Jacqueline Huvanandana

Prediction of intraventricular haemorrhage in preterm infants using time series analysis of blood pressure and respiratory signals

Despite the decline in mortality rates of extremely preterm infants, intraventricular haemorrhage (IVH) remains common in survivors. The need for resuscitation and cardiorespiratory management, particularly within the first 24 hours of life, are important factors in the incidence and timing of IVH. Variability analyses of heart rate and blood pressure data has demonstrated potential approaches to predictive monitoring. In this study, we investigated the early identification of infants at a high risk of developing IVH, using time series analysis of blood pressure and respiratory data. We also explore approaches to improving model performance, such as the inclusion of multiple variables and signal pre-processing to enhance the results from detrended fluctuation analysis. Of the models we evaluated, the highest area under receiver-operator characteristic curve (5th, 95th percentile) achieved was 0.921 (0.82, 1.00) by mean diastolic blood pressure and the long-term scaling exponent of pulse interval (PI α2), exhibiting a sensitivity of >90% at a specificity of 75%. Following evaluation in a larger population, our approach may be useful in predictive monitoring to identify infants at high risk of developing IVH, offering caregivers more time to adjust intensive care treatment.

Reducing false arrhythmia alarms in the ICU using multimodal signals and robust QRS detection

This study developed algorithms to decrease the arrhythmia false alarms in the ICU by processing multimodal signals of photoplethysmography (PPG), arterial blood pressure (ABP), and two ECG signals. The goal was to detect the five critical arrhythmias comprising asystole (ASY), extreme bradycardia (EBR), extreme tachycardia (ETC), ventricular tachycardia (VTA), and ventricular flutter or fibrillation (VFB). The different characteristics of the arrhythmias suggested the application of individual signal processing for each alarm and the combination of the algorithms to enhance false alarm detection. Thus, different features and signal processing techniques were used for each arrhythmia type. The ECG signals were first processed to reduce the signal interference. Then, a Hilbert-transform based QRS detector algorithm was utilized to identify the QRS complexes, which were then processed to determine the instantaneous heart rate. The pulsatile signals (PPG and ABP) were processed to discover the pulse onset of beats which were then employed to measure the heart rate. The signal quality index (SQI) of the signals was implemented to verify the integrity of the heart rate information. The overall score obtained by our algorithms in the 2015 Computing in Cardiology Challenge was a score of 74.03% for retrospective and 69.92% for real-time analysis.

Reducing false arrhythmia alarms in the ICU by Hilbert QRS detection

In this study, we develop algorithms that reduce the arrhythmia false alarms in the ICU by processing the four signals of Photoplethysmography (PPG), arterial blood pressure (ABP), ECG Lead II, and Augmented right arm ECG. Our algorithms detect five arrhythmias including asystole, extreme bradycardia, extreme tachycardia, ventricular tachycardia (VT), and ventricular flutter or fibrillation (VF). Real time algorithm is provided. Our processing proceeded as follows. Firstly, preprocessing was applied to the ECG signals by two median filters in order to remove the baseline wander and high-frequency noise. Then a Hilbert-transform based QRS detector algorithm was used to detect R waves from the ECG signals. Following this, RR intervals were calculated from the available ECG signals. Pulse onset points of the pulsatile signals (PPG and ABP) were also detected and the signal quality index (SQI) of the four signals was measured. The ECG based RR intervals were combined with the pulsatile signal based RR intervals using the algorithms provided by the CinC2015 competition organizers. The combined RR intervals were thresholded at the clinically important values for the five arrhythmias. Template matching was used to detect ventricular tachycardia (VT) and power spectrum of ECG signals and identifying the VF frequency components employed to investigate ventricular fibrillation. Our highest overall result was a 98% True Positive Rate (TPR), 66% True Negative Rate (TNR) with a score of 74.03% for the retrospective algorithm. For the realtime algorithm, we achieved a 98% TPR, 65% TNR and a score of 69.92%.

