Medicine (Austin & Northern Health) - Research Publications
Permanent URI for this collection
Search Results
1 - 10 of 48
-
ItemUncertainty-aware non-invasive patient–ventilator asynchrony detection using latent Gaussian mixture generative classifier with noisy label correction(Springer Science and Business Media LLC, 2024-01-01)Abstract Patient–ventilator asynchrony (PVA) refers to instances where a mechanical ventilator’s cycles are desynchronised from the patient’s breathing efforts, and may result in patient discomfort and potential ineffective ventilation. Typically, they are identified with constant monitoring by trained clinicians. Such expertise is often limited; therefore, it is desirable to automate PVA detection with machine learning methods. However, there are three major challenges to applying machine learning to the problem: data collected from non-invasive ventilation are often noisy, there exists high variability between patients or between setting changes, and manual annotations of PVA events are not always consistent. To produce meaningful inference from such noisy data, a model needs to not only provide a measure of uncertainty, but also take into account potential inconsistencies in the training signal it is based on. In this work, we propose a conditional latent Gaussian mixture generative classifier with noisy label correction, which is capable of capturing variations within and between classes, providing well-calibrated class probabilities, detecting unlikely input instances that deviates from training data, while also taking into account possible mislabelling of event classes. We show that our model is able to match the performance of a well-tuned gradient boosting classifier, but also produce better calibrated predictions and smaller performance variability between patients.
-
ItemIdentification of Patient Ventilator Asynchrony in Physiological Data Through Integrating Machine-Learning(SCITEPRESS - Science and Technology Publications, 2024)Patient Ventilator Asynchrony (PVA) occurs where a mechanical ventilator aiding a patient's breathing falls out of synchronisation with their breathing pattern. This de-synchronisation may cause patient distress and can lead to long-term negative clinical outcomes. Research into the causes and possible mitigations of PVA is currently conducted by clinical domain experts using manual methods, such as parsing entire sleep hypnograms visually, and identifying and tagging instances of PVA that they find. This process is very labour-intensive and can be error prone. This project aims to make this analysis more efficient, by using machine-learning approaches to automatically parse, classify, and suggest instances of PVA for ultimate confirmation by domain experts. The solution has been developed based on a retrospective dataset of intervention and control patients that were recruited to a non-invasive ventilation study. This achieves a specificity metric of over 90%. This paper describes the process of integrating the output of the machine learning into the bedside clinical monitoring system for production use in anticipation of a future clinical trial.
-
ItemA brief assessment of eye blink drowsiness immediately prior to or following driving detects drowsiness related driving impairment(Wiley, 2023-06)Drowsy driving is a major cause of fatal and serious injury motor vehicle accidents. The inability objectively to assess drowsiness has hindered the assessment of fitness to drive and the development of drowsy driving regulations. This study evaluated whether spontaneous eye blink parameters measured briefly pre- and post-drive could be used to detect drowsy driving impairment. Twelve healthy participants (6 female) drove an instrumented vehicle for 2 h on a closed-loop track during a rested (8-10 h awake) and an extended wake condition (32-34 h awake). Pre- and post-drive, the participants completed a 5 min eye blink task, a psychomotor vigilance task (PVT), and the Karolinska sleepiness scale (KSS). Whole drive impairment was defined as >3.5 lane departures per hour. Severe end of drive impairment was defined as ≥2 lane departures in the last 15 min. The pre-drive % of time with eyes closed best predicted the whole drive impairment (area under the curve [AUC] 0.87). KSS had similar prediction ability (AUC 0.85), while PVT reaction time (AUC 0.72) was less accurate. The composite eye blink parameter, the Johns drowsiness scale was the best retrospective detector of severe end of drive impairment (AUC 0.99). The PVT reaction time (AUC 0.92) and the KSS (AUC 0.93) were less accurate. Eye blink parameters detected drowsy driving impairment with an accuracy that was similar to, or marginally better than, PVT and KSS. As eye blink measures are simple to measure, are objective and have high accuracy, they present an ideal option for the assessment of fitness for duty and roadside drowsiness.
