A machine-learning exploration of the exposome from preconception in early childhood atopic eczema, rhinitis and wheeze development.

BACKGROUND: Most previous research on the environmental epidemiology of childhood atopic eczema, rhinitis and wheeze is limited in the scope of risk factors studied. Our study adopted a machine learning approach to explore the role of the exposome starting already in the preconception phase. METHODS: We performed a combined analysis of two multi-ethnic Asian birth cohorts, the Growing Up in Singapore Towards healthy Outcomes (GUSTO) and the Singapore PREconception Study of long Term maternal and child Outcomes (S-PRESTO) cohorts. Interviewer-administered questionnaires were used to collect information on demography, lifestyle and childhood atopic eczema, rhinitis and wheeze development. Data training was performed using XGBoost, genetic algorithm and logistic regression models, and the top variables with the highest importance were identified. Additive explanation values were identified and inputted into a final multiple logistic regression model. Generalised structural equation modelling with maternal and child blood micronutrients, metabolites and cytokines was performed to explain possible mechanisms. RESULTS: The final study population included 1151 mother-child pairs. Our findings suggest that these childhood diseases are likely programmed in utero by the preconception and pregnancy exposomes through inflammatory pathways. We identified preconception alcohol consumption and maternal depressive symptoms during pregnancy as key modifiable maternal environmental exposures that increased eczema and rhinitis risk. Our mechanistic model suggested that higher maternal blood neopterin and child blood dimethylglycine protected against early childhood wheeze. After birth, early infection was a key driver of atopic eczema and rhinitis development. CONCLUSION: Preconception and antenatal exposomes can programme atopic eczema, rhinitis and wheeze development in utero. Reducing maternal alcohol consumption during preconception and supporting maternal mental health during pregnancy may prevent atopic eczema and rhinitis by promoting an optimal antenatal environment. Our findings suggest a need to include preconception environmental exposures in future research to counter the earliest precursors of disease development in children.

Deep learning for temporal data representation in electronic health records: A systematic review of challenges and methodologies.

OBJECTIVE: Temporal electronic health records (EHRs) contain a wealth of information for secondary uses, such as clinical events prediction and chronic disease management. However, challenges exist for temporal data representation. We therefore sought to identify these challenges and evaluate novel methodologies for addressing them through a systematic examination of deep learning solutions. METHODS: We searched five databases (PubMed, Embase, the Institute of Electrical and Electronics Engineers [IEEE] Xplore Digital Library, the Association for Computing Machinery [ACM] Digital Library, and Web of Science) complemented with hand-searching in several prestigious computer science conference proceedings. We sought articles that reported deep learning methodologies on temporal data representation in structured EHR data from January 1, 2010, to August 30, 2020. We summarized and analyzed the selected articles from three perspectives: nature of time series, methodology, and model implementation. RESULTS: We included 98 articles related to temporal data representation using deep learning. Four major challenges were identified, including data irregularity, heterogeneity, sparsity, and model opacity. We then studied how deep learning techniques were applied to address these challenges. Finally, we discuss some open challenges arising from deep learning. CONCLUSION: Temporal EHR data present several major challenges for clinical prediction modeling and data utilization. To some extent, current deep learning solutions can address these challenges. Future studies may consider designing comprehensive and integrated solutions. Moreover, researchers should incorporate clinical domain knowledge into study designs and enhance model interpretability to facilitate clinical implementation.

Effect of Training Data Volume on Performance of Convolutional Neural Network Pneumothorax Classifiers.

Large datasets with high-quality labels required to train deep neural networks are challenging to obtain in the radiology domain. This work investigates the effect of training dataset size on the performance of deep learning classifiers, focusing on chest radiograph pneumothorax detection as a proxy visual task in the radiology domain. Two open-source datasets (ChestX-ray14 and CheXpert) comprising 291,454 images were merged and convolutional neural networks trained with stepwise increase in training dataset sizes. Model iterations at each dataset volume were evaluated on an external test set of 525 emergency department chest radiographs. Learning curve analysis was performed to fit the observed AUCs for all models generated. For all three network architectures tested, model AUCs and accuracy increased rapidly from 2 × 103 to 20 × 103 training samples, with more gradual increase until the maximum training dataset size of 291 × 103 images. AUCs for models trained with the maximum tested dataset size of 291 × 103 images were significantly higher than models trained with 20 × 103 images: ResNet-50: AUC20k = 0.86, AUC291k = 0.95, p < 0.001; DenseNet-121 AUC20k = 0.85, AUC291k = 0.93, p < 0.001; EfficientNet AUC20k = 0.92, AUC 291 k = 0.98, p < 0.001. Our study established learning curves describing the relationship between dataset training size and model performance of deep learning convolutional neural networks applied to a typical radiology binary classification task. These curves suggest a point of diminishing performance returns for increasing training data volumes, which algorithm developers should consider given the high costs of obtaining and labelling radiology data.

