Machine learning clinical prediction models for acute kidney injury: the impact of baseline creatinine on prediction efficacy.

BACKGROUND: There are many Machine Learning (ML) models which predict acute kidney injury (AKI) for hospitalised patients. While a primary goal of these models is to support clinical decision-making, the adoption of inconsistent methods of estimating baseline serum creatinine (sCr) may result in a poor understanding of these models' effectiveness in clinical practice. Until now, the performance of such models with different baselines has not been compared on a single dataset. Additionally, AKI prediction models are known to have a high rate of false positive (FP) events regardless of baseline methods. This warrants further exploration of FP events to provide insight into potential underlying reasons. OBJECTIVE: The first aim of this study was to assess the variance in performance of ML models using three methods of baseline sCr on a retrospective dataset. The second aim was to conduct an error analysis to gain insight into the underlying factors contributing to FP events. MATERIALS AND METHODS: The Intensive Care Unit (ICU) patients of the Medical Information Mart for Intensive Care (MIMIC)-IV dataset was used with the KDIGO (Kidney Disease Improving Global Outcome) definition to identify AKI episodes. Three different methods of estimating baseline sCr were defined as (1) the minimum sCr, (2) the Modification of Diet in Renal Disease (MDRD) equation and the minimum sCr and (3) the MDRD equation and the mean of preadmission sCr. For the first aim of this study, a suite of ML models was developed for each baseline and the performance of the models was assessed. An analysis of variance was performed to assess the significant difference between eXtreme Gradient Boosting (XGB) models across all baselines. To address the second aim, Explainable AI (XAI) methods were used to analyse the XGB errors with Baseline 3. RESULTS: Regarding the first aim, we observed variances in discriminative metrics and calibration errors of ML models when different baseline methods were adopted. Using Baseline 1 resulted in a 14% reduction in the f1 score for both Baseline 2 and Baseline 3. There was no significant difference observed in the results between Baseline 2 and Baseline 3. For the second aim, the FP cohort was analysed using the XAI methods which led to relabelling data with the mean of sCr in 180 to 0 days pre-ICU as the preferred sCr baseline method. The XGB model using this relabelled data achieved an AUC of 0.85, recall of 0.63, precision of 0.54 and f1 score of 0.58. The cohort size was 31,586 admissions, of which 5,473 (17.32%) had AKI. CONCLUSION: In the absence of a widely accepted method of baseline sCr, AKI prediction studies need to consider the impact of different baseline methods on the effectiveness of ML models and their potential implications in real-world implementations. The utilisation of XAI methods can be effective in providing insight into the occurrence of prediction errors. This can potentially augment the success rate of ML implementation in routine care.

Systematic review of externally validated machine learning models for predicting acute kidney injury in general hospital patients.

Acute kidney injury (AKI) is one of the most common and consequential complications among hospitalized patients. Timely AKI risk prediction may allow simple interventions that can minimize or avoid the harm associated with its development. Given the multifactorial and complex etiology of AKI, machine learning (ML) models may be best placed to process the available health data to generate accurate and timely predictions. Accordingly, we searched the literature for externally validated ML models developed from general hospital populations using the current definition of AKI. Of 889 studies screened, only three were retrieved that fit these criteria. While most models performed well and had a sound methodological approach, the main concerns relate to their development and validation in populations with limited diversity, comparable digital ecosystems, use of a vast number of predictor variables and over-reliance on an easily accessible biomarker of kidney injury. These are potentially critical limitations to their applicability in diverse socioeconomic and cultural settings, prompting a need for simpler, more transportable prediction models which can offer a competitive advantage over the current tools used to predict and diagnose AKI.

A multi-country analysis of COVID-19 hospitalizations by vaccination status.

BACKGROUND: Individuals vaccinated against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), when infected, can still develop disease that requires hospitalization. It remains unclear whether these patients differ from hospitalized unvaccinated patients with regard to presentation, coexisting comorbidities, and outcomes. METHODS: Here, we use data from an international consortium to study this question and assess whether differences between these groups are context specific. Data from 83,163 hospitalized COVID-19 patients (34,843 vaccinated, 48,320 unvaccinated) from 38 countries were analyzed. FINDINGS: While typical symptoms were more often reported in unvaccinated patients, comorbidities, including some associated with worse prognosis in previous studies, were more common in vaccinated patients. Considerable between-country variation in both in-hospital fatality risk and vaccinated-versus-unvaccinated difference in this outcome was observed. CONCLUSIONS: These findings will inform allocation of healthcare resources in future surges as well as design of longer-term international studies to characterize changes in clinical profile of hospitalized COVID-19 patients related to vaccination history. FUNDING: This work was made possible by the UK Foreign, Commonwealth and Development Office and Wellcome (215091/Z/18/Z, 222410/Z/21/Z, 225288/Z/22/Z, and 220757/Z/20/Z); the Bill & Melinda Gates Foundation (OPP1209135); and the philanthropic support of the donors to the University of Oxford's COVID-19 Research Response Fund (0009109). Additional funders are listed in the "acknowledgments" section.