-
ItemProlonged Eyelid Closure Episodes during Sleep Deprivation in Professional Drivers(AMER ACAD SLEEP MEDICINE, 2016)STUDY OBJECTIVES: Real life ocular measures of drowsiness use average blink duration, amplitude and velocity of eyelid movements to reflect drowsiness in drivers. However, averaged data may conceal the variability in duration of eyelid closure episodes, and more prolonged episodes that indicate higher levels of drowsiness. The current study aimed to describe the frequency and duration of prolonged eyelid closure episodes during acute sleep deprivation. METHODS: Twenty male professional drivers (mean age ± standard deviation = 41.9 ± 8.3 years) were recruited from the Transport Workers Union newsletter and newspaper advertisements in Melbourne, Australia. Each participant underwent 24 hours of sleep deprivation and completed a simulated driving task (AusEd), the Psychomotor Vigilance Task, and the Karolinska Sleepiness Scale. Eyelid closure episodes during the driving task were recorded and analyzed manually from digital video recordings. RESULTS: Eyelid closure episodes increased in frequency and duration with a median of zero s/h of eyelid closure after 3 h increasing to 34 s/h after 23 h awake. Eyelid closure episodes were short and infrequent from 3 to 14 h of wakefulness. After 17 h of sleep deprivation, longer and more frequent eyelid closure episodes began to occur. Episodes lasting from 7 seconds up to 18 seconds developed after 20 h of wakefulness. Length of eyelid closure episodes was moderately to highly correlated with the standard deviation of lateral lane position, braking reaction time, crashes, impaired vigilance, and subjective sleepiness. CONCLUSIONS: The frequency and duration of episodes of prolonged eyelid closure increases during acute sleep deprivation, with very prolonged episodes after 17 hours awake. Automated devices that assess drowsiness using averaged measures of eyelid closure episodes need to be able to detect prolonged eyelid closure episodes that occur during more severe sleep deprivation.
-
ItemSilver nitrate therapy for persistent tracheocutaneous fistula following prolonged tracheostomy and invasive ventilation: A case report(WILEY, 2022-11)We report the case of a man with severe Guillain-Barré syndrome who developed a persistent tracheocutaneous fistula (TCF) following prolonged tracheostomy and mechanical ventilation. Following tracheostomy decannulation, the TCF had a deleterious effect on non-invasive positive pressure ventilation efficacy and ability to effectively clear airway secretions due to air leaking from the patent stoma. This case highlights a non-surgical approach to TCF management that is not well-described in the literature and presents an alternative management option for cohorts of patients in which the risk associated with surgical interventions may be undesirable.
-
ItemRapidly and slowly progressive neuromuscular disease: differences in pulmonary function, respiratory tract infections and response to lung volume recruitment therapy (LVR)(BMJ PUBLISHING GROUP, 2022-12)INTRODUCTION: Reduced lung volumes are a hallmark of respiratory muscle weakness in neuromuscular disease (NMD). Low respiratory system compliance (Crs) may contribute to restriction and be amenable to lung volume recruitment (LVR) therapy. This study evaluated respiratory function and the immediate impact of LVR in rapidly progressive compared to slowly progressive NMD. METHODS: We compared vital capacity (VC), static lung volumes, maximal inspiratory and expiratory pressures (MIP, MEP), Crs and peak cough flow (PCF) in 80 adult participants with motor neuron disease ('MND'=27) and more slowly progressive NMDs ('other NMD'=53), pre and post a single session of LVR. Relationships between respiratory markers and a history of respiratory tract infections (RTI) were examined. RESULTS: Participants with other NMD had lower lung volumes and Crs but similar reduction in respiratory muscle strength compared with participants with MND (VC=1.30±0.77 vs 2.12±0.75 L, p<0.001; Crs=0.0331±0.0245 vs 0.0473±0.0241 L/cmH2O, p=0.024; MIP=39.8±21.3 vs 37.8±19.5 cmH2O). More participants with other NMD reported an RTI in the previous year (53% vs 22%, p=0.01). The likelihood of having a prior RTI was associated with baseline VC (%predicted) (OR=1.03 (95% CI 1.00 to 1.06), p=0.029). Published thresholds (VC<1.1 L or PCF<270 L/min) were, however, not associated with prior RTI.A single session of LVR improved Crs (mean (95% CI) increase = 0.0038 (0.0001 to 0.0075) L/cmH2O, p=0.047) but not VC. CONCLUSION: These findings corroborate the hypothesis that ventilatory restriction in NMD is related to weakness initially with respiratory system stiffness potentiating lung volume loss in slowly progressive disease. A single session of LVR can improve Crs. A randomised controlled trial of regular LVR is needed to assess longer-term effects.