Leveraging open data to reconstruct the Singapore Housing Index and other building-level markers of socioeconomic status for health services research.

BACKGROUND: Socioeconomic status (SES) is an important determinant of health, and SES data is an important confounder to control for in epidemiology and health services research. Individual level SES measures are cumbersome to collect and susceptible to biases, while area level SES measures may have insufficient granularity. The 'Singapore Housing Index' (SHI) is a validated, building level SES measure that bridges individual and area level measures. However, determination of the SHI has previously required periodic data purchase and manual parsing. In this study, we describe a means of SHI determination for public housing buildings with open government data, and validate this against the previous SHI determination method. METHODS: Government open data sources (e.g. DATA: gov.sg, Singapore Land Authority OneMAP API, Urban Redevelopment Authority API) were queried using custom Python scripts. Data on residential public housing block address and composition from the HDB Property Information dataset (data.gov.sg) was matched to postal code and geographical coordinates via OneMAP API calls. The SHI was calculated from open data, and compared to the original SHI dataset that was curated from non-open data sources in 2018. RESULTS: Ten thousand seventy-seven unique residential buildings were identified from open data. OneMAP API calls generated valid geographical coordinates for all (100%) buildings, and valid postal code for 10,012 (99.36%) buildings. There was an overlap of 10,011 buildings between the open dataset and the original SHI dataset. Intraclass correlation coefficient was 0.999 for the two sources of SHI, indicating almost perfect agreement. A Bland-Altman plot analysis identified a small number of outliers, and this revealed 5 properties that had an incorrect SHI assigned by the original dataset. Information on recently transacted property prices was also obtained for 8599 (85.3%) of buildings. CONCLUSION: SHI, a useful tool for health services research, can be accurately reconstructed using open datasets at no cost. This method is a convenient means for future researchers to obtain updated building-level markers of socioeconomic status for policy and research.

Systematic review on the definition and predictors of severe Clostridiodes difficile infection.

Clostridiodes difficile infection (CDI) is one of the most common hospital-acquired infections with high mortality rates. Optimal management of CDI depends on early recognition of severity. However, currently, there is no acceptable standard of prediction. We reviewed severe CDI predictors in published literature and its definition according to clinical guidelines. We systematically reviewed studies describing clinical predictors for severe CDI in medical databases (Cochrane, EMBASE, Global Health Library, and MEDLINE/PubMed). They were independently evaluated by two reviewers. Six hundred thirty-three titles and abstracts were screened, and 31 studies were included. We excluded studies that were restricted to a specific patient population. There were 16 articles that examined mortality in CDI, as compared with 15 articles investigating non-mortality outcomes of CDI. The commonest risk factors identified were comorbidities, white blood cell count, serum albumin level, age, serum creatinine level and intensive care unit admission. Generally, the studies had small patient populations, were retrospective in nature, and mostly from Western centers. The commonest severe CDI criteria in clinical guidelines were raised white blood cell count, followed by low serum albumin and raised serum creatinine levels. There was no commonly agreed upon definition of severe CDI severity in the literature. Current clinical guidelines' definitions for severe CDI are heterogeneous. Hence, there is a need for prospective multi-center studies using standardized protocol for biospecimen investigation collection and shared data on outcomes of patients in order to devise a universally accepted definition for severe CDI.

The impact of high frequency oscillatory ventilation on mortality in paediatric acute respiratory distress syndrome.

