphone.
Photoplethysmogram (PPG) signals acquired by smartphone cameras are weaker than those acquired by dedicated pulse oximeters. Furthermore, the signals have lower sampling rates, have notches in the waveform and are more severely affected by baseline drift, leading to specific morphological characteristics. This paper introduces a new feature, the inverted triangular area, to address these specific characteristics. ⋯ We collected data from 24 volunteers and compared our algorithm in peak detection with two competing algorithms designed for PPG signals, Incremental-Merge Segmentation (IMS) and Adaptive Thresholding (ADT). A sensitivity of 98.0% and a positive predictive value of 98.8% were obtained, which were 7.7% higher than the IMS algorithm in sensitivity, and 8.3% higher than the ADT algorithm in positive predictive value. The experimental results confirmed the applicability of the proposed method.
-
Conf Proc IEEE Eng Med Biol Soc · Jan 2014
Respiratory rate estimation from the oscillometric waveform obtained from a non-invasive cuff-based blood pressure device.
The presence of respiratory activity in the electrocardiogram (ECG), the pulse oximeter's photoplethysmo-graphic and continuous arterial blood pressure signals is a well-documented phenomenon. In this paper, we demonstrate that such information is also present in the oscillometric signal acquired from automatic non-invasive blood pressure monitors, and may be used to estimate the vital sign respiratory rate (RR). ⋯ Results demonstrated a good RR estimation accuracy of our method when compared to the reference values extracted from the reference respiration waveforms (mean absolute error of 2.69 breaths/min), which is comparable to existing methods in the literature that extract RR from other physiological signals. The proposed method has been implemented in Java on the Android device for use in an mHealth platform.
-
Conf Proc IEEE Eng Med Biol Soc · Jan 2014
Comparative StudyComparison of methods for determining pulse arrival time from Doppler and photoplethysmography signals.
The aim of this study was to compare three foot-finding methods applied to ultrasound Doppler and photoplethysmographic (PPG) signals: maximum 1st derivative, maximum 2nd derivative and an 'intersecting tangents' method. The pulse arrival times of each method were compared. ⋯ The results show that the maximum 1st derivative method produced significantly larger pulse arrival times than the other two methods. The intersecting tangents method produced greatest precision for cardiac periods compared with ECG than maximum 1st or 2nd derivatives for both Doppler (r(2) = 0.975) and PPG (r(2) = 0.987) signals.
-
Conf Proc IEEE Eng Med Biol Soc · Jan 2014
Investigation of photoplethysmography and near infrared spectroscopy for the assessment of tissue blood perfusion.
Pulse Oximetry (PO) and Near Infrared Spectroscopy (NIRS) are among the most widely adopted optical techniques for the assessment of tissue perfusion. PO estimates arterial oxygen saturation (SpO2) by exploiting light attenuations due to pulsatile arterial blood (AC) and constant absorbers (DC) at two different wavelengths. NIRS processes the attenuations of at least two wavelengths to calculate concentrations of Deoxygenated ([HHb]), Oxygenated ([HbO2]), Total Haemoglobin ([tHb]) and Tissue Oxygenation Index (TOI). ⋯ The system adopts both Pulse Oximetry and NIRS principles to calculate SpO2, [HHb], and [HbO2] and [tHb]. The system has been evaluated on the forearm of 10 healthy volunteers during cuff-induced vascular occlusions. The presented system was able to estimate SpO2, [HHb], [HbO2] and [tHb], showing good agreement with state-of-the-art NIRS and conventional PO.