-
ItemNo Preview AvailablePost-stroke fatigue is associated with resting state posterior hypoactivity and prefrontal hyperactivity(SAGE PUBLICATIONS LTD, 2022-10-01)BACKGROUND: Fatigue is associated with poor functional outcomes and increased mortality following stroke. Survivors identify fatigue as one of their key unmet needs. Despite the growing body of research into post-stroke fatigue, the specific neural mechanisms remain largely unknown. AIM: This observational study aimed to identify resting state brain activity markers of post-stroke fatigue. METHOD: Sixty-three stroke survivors (22 women; age 30-89 years; mean 67.5 ± 13.4 years) from the Cognition And Neocortical Volume After Stroke study, a cohort study examining cognition, mood, and brain volume in stroke survivors following ischemic stroke, underwent brain imaging three months post-stroke, including a 7-minute resting state functional magnetic resonance imaging. We calculated the fractional amplitude of low-frequency fluctuations, which is measured at the whole-brain level and can detect altered spontaneous neural activity of specific regions. RESULTS: Forty-five participants reported experiencing post-stroke fatigue as measured by an item on the Patient Health Questionnaire-9. Fatigued compared to non-fatigued participants demonstrated significantly lower resting-state activity in the calcarine cortex (p < 0.001, cluster-corrected pFDR = 0.009, k = 63) and lingual gyrus (p < 0.001, cluster-corrected pFDR = 0.025, k = 42) and significantly higher activity in the medial prefrontal cortex (p < 0.001, cluster-corrected pFDR = 0.03, k = 45). CONCLUSIONS: Post-stroke fatigue is associated with posterior hypoactivity and prefrontal hyperactivity reflecting dysfunction within large-scale brain systems such as fronto-striatal-thalamic and frontal-occipital networks. These systems in turn might reflect a relationship between post-stroke fatigue and abnormalities in executive and visual functioning. This whole-brain resting-state study provides new targets for further investigation of post-stroke fatigue beyond the lesion approach.
-
ItemMaternal sleep behaviours preceding fetal heart rate events on cardiotocography(WILEY, 2022-04)In Australia, a significant proportion of stillbirths remain unexplained. Recent research has highlighted nocturnal maternal behaviours as potentially modifiable contributors. This study determined whether sleep-related behaviours including sleep position and sleep-disordered breathing adversely affect fetuses overnight, in both uncomplicated pregnancies and those at increased risk due to hypertensive disorders or fetal growth restriction (FGR). All participants underwent polysomnography with time-synchronized fetal heart rate (FHR) monitoring (cardiotocography - CTG) in late pregnancy. CTGs were analysed for abnormal FHR events, including decelerations and reduced variability, by two blinded observers and exported into the sleep study to temporally align FHR events with sleep behaviours. For each FHR event, 10 control epochs with normal FHR were randomly selected for the same participant. Conditional logistic regression assessed the relationships between FHR events and sleep behaviours. From 116 participants, 52 had a total of 129 FHR events overnight; namely prolonged decelerations and prolonged periods of reduced variability. Significantly more FHR events were observed in women with FGR and/or a hypertensive disorder compared with uncomplicated pregnancies (P = 0.006). FHR events were twice as likely to be preceded by a change in body position within the previous 5 min, compared with control epochs (P = 0.007), particularly in hypertensive pregnancies both with and without FGR. Overall, FHR events were not temporally related to supine body position, respiratory events or snoring. Our results indicate that most fetuses tolerate sleep-related stressors, but further research is needed to identify the interplay of maternal and fetal conditions putting the fetus at risk overnight. KEY POINTS: Maternal sleep behaviours including supine position and sleep-disordered breathing are potential contributors to stillbirth but much of this work is based on self-reported data. Using time-synchronized polysomnography and cardiotocography, we found that nocturnal fetal heart rate decelerations were more likely to be preceded by a change in body position compared with epochs containing normal fetal heart rate, particularly in hypertensive pregnancies with or without fetal growth restriction. There was no temporal relationship between maternal sleeping position, snoring or apnoeic events and an abnormal fetal heart rate overnight. We conclude that most fetuses can tolerate sleep-related stressors with no evidence of fetal heart rate changes indicating compromised wellbeing. Further work needs to identify how sleep behaviours contribute to stillbirth risk and how these intersect with underlying maternal and fetal conditions.