BACKGROUND: High-frequency oscillatory ventilation (HFOV) use was associated with greater mortality in adult acute respiratory distress syndrome (ARDS). Nevertheless, HFOV is still frequently used as rescue therapy in paediatric acute respiratory distress syndrome (PARDS). In view of the limited evidence for HFOV in PARDS and evidence demonstrating harm in adult patients with ARDS, we hypothesized that HFOV use compared to other modes of mechanical ventilation is associated with increased mortality in PARDS. METHODS: Patients with PARDS from 10 paediatric intensive care units across Asia from 2009 to 2015 were identified. Data on epidemiology and clinical outcomes were collected. Patients on HFOV were compared to patients on other modes of ventilation. The primary outcome was 28-day mortality and secondary outcomes were 28-day ventilator- (VFD) and intensive care unit- (IFD) free days. Genetic matching (GM) method was used to analyse the association between HFOV treatment with the primary outcome. Additionally, we performed a sensitivity analysis, including propensity score (PS) matching, inverse probability of treatment weighting (IPTW) and marginal structural modelling (MSM) to estimate the treatment effect. RESULTS: A total of 328 patients were included. In the first 7 days of PARDS, 122/328 (37.2%) patients were supported with HFOV. There were significant differences in baseline oxygenation index (OI) between the HFOV and non-HFOV groups (18.8 [12.0, 30.2] vs. 7.7 [5.1, 13.1] respectively; p < 0.001). A total of 118 pairs were matched in the GM method which found a significant association between HFOV with 28-day mortality in PARDS [odds ratio 2.3, 95% confidence interval (CI) 1.3, 4.4, p value 0.01]. VFD was indifferent between the HFOV and non-HFOV group [mean difference - 1.3 (95%CI - 3.4, 0.9); p = 0.29] but IFD was significantly lower in the HFOV group [- 2.5 (95%CI - 4.9, - 0.5); p = 0.03]. From the sensitivity analysis, PS matching, IPTW and MSM all showed consistent direction of HFOV treatment effect in PARDS. CONCLUSION: The use of HFOV was associated with increased 28-day mortality in PARDS. This study suggests caution but does not eliminate equivocality and a randomized controlled trial is justified to examine the true association.

Reinforcement Learning for Clinical Decision Support in Critical Care: Comprehensive Review.

BACKGROUND: Decision support systems based on reinforcement learning (RL) have been implemented to facilitate the delivery of personalized care. This paper aimed to provide a comprehensive review of RL applications in the critical care setting. OBJECTIVE: This review aimed to survey the literature on RL applications for clinical decision support in critical care and to provide insight into the challenges of applying various RL models. METHODS: We performed an extensive search of the following databases: PubMed, Google Scholar, Institute of Electrical and Electronics Engineers (IEEE), ScienceDirect, Web of Science, Medical Literature Analysis and Retrieval System Online (MEDLINE), and Excerpta Medica Database (EMBASE). Studies published over the past 10 years (2010-2019) that have applied RL for critical care were included. RESULTS: We included 21 papers and found that RL has been used to optimize the choice of medications, drug dosing, and timing of interventions and to target personalized laboratory values. We further compared and contrasted the design of the RL models and the evaluation metrics for each application. CONCLUSIONS: RL has great potential for enhancing decision making in critical care. Challenges regarding RL system design, evaluation metrics, and model choice exist. More importantly, further work is required to validate RL in authentic clinical environments.

The Effect of ARDS on Survival: Do Patients Die From ARDS or With ARDS?

OBJECTIVE:: To investigate the contribution of acute respiratory distress syndrome (ARDS) in of itself to mortality among ventilated patients. DESIGN AND SETTING:: A longitudinal retrospective study of ventilated intensive care unit (ICU) patients. PATIENTS:: The analysis included patients ventilated for more than 48 hours. Patients were classified as having ARDS on admission (early-onset ARDS), late-onset ARDS (ARDS not present during the first 24 hours of admission), or no ARDS. Primary outcomes were mortality at 28 days, and secondary outcomes were 2-year mortality rate from ICU admission. RESULTS:: A total of 1411 ventilated patients were enrolled: 41% had ARDS on admission, 28.5% developed ARDS during their ICU stay, and 30.5% did not meet the ARDS criteria prior to ICU discharge or death. The non-ARDS group was used as the control. We also divided the cohort based on the severity of ARDS. After adjusting for covariates, mortality risk at 28 days was not significantly different among the different groups. Both early- and late-onset ARDS as well as the severity of ARDS were found to be significant risk factors for 2 years from ICU survival. CONCLUSION:: Among patients who were ventilated on ICU admission, neither the presence, the severity, or the timing of ARDS contribute independently to the short-term mortality risk. However, acute respiratory distress syndrome does contribute significantly to 2-year mortality risk. This suggests that patients may not die acutely from ARDS itself but rather from the primary disease, and during the acute phase of ARDS, clinicians should focus on improving treatment strategies for the diseases that led to ARDS.