How do different brands of size 1 laryngeal mask airway compare with face mask ventilation in a dedicated laryngeal mask airway teaching manikin

Background International neonatal resuscitation guidelines recommend the use of laryngeal mask airway (LMA) with newborn infants (≥34 weeks’ gestation or >2 kg weight) when bag-mask ventilation (BMV) or tracheal intubation is unsuccessful. Previous publications do not allow broad LMA device comparison. Objective To compare delivered ventilation of seven brands of size 1 LMA devices with two brands of face mask using self-inflating bag (SIB). Design 40 experienced neonatal staff provided inflation cycles using SIB with positive end expiratory pressure (PEEP) (5 cmH2O) to a specialised newborn/infant training manikin randomised for each LMA and face mask. All subjects received prior education in LMA insertion and BMV. Results 12 415 recorded inflations for LMAs and face masks were analysed. Leak detected was lowest with i-gel brand, with a mean of 5.7% compared with face mask (triangular 42.7, round 35.7) and other LMAs (45.5–65.4) (p<0.001). Peak inspiratory pressure was higher with i-gel, with a mean of 28.9 cmH2O compared with face mask (triangular 22.8, round 25.8) and other LMAs (14.3–22.0) (p<0.001). PEEP was higher with i-gel, with a mean of 5.1 cmH2O compared with face mask (triangular 3.0, round 3.6) and other LMAs (0.6–2.6) (p<0.001). In contrast to other LMAs examined, i-gel had no insertion failures and all users found i-gel easy to use. Conclusion This study has shown dramatic performance differences in delivered ventilation, mask leak and ease of use among seven different brands of LMA tested in a manikin model. This coupled with no partial or complete insertion failures and ease of use suggests i-gel LMA may have an expanded role with newborn resuscitation as a primary resuscitation device.

Improvement of near infrared body fat sensing at 45-degree source-detector position angle

While other studies into sensing the thickness of subcutaneous fat have configured the axes of both the source and detector to be at a 90-degree angle to the surface being measured, we introduce a source-detector set up at 45-degree angle. We apply Gamos computer simulation and ex-vivo phantom experiment to compare the near infrared response between 90-degree and 45-degree position angles at 930 nm. A single fat layer is used as a sample with thickness at 1 mm to 15 mm in 1 mm intervals in the simulations while the thickness in the experiments is from 3 mm to 15 mm in 3 mm intervals utilising ground beef fat. Near infrared intensities exhibit logarithmic response with the maximum thickness detection at 9 mm. The 45-degree source-detector configuration produces a higher detected intensity throughout the range of thickness than that of 90-degree in both simulations and experiments. The sensitivities of the source-detector at 45-degree and at 90-degree are 15.41% and 5.6% respectively in the simulations. Meanwhile in the experiments, the sensitivity of the 45-degree is 2.364% while 1.712% of the 90-degree. With the higher performance of the source-detector at 45-degree position angle, it suggests to be implemented in real near infrared body fat measurement device.

Length-free near infrared measurement of newborn malnutrition.

Under-nutrition in neonates can cause immediate mortality, impaired cognitive development and early onset adult disease. Body fat percentage measured using air-displacement-plethysmography has been found to better indicate under-nutrition than conventional birth weight percentiles. However, air-displacement-plethysmography equipment is expensive and non-portable, so is not suited for use in developing communities where the burden is often the greatest. We proposed a new body fat measurement technique using a length-free model with near-infrared spectroscopy measurements on a single site of the body - the thigh. To remove the need for length measurement, we developed a model with five discrete wavelengths and a sex parameter. The model was developed using air-displacement-plethysmography measurements in 52 neonates within 48 hours of birth. We identified instrumentation required in a low-cost LED-based screening device and incorporated a receptor device that can increase the amount of light collected. This near-infrared method may be suitable as a low cost screening tool for detecting body fat levels and monitoring nutritional interventions for malnutrition in neonates and young children in resource-constrained communities.