-
ItemFetal heart rate events during sleep, and the impact of sleep disordered breathing, in pregnancies complicated by preterm fetal growth restriction: An exploratory observational case-control study(WILEY, 2022-12)OBJECTIVES: To evaluate fetal heart rate (FHR) patterns during sleep in pregnancies complicated by preterm fetal growth restriction (FGR). To determine whether co-existing sleep-disordered breathing (SDB) impacts on acute FHR events or perinatal outcome. DESIGN: Observational case control study. SETTING AND POPULATION: Women with preterm FGR and gestation-matched well grown controls (estimated fetal weight above the 10th percentile with normal Doppler studies); tertiary maternity hospital, Australia. METHODS: A polysomnogram, a test used to measure sleep patterns and diagnose sleep disorders, and concurrent cardiotocography (CTG), were analysed for respiratory events and FHR changes. MAIN OUTCOME MEASURES: Frequency of FHR events overnight in FGR cases versus controls and in those with or without SDB. RESULTS: Twenty-nine patients with preterm FGR and 29 controls (median estimated fetal weight 1st versus 60th percentile, P < 0.001) underwent polysomnography with concurrent CTG at a mean gestation of 30.2 weeks. The median number of FHR events per night was higher among FGR cases than among controls (3.0 events, interquartile range [IQR] 1.0-4.0, versus 1.0 [IQR 0-1.0]; P < 0.001). Women with pregnancies complicated by preterm FGR were more likely than controls to be nulliparous, receive antihypertensive medications, be supine at sleep onset, and to sleep supine (32.9% of total sleep time versus 18.3%, P = 0.03). SDB was common in both FGR and control pregnancies (48% versus 38%, respectively, P = 0.55) but was generally mild and not associated with an increase in overnight FHR events or adverse perinatal outcome. CONCLUSIONS: Acute FHR events overnight are more common in pregnancies complicated by preterm FGR than in pregnancies with normal fetal growth. Mild SDB was common in late pregnancy and well tolerated, even by fetuses with preterm FGR. TWEETABLE ABSTRACT: Mild sleep-disordered breathing seems well tolerated even by highly vulnerable fetuses.
-
ItemNo Preview AvailablePolysomnographic analysis of maternal sleep position and its relationship to pregnancy complications and sleep-disordered breathing(OXFORD UNIV PRESS INC, 2022-04-11)Links between supine "going to sleep" position and stillbirth risk have led to campaigns regarding safe maternal sleep position. This study profiles the distribution of sleep positions overnight and relationships to sleep onset position during pregnancy, and the relationships between supine sleep, sleep-disordered breathing (SDB), and pregnancy outcomes. Data from three prospective cohort studies evaluating SDB in healthy and complicated pregnancies were pooled. All participants underwent one night of polysomnography in late pregnancy and birth outcome data were collected. 187 women underwent polysomnography at a median gestation of 34 weeks'. The left lateral position was preferred for falling asleep (52%) compared to supine (14%), but sleep onset position was the dominant sleep position overnight in only half (54%) of women. The median percentage of sleep time in the supine position was 24.2%; women who fell asleep supine spent more time supine overnight compared to those who began non-supine (48.0% (30.0,65.9) vs. 22.6% (5.7,32.2), p < .001). Women with growth-restricted fetuses were more likely to fall asleep supine than those with well-grown fetuses (36.6% vs. 7.5%, p < .001). Positional SDB was observed in 46% of those with an RDI ≥ 5. Sleep onset position was the dominant position overnight for half of the sample, suggesting that sleep onset position is not always a reliable indicator of body position overnight. Supine sleep was related to fetal growth restriction and birthweight at delivery, though causality cannot be inferred. It is critical that we pursue research into verifying the important relationship between supine sleep and increased stillbirth risk, and the mechanisms behind it.
- 1 (current)
- 2
- 3
- 4
- 5