Quantifying the Mortality Impact of Do-Not-Resuscitate Orders in the ICU.

OBJECTIVES: We quantified the 28-day mortality effect of preexisting do-not-resuscitate orders in ICUs. DESIGN: Longitudinal, retrospective study of patients admitted to five ICUs at a tertiary university medical center (Beth Israel Deaconess Medical Center, BIDMC, Boston, MA) between 2001 and 2008. INTERVENTION: None. PATIENTS: Two cohorts were defined: patients with do not resuscitate advance directives on day 1 of ICU admission and a control group comprising patients with no limitations of level of care on ICU day 1 (full code). MEASUREMENTS AND MAIN RESULTS: The primary outcome was mortality at 28 days after ICU admission. Of 19,007 ICU patients, 1,239 patients (6.5%) had a do-not-resuscitate order on the first day of ICU admission and survived 48 hours in the ICU. We matched those do-not-resuscitate patients with 2,402 patients with full-code status. Twenty-eight day and 1-year mortality were both significantly higher in the do-not-resuscitate group (33.9% vs 18.4% and 60.7% vs 40.2%; p < 0.001, respectively). CONCLUSION: Do-not-resuscitate status is an independent risk factor for ICU mortality. This may reflect severity of illness not captured by other clinical factors, but the perceptions of the treating team related to do-not-resuscitate status could also be causally responsible for increased mortality in patients with do-not-resuscitate status.

Obesity, Acute Kidney Injury, and Mortality in Critical Illness.

OBJECTIVES: Although obesity is associated with risk for chronic kidney disease and improved survival, less is known about the associations of obesity with risk of acute kidney injury and post acute kidney injury mortality. DESIGN: In a single-center inception cohort of almost 15,000 critically ill patients, we evaluated the association of obesity with acute kidney injury and acute kidney injury severity, as well as in-hospital and 1-year survival. Acute kidney injury was defined using the Kidney Disease Outcome Quality Initiative criteria. MEASUREMENTS AND MAIN RESULTS: The acute kidney injury prevalence rates for normal, overweight, class I, II, and III obesity were 18.6%, 20.6%, 22.5%, 24.3%, and 24.0%, respectively, and the adjusted odds ratios of acute kidney injury were 1.18 (95% CI, 1.06-1.31), 1.35 (1.19-1.53), 1.47 (1.25-1.73), and 1.59 (1.31-1.87) when compared with normal weight, respectively. Each 5-kg/m² increase in body mass index was associated with a 10% risk (95% CI, 1.06-1.24; p < 0.001) of more severe acute kidney injury. Within-hospital and 1-year survival rates associated with the acute kidney injury episodes were similar across body mass index categories. CONCLUSION: Obesity is a risk factor for acute kidney injury, which is associated with increased short- and long-term mortality.

Sodium modelling to reduce intradialytic hypotension during haemodialysis for acute kidney injury in the intensive care unit.

AIM: Intradialytic hypotension often complicates haemodialysis for patients with acute kidney injury (AKI) and may impact renal recovery. Sodium modelling is sometimes used as prophylaxis against intradialytic hypotension in the chronic haemodialysis population, but there is little evidence for its use among critically ill patients with AKI. METHODS: A retrospective cohort with AKI requiring intermittent haemodialysis in the intensive care unit from 2001 to 2008 was used to study the association of prophylactic sodium modelling and multiple outcomes. Outcomes included a composite of in-hospital death or dialysis dependence at hospital discharge, as well as intradialytic hypotension, ultrafiltration goal achievement and net ultrafiltration volume. Associations were estimated using logistic regression, mixed linear models and generalized estimating equations adjusting for demographic and clinical characteristics. RESULTS: One hundred and ninety-one individuals who underwent 892 sessions were identified; sodium modelling was prescribed in 27.1% of the sessions. In adjusted analyses, sodium modelling was not significantly associated with intradialytic hypotension (P = 0.67) or with the ultrafiltration goal achievement (P = 0.06). Sodium modelling during the first dialysis session was numerically associated with lower risk for the composite of in-hospital death or dialysis dependence: adjusted odds ratio (95% confidence interval) 0.39 (0.15-1.02; P = 0.06); however, this association did not reach statistical significance. CONCLUSION: We did not observe statistically significant associations between sodium modelling and improved outcomes among AKI patients receiving intermittent dialysis in the intensive care unit. However, suggestive findings warrant further study.