An anthropometric approach to characterising neonatal morbidity and body composition, using air displacement plethysmography as a criterion method

Background With the greatest burden of infant undernutrition and morbidity in low and middle income countries (LMICs), there is a need for suitable approaches to monitor infants in a simple, low-cost and effective manner. Anthropometry continues to play a major role in characterising growth and nutritional status. Methods We developed a range of models to aid in identifying neonates at risk of malnutrition. We first adopted a logistic regression approach to screen for a composite neonatal morbidity, low and high body fat (BF%) infants. We then developed linear regression models for the estimation of neonatal fat mass as an assessment of body composition and nutritional status. Results We fitted logistic regression models combining up to four anthropometric variables to predict composite morbidity and low and high BF% neonates. The greatest area under receiver-operator characteristic curves (AUC with 95% confidence intervals (CI)) for identifying composite morbidity was 0.740 (0.63, 0.85), resulting from the combination of birthweight, length, chest and mid-thigh circumferences. The AUCs (95% CI) for identifying low and high BF% were 0.827 (0.78, 0.88) and 0.834 (0.79, 0.88), respectively. For identifying composite morbidity, BF% as measured via air displacement plethysmography showed strong predictive ability (AUC 0.786 (0.70, 0.88)), while birthweight percentiles had a lower AUC (0.695 (0.57, 0.82)). Birthweight percentiles could also identify low and high BF% neonates with AUCs of 0.792 (0.74, 0.85) and 0.834 (0.79, 0.88). We applied a sex-specific approach to anthropometric estimation of neonatal fat mass, demonstrating the influence of the testing sample size on the final model performance. Conclusions These models display potential for further development and evaluation in LMICs to detect infants in need of further nutritional management, especially where traditional methods of risk management such as birthweight for gestational age percentiles may be variable or non-existent, or unable to detect appropriately grown, low fat newborns.

Cardiovascular impact of intravenous caffeine in preterm infants

To evaluate the acute effect of intravenous caffeine on heart rate and blood pressure variability in preterm infants.

Advanced analyses of physiological signals in the neonatal intensive care unit

Management and monitoring of infants within the neonatal intensive care unit represents a unique challenge. It involves an array of life-threatening diseases, procedures with potentially lifelong impacts, co-morbidities associated with preterm birth and risk of infection from prolonged exposure to the hospital environment. With the integration of monitoring systems and increasing accessibility of high-resolution data, there is a growing interest in the utility of advanced data analyses in predictive monitoring and characterising patterns of disease. Such analyses may offer an opportunity to identify infants at high risk of certain conditions and to detect the onset of disease prior to manifestation of clinical signs. This allows caregivers more time to respond and mitigate any abnormal or potentially fatal changes. We review techniques for variability analysis as they have been or have the potential to be applied to neonatal intensive care, the disease conditions in which they have been tested, and technical as well as clinical challenges relevant to their application.

Logistic regression models for predicting intraventricular haemorrhage in preterm infants using respiratory and blood pressure signals

Despite the decline in mortality rates for extremely preterm infants, intraventricular haemorrhage (IVH) remains a threat to their survival. In this study, we sought to explore logistic regression models for predicting IVH as they would be applied in a clinical setting, using features derived from respiratory and blood pressure signals. Calculated predictors included mean (μ) and the short- and long-term scaling exponents (α1, α2) from detrended fluctuation analysis. The model fitted with short-term scaling exponent (α1) of the beat-to-beat diastolic blood pressure (DBP) exhibited an area under receiver-operator characteristic curve (AUC) of 0.788 (0.62, 0.96), with a sensitivity of approximately 0.875 at a specificity of 0.75. Of the multivariable models explored, the highest AUC was 0.831 (0.66, 1.00), combining μDBP with α1 of the beat-to-beat systolic blood pressure (SBP).