Prediction of extubation failure in the paediatric cardiac ICU using machine learning and high-frequency physiologic data

Background: Cardiac intensivists frequently assess patient readiness to wean off mechanical ventilation with an extubation readiness trial despite it being no more effective than clinician judgement alone. We evaluated the utility of high-frequency physiologic data and machine learning for improving the prediction of extubation failure in children with cardiovascular disease. Methods: This was a retrospective analysis of clinical registry data and streamed physiologic extubation readiness trial data from one paediatric cardiac ICU (12/2016-3/2018). We analysed patients’ final extubation readiness trial. Machine learning methods (classification and regression tree, Boosting, Random Forest) were performed using clinical/demographic data, physiologic data, and both datasets. Extubation failure was defined as reintubation within 48 hrs. Classifier performance was assessed on prediction accuracy and area under the receiver operating characteristic curve. Results: Of 178 episodes, 11.2% (N = 20) failed extubation. Using clinical/demographic data, our machine learning methods identified variables such as age, weight, height, and ventilation duration as being important in predicting extubation failure. Best classifier performance with this data was Boosting (prediction accuracy: 0.88; area under the receiver operating characteristic curve: 0.74). Using physiologic data, our machine learning methods found oxygen saturation extremes and descriptors of dynamic compliance, central venous pressure, and heart/respiratory rate to be of importance. The best classifier in this setting was Random Forest (prediction accuracy: 0.89; area under the receiver operating characteristic curve: 0.75). Combining both datasets produced classifiers highlighting the importance of physiologic variables in determining extubation failure, though predictive performance was not improved. Conclusion: Physiologic variables not routinely scrutinised during extubation readiness trials were identified as potential extubation failure predictors. Larger analyses are necessary to investigate whether these markers can improve clinical decision-making.

Insomnia with physiological hyperarousal is associated with lower weight: a novel finding and its clinical implications

Previous studies on the association of insomnia with body mass index (BMI) have been controversial. Physiological hyperarousal, the key pathological mechanism of insomnia, may be an important reason for different findings. We explored whether insomnia with physiological hyperarousal measured by the multiple sleep latency test (MSLT) is associated with body-weight differences. A total of 185 normal sleepers and 440 insomniacs were included in this study. Insomnia was defined by standard diagnostic criteria with symptoms lasting ≥6 months. All subjects underwent one night of laboratory polysomnography followed by a standard MSLT. We used the median MSLT value (i.e., ≥14 min) to define physiological hyperarousal. BMI was based on measured height (cm) and weight (kg) during the subjects’ sleep laboratory visit. BMI > 25 kg/m2 was defined as overweight, while BMI < 18.5 kg/m2 was defined as underweight. After controlling for confounders, the odds of lower weight rather than overweight were significantly increased among insomnia patients with increased MSLT: insomnia with MSLT 14–17 min and MSLT > 17 min increased the odds of lower weight by approximately 89% (OR = 1.89, 95% CI 1.00–4.85) and 273% (OR = 3.73, 95% CI 1.51–9.22) compared with normal sleepers, respectively. In contrast, insomnia in patients with MSLT 11–14 min and 8–11 min was not different from normal sleepers in terms of body weight. Insomnia associated with physiological hyperarousal, the most severe phenotype of chronic insomnia, is associated with higher odds of lower weight and underweight compared with normal sleepers. This is a novel finding consistent with previous physiologic data and has significant clinical implications.

Artificial Intelligence May Predict Early Sepsis After Liver Transplantation

Background: Sepsis, post-liver transplantation, is a frequent challenge that impacts patient outcomes. We aimed to develop an artificial intelligence method to predict the onset of post-operative sepsis earlier.Methods: This pilot study aimed to identify “physiomarkers” in continuous minute-by-minute physiologic data streams, such as heart rate, respiratory rate, oxygen saturation (SpO2), and blood pressure, to predict the onset of sepsis. The model was derived from a cohort of 5,748 transplant and non-transplant patients across intensive care units (ICUs) over 36 months, with 92 post-liver transplant patients who developed sepsis.Results: Using an alert timestamp generated with the Third International Consensus Definition of Sepsis (Sepsis-3) definition as a reference point, we studied up to 24 h of continuous physiologic data prior to the event, totaling to 8.35 million data points. One hundred fifty-five features were generated using signal processing and statistical methods. Feature selection identified 52 highly ranked features, many of which included blood pressures. An eXtreme Gradient Boost (XGB) classifier was then trained on the ranked features by 5-fold cross validation on all patients (n = 5,748). We identified that the average sensitivity, specificity, positive predictive value (PPV), and area under the receiver-operator curve (AUC) of the model after 100 iterations was 0.94 ± 0.02, 0.9 ± 0.02, 0.89 ± 0.01, respectively, and 0.97 ± 0.01 for predicting sepsis 12 h before meeting criteria.Conclusion: The data suggest that machine learning/deep learning can be applied to continuous streaming data in the transplant ICU to monitor patients and possibly predict sepsis.