Outcome Prediction for Patients with Traumatic Brain Injury with Dynamic Features from Intracranial Pressure and Arterial Blood Pressure Signals: A Gaussian Process Approach.

Previous work has been demonstrated that tracking features describing the dynamic and time-varying patterns in brain monitoring signals provide additional predictive information beyond that derived from static features based on snapshot measurements. To achieve more accurate predictions of outcomes of patients with traumatic brain injury (TBI), we proposed a statistical framework to extract dynamic features from brain monitoring signals based on the framework of Gaussian processes (GPs). GPs provide an explicit probabilistic, nonparametric Bayesian approach to metric regression problems. This not only provides probabilistic predictions, but also gives the ability to cope with missing data and infer model parameters such as those that control the function's shape, noise level and dynamics of the signal. Through experimental evaluation, we have demonstrated that dynamic features extracted from GPs provide additional predictive information in addition to the features based on the pressure reactivity index (PRx). Significant improvements in patient outcome prediction were achieved by combining GP-based and PRx-based dynamic features. In particular, compared with the a baseline PRx-based model, the combined model achieved over 30 % improvement in prediction accuracy and sensitivity and over 20 % improvement in specificity and the area under the receiver operating characteristic curve.

Hyperdynamic left ventricular ejection fraction in the intensive care unit.

INTRODUCTION: Limited information exists on the etiology, prevalence, and significance of hyperdynamic left ventricular ejection fraction (HDLVEF) in the intensive care unit (ICU). Our aim in the present study was to compare characteristics and outcomes of patients with HDLVEF with those of patients with normal left ventricular ejection fraction in the ICU using a large, public, deidentified critical care database. METHODS: We conducted a longitudinal, single-center, retrospective cohort study of adult patients who underwent echocardiography during a medical or surgical ICU admission at the Beth Israel Deaconess Medical Center using the Multiparameter Intelligent Monitoring in Intensive Care II database. The final cohort had 2867 patients, of whom 324 had HDLVEF, defined as an ejection fraction >70%. Patients with an ejection fraction <55% were excluded. RESULTS: Compared with critically ill patients with normal left ventricular ejection fraction, the finding of HDLVEF in critically ill patients was associated with female sex, increased age, and the diagnoses of hypertension and cancer. Patients with HDLVEF had increased 28-day mortality compared with those with normal ejection fraction in multivariate logistic regression analysis adjusted for age, sex, Sequential Organ Failure Assessment score, Elixhauser score for comorbidities, vasopressor use, and mechanical ventilation use (odds ratio 1.38, 95% confidence interval 1.039-1.842, p =0.02). CONCLUSIONS: The presence of HDLVEF portended increased 28-day mortality, and may be helpful as a gravity marker for prognosis in patients admitted to the ICU. Further research is warranted to gain a better understanding of how these patients respond to common interventions in the ICU and to determine if pharmacologic modulation of HDLVEF improves outcomes.

Novel SNP improves differential survivability and mortality in non-small cell lung cancer patients.

BACKGROUND: Non-small cell lung cancer (NSCLC) is a major cause of cancer-related death worldwide due to poor patient prognosis and clinical outcome. Here, we studied the genetic variations underlying NSCLC pathogenesis based on their association to patient outcome after gemcitabine therapy. RESULTS: Bioinformatics analysis was used to investigate possible effects of POLA2 G583R (POLA2+1747 GG/GA, dbSNP ID: rs487989) in terms of protein function. Using biostatistics, POLA2+1747 GG/GA (rs487989, POLA2 G583R) was identified as strongly associated with mortality rate and survival time among NSCLC patients. It was also shown that POLA2+1747 GG/GA is functionally significant for protein localization via green fluorescent protein (GFP)-tagging and confocal laser scanning microscopy analysis. The single nucleotide polymorphism (SNP) causes DNA polymerase alpha subunit B to localize in the cytoplasm instead of the nucleus. This inhibits DNA replication in cancer cells and confers a protective effect in individuals with this SNP. CONCLUSIONS: The results suggest that POLA2+1747 GG/GA may be used as a prognostic biomarker of patient outcome in NSCLC pathogenesis.

The effect of age and clinical circumstances on the outcome of red blood cell transfusion in critically ill patients.

INTRODUCTION: Whether red blood cell (RBC) transfusion is beneficial remains controversial. In both retrospective and prospective evaluations, transfusion has been associated with adverse, neutral, or protective effects. These varying results likely stem from a complex interplay between transfusion, patient characteristics, and clinical context. The objective was to test whether age, comorbidities, and clinical context modulate the effect of transfusion on survival. METHODS: By using the multiparameter intelligent monitoring in intensive care II database (v. 2.6), a retrospective analysis of 9,809 critically ill patients, we evaluated the effect of RBC transfusion on 30-day and 1-year mortality. Propensity score modeling and logistic regression adjusted for known confounding and assessed the independent effect of transfusion on 30-day and 1-year mortality. Sensitivity analysis was performed by using 3,164 transfused and non-transfused pairs, matched according the previously validated propensity model for RBC transfusion. RESULTS: RBC transfusion did not affect 30-day or 1-year mortality in the overall cohort. Patients younger than 55 years had increased odds of mortality (OR, 1.71; P < 0.01) with transfusion. Patients older than 75 years had lower odds of 30-day and 1-year mortality (OR, 0.70; P < 0.01) with transfusion. Transfusion was associated with worse outcome among patients undergoing cardiac surgery (OR, 2.1; P < 0.01). The propensity-matched population corroborated findings identified by regression adjustment. CONCLUSION: A complex relation exists between RBC transfusion and clinical outcome. Our results show that transfusion is associated with improved outcomes in some cohorts and worse outcome in others, depending on comorbidities and patient characteristics. As such, future investigations and clinical decisions evaluating the value of transfusion should account for variations in baseline characteristics and clinical context.

A review of uncertainty quantification in medical image analysis: Probabilistic and non-probabilistic methods.

The comprehensive integration of machine learning healthcare models within clinical practice remains suboptimal, notwithstanding the proliferation of high-performing solutions reported in the literature. A predominant factor hindering widespread adoption pertains to an insufficiency of evidence affirming the reliability of the aforementioned models. Recently, uncertainty quantification methods have been proposed as a potential solution to quantify the reliability of machine learning models and thus increase the interpretability and acceptability of the results. In this review, we offer a comprehensive overview of the prevailing methods proposed to quantify the uncertainty inherent in machine learning models developed for various medical image tasks. Contrary to earlier reviews that exclusively focused on probabilistic methods, this review also explores non-probabilistic approaches, thereby furnishing a more holistic survey of research pertaining to uncertainty quantification for machine learning models. Analysis of medical images with the summary and discussion on medical applications and the corresponding uncertainty evaluation protocols are presented, which focus on the specific challenges of uncertainty in medical image analysis. We also highlight some potential future research work at the end. Generally, this review aims to allow researchers from both clinical and technical backgrounds to gain a quick and yet in-depth understanding of the research in uncertainty quantification for medical image analysis machine learning models.

Frailty Modeling using Machine Learning Methodologies: A Systematic Review with Discussions on Outstanding Questions.

Studying frailty is crucial for enhancing the health and quality of life among older adults, refining healthcare delivery methods, and tackling the obstacles linked to an aging demographic. Approaches to frailty modeling often utilise simple analytic techniques rather than available advanced machine learning methods, which may be sub-optimal. There is no large-scale systematic review on applications of machine learning methods on frailty modeling. In this study we explore the use of machine learning methods to predict or classify frailty in older persons in routinely collected data. We reviewed 181 research articles, and categorised analytic methods into three categories: generalised linear models, survival models, and non-linear models. These methods have a moderate agreement with existing frailty scores and predictive validity for adverse outcomes. Limited evidence suggests that non-linear methods outperform generalised linear methods. The top-three predictor/input variables are specific diagnosis or groups of diagnoses, functional performance (e.g., ADLs), and impaired cognition. Mortality, hospital admissions and prolonged hospital stay are the mainly predicted outcomes. Most studies utilise classical machine learning methods with cross-sectional data. Longitudinal data collected by wearable sensors have been used for frailty modeling. We also discuss the opportunities to use more advanced machine learning methods with high dimensional longitudinal data for more personalised and accessible frailty tools.

Predicting intubation for intensive care units patients: A deep learning approach to improve patient management.