Wireless versus routine physiologic monitoring after cesarean delivery to reduce maternal morbidity and mortality in a resource-limited setting: protocol of type 2 hybrid effectiveness-implementation study

Background Women in sub-Saharan Africa have the highest rates of morbidity and mortality during childbirth globally. Despite increases in facility-based childbirth, gaps in quality of care at facilities have limited reductions in maternal deaths. Infrequent physiologic monitoring of women around childbirth is a major gap in care that leads to delays in life-saving interventions for women experiencing complications. Methods We will conduct a type-2 hybrid effectiveness-implementation study over 12 months to evaluate using a wireless physiologic monitoring system to detect and alert clinicians of abnormal vital signs in women for 24 h after undergoing emergency cesarean delivery at a tertiary care facility in Uganda. We will provide physiologic data (heart rate, respiratory rate, temperature and blood pressure) to clinicians via a smartphone-based application with alert notifications if monitored women develop predefined abnormalities in monitored physiologic signs. We will alternate two-week intervention and control time periods where women and clinicians use the wireless monitoring system during intervention periods and current standard of care (i.e., manual vital sign measurement when clinically indicated) during control periods. Our primary outcome for effectiveness is a composite of severe maternal outcomes per World Health Organization criteria (e.g. death, cardiac arrest, jaundice, shock, prolonged unconsciousness, paralysis, hysterectomy). Secondary outcomes include maternal mortality rate, and case fatality rates for postpartum hemorrhage, hypertensive disorders, and sepsis. We will use the RE-AIM implementation framework to measure implementation metrics of the wireless physiologic system including Reach (proportion of eligible women monitored, length of time women monitored), Efficacy (proportion of women with monitoring according to Uganda Ministry of Health guidelines, number of appropriate alerts sent), Adoption (proportion of clinicians utilizing physiologic data per shift, clinical actions in response to alerts), Implementation (fidelity to monitoring protocol), Maintenance (sustainability of implementation over time). We will also perform in-depth qualitative interviews with up to 30 women and 30 clinicians participating in the study. Discussion This is the first hybrid-effectiveness study of wireless physiologic monitoring in an obstetric population. This study offers insights into use of wireless monitoring systems in low resource-settings, as well as normal and abnormal physiologic parameters among women delivering by cesarean. Trial registration ClinicalTrials.gov, NCT04060667. Registered on 08/01/2019.

Correlating Physiologic Measures of Stress: Exploring Dyads in Clinical Surgical Teams

Clinical teams are subject to stress from various sources, including the technical and cognitive challenges of providing care in high stakes environments. Existing analytic approaches are limited in their ability to study the interdependence of team member stress. This study explores the correlation of a physiologic marker of stress, blood pulse wave, between members of a working surgical team. We propose an area overlap method as a means of evaluating blood pulse wave time-series correlation as a function of time. This is a stepwise approach to the collection and analysis of a large volume of continuous physiologic data from paired team members in a clinical setting. This method was applied to thirteen surgical team dyads with similar results to Pearson correlation. The area overlap method allows for improved exploration of temporal correlation within dyads but, in its current form, does not identify directionality of correlation.

Pseudocholinesterase Deficiency Considerations: A Case Study

Pseudocholinesterase deficiency, sometimes called butyrylcholinesterase deficiency, is a rare disorder in which the neuromuscular blocking drugs succinylcholine and mivacurium cannot be metabolized properly in the blood plasma. This disorder can either be acquired as a result of certain comorbidities or it can be inherited genetically. Anesthesia providers must understand the pathophysiology of pseudocholinesterase deficiency and be prepared to safely and effectively manage patients who show signs and symptoms consistent with the disorder after the use of the indicated neuromuscular blocking drugs. This article summarizes the pharmacologic and physiologic data relevant to understanding the basic pathophysiology associated with pseudocholinesterase deficiency and illustrates a case study of a young woman suspected of having the disorder after a prolonged delay in emergence from general anesthesia.

Blood product transfusion during air medical transport: A needs assessment

ABSTRACTObjectivesEarly administration of blood products to patients with hemorrhagic shock has a positive impact on morbidity and mortality. Smaller hospitals may have limited supply of blood, and air medical systems may not carry blood. The primary outcome is to quantify the number of patients meeting established physiologic criteria for blood product administration and to identify which patients receive and which ones do not receive it due to lack of availability locally.MethodsElectronic patient care records were used to identify a retrospective cohort of patients undergoing emergent air medical transport in Ontario, Canada, who are likely to require blood. Presenting problems for blood product administration were identified. Physiologic data were extracted with criteria for transfusion used to identify patients where blood product administration is indicated.ResultsThere were 11,520 emergent patient transports during the study period, with 842 (7.3%) where blood product administration was considered. Of these, 290 met established physiologic criteria for blood products, with 167 receiving blood, of which 57 received it at a hospital with a limited supply. The mean number of units administered per patient was 3.5. The remaining 123 patients meeting criteria did not receive product because none was unavailable.ConclusionIndications for blood product administration are present in 2.5% of patients undergoing time-sensitive air medical transport. Air medical services can enhance access to potentially lifesaving therapy in patients with hemorrhagic shock by carrying blood products, as blood may be unavailable or in limited supply locally in the majority of patients where it is indicated.