OBJECTIVE: For patients in the Intensive Care Unit (ICU), the timing of intubation has a significant association with patients' outcomes. However, accurate prediction of the timing of intubation remains an unsolved challenge due to the noisy, sparse, heterogeneous, and unbalanced nature of ICU data. In this study, our objective is to develop a workflow for pre-processing ICU data and to develop a customized deep learning model to predict the need for intubation. METHODS: To improve the prediction accuracy, we transform the intubation prediction task into a time series classification task. We carefully design a sequence of data pre-processing steps to handle the multimodal noisy data. Firstly, we discretize the sequential data and address missing data using interpolation. Next, we employ a sampling strategy to address data imbalance and standardize the data to facilitate faster model convergence. Furthermore, we employ the feature selection technique and propose an ensemble model to combine features learned by different deep learning models. RESULTS: The performance is evaluated on Medical Information Mart for Intensive Care (MIMIC)-III, an ICU dataset. Our proposed Deep Feature Fusion method achieves an area under the curve (AUC) of the receiver operating curve (ROC) of 0.8953, surpassing the performance of other deep learning and traditional machine learning models. CONCLUSION: Our proposed Deep Feature Fusion method proves to be a viable approach for predicting intubation and outperforms other deep learning and classical machine learning models. The study confirms that high-frequency time-varying indicators, particularly Mean Blood Pressure (MeanBP) and peripheral oxygen saturation (SpO2), are significant risk factors for predicting intubation.

The SingHealth Perioperative and Anesthesia Subject Area Registry (PASAR), a large-scale perioperative data mart and registry.

BACKGROUND: To enhance perioperative outcomes, a perioperative registry that integrates high-quality real-world data throughout the perioperative period is essential. Singapore General Hospital established the Perioperative and Anesthesia Subject Area Registry (PASAR) to unify data from the preoperative, intraoperative, and postoperative stages. This study presents the methodology employed to create this database. METHODS: Since 2016, data from surgical patients have been collected from the hospital electronic medical record systems, de-identified, and stored securely in compliance with privacy and data protection laws. As a representative sample, data from initiation in 2016 to December 2022 were collected. RESULTS: As of December 2022, PASAR data comprise 26 tables, encompassing 153,312 patient admissions and 168,977 operation sessions. For this period, the median age of the patients was 60.0 years, sex distribution was balanced, and the majority were Chinese. Hypertension and cardiovascular comorbidities were also prevalent. Information including operation type and time, intensive care unit (ICU) length of stay, and 30-day and 1-year mortality rates were collected. Emergency surgeries resulted in longer ICU stays, but shorter operation times than elective surgeries. CONCLUSIONS: The PASAR provides a comprehensive and automated approach to gathering high-quality perioperative patient data.

Association of body temperature and mortality in critically ill patients: an observational study using two large databases.

BACKGROUND: Body temperature (BT) is routinely measured and can be controlled in critical care settings. BT can impact patient outcome, but the relationship between BT and mortality has not been well-established. METHODS: A retrospective cohort study was conducted based on the MIMIC-IV (N = 43,537) and eICU (N = 75,184) datasets. The primary outcome and exposure variables were hospital mortality and first 48-h median BT, respectively. Generalized additive models were used to model the associations between exposures and outcomes, while adjusting for patient age, sex, APS-III, SOFA, and Charlson comorbidity scores, temperature gap, as well as ventilation, vasopressor, steroids, and dialysis usage. We conducted subgroup analysis according to ICU setting, diagnoses, and demographics. RESULTS: Optimal BT was 37 °C for the general ICU and subgroup populations. A 10% increase in the proportion of time that BT was within the 36-38 °C range was associated with reduced hospital mortality risk in both MIMIC-IV (OR 0.91; 95% CI 0.90-0.93) and eICU (OR 0.86; 95% CI 0.85-0.87). On the other hand, a 10% increase in the proportion of time when BT < 36 °C was associated with increased mortality risk in both MIMIC-IV (OR 1.08; 95% CI 1.06-1.10) and eICU (OR 1.18; 95% CI 1.16-1.19). Similarly, a 10% increase in the proportion of time when BT > 38 °C was associated with increased mortality risk in both MIMIC-IV (OR 1.09; 95% CI 1.07-1.12) and eICU (OR 1.09; 95% CI 1.08-1.11). All patient subgroups tested consistently showed an optimal temperature within the 36-38 °C range. CONCLUSIONS: A BT of 37 °C is associated with the lowest mortality risk among ICU patients. Further studies to explore the causal relationship between the optimal BT and mortality should be conducted and may help with establishing guidelines for active BT management in critical care settings.