Physical Grounds for Causal Perspectivalism.

We ground the asymmetry of causal relations in the internal physical states of a special kind of open and irreversible physical system, a causal agent. A causal agent is an autonomous physical system, maintained in a steady state, far from thermal equilibrium, with special subsystems: sensors, actuators, and learning machines. Using feedback, the learning machine, driven purely by thermodynamic constraints, changes its internal states to learn probabilistic functional relations inherent in correlations between sensor and actuator records. We argue that these functional relations just are causal relations learned by the agent, and so such causal relations are simply relations between the internal physical states of a causal agent. We show that learning is driven by a thermodynamic principle: the error rate is minimised when the dissipated power is minimised. While the internal states of a causal agent are necessarily stochastic, the learned causal relations are shared by all machines with the same hardware embedded in the same environment. We argue that this dependence of causal relations on such 'hardware' is a novel demonstration of causal perspectivalism.

Association of Country Income Level With the Characteristics and Outcomes of Critically Ill Patients Hospitalized With Acute Kidney Injury and COVID-19.

Introduction: Acute kidney injury (AKI) has been identified as one of the most common and significant problems in hospitalized patients with COVID-19. However, studies examining the relationship between COVID-19 and AKI in low- and low-middle income countries (LLMIC) are lacking. Given that AKI is known to carry a higher mortality rate in these countries, it is important to understand differences in this population. Methods: This prospective, observational study examines the AKI incidence and characteristics of 32,210 patients with COVID-19 from 49 countries across all income levels who were admitted to an intensive care unit during their hospital stay. Results: Among patients with COVID-19 admitted to the intensive care unit, AKI incidence was highest in patients in LLMIC, followed by patients in upper-middle income countries (UMIC) and high-income countries (HIC) (53%, 38%, and 30%, respectively), whereas dialysis rates were lowest among patients with AKI from LLMIC and highest among those from HIC (27% vs. 45%). Patients with AKI in LLMIC had the largest proportion of community-acquired AKI (CA-AKI) and highest rate of in-hospital death (79% vs. 54% in HIC and 66% in UMIC). The association between AKI, being from LLMIC and in-hospital death persisted even after adjusting for disease severity. Conclusions: AKI is a particularly devastating complication of COVID-19 among patients from poorer nations where the gaps in accessibility and quality of healthcare delivery have a major impact on patient outcomes.

ISARIC-COVID-19 dataset: A Prospective, Standardized, Global Dataset of Patients Hospitalized with COVID-19.

The International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC) COVID-19 dataset is one of the largest international databases of prospectively collected clinical data on people hospitalized with COVID-19. This dataset was compiled during the COVID-19 pandemic by a network of hospitals that collect data using the ISARIC-World Health Organization Clinical Characterization Protocol and data tools. The database includes data from more than 705,000 patients, collected in more than 60 countries and 1,500 centres worldwide. Patient data are available from acute hospital admissions with COVID-19 and outpatient follow-ups. The data include signs and symptoms, pre-existing comorbidities, vital signs, chronic and acute treatments, complications, dates of hospitalization and discharge, mortality, viral strains, vaccination status, and other data. Here, we present the dataset characteristics, explain its architecture and how to gain access, and provide tools to facilitate its use.

Early short course of neuromuscular blocking agents in patients with COVID-19 ARDS: a propensity score analysis.

BACKGROUND: The role of neuromuscular blocking agents (NMBAs) in coronavirus disease 2019 (COVID-19) acute respiratory distress syndrome (ARDS) is not fully elucidated. Therefore, we aimed to investigate in COVID-19 patients with moderate-to-severe ARDS the impact of early use of NMBAs on 90-day mortality, through propensity score (PS) matching analysis. METHODS: We analyzed a convenience sample of patients with COVID-19 and moderate-to-severe ARDS, admitted to 244 intensive care units within the COVID-19 Critical Care Consortium, from February 1, 2020, through October 31, 2021. Patients undergoing at least 2 days and up to 3 consecutive days of NMBAs (NMBA treatment), within 48 h from commencement of IMV were compared with subjects who did not receive NMBAs or only upon commencement of IMV (control). The primary objective in the PS-matched cohort was comparison between groups in 90-day in-hospital mortality, assessed through Cox proportional hazard modeling. Secondary objectives were comparisons in the numbers of ventilator-free days (VFD) between day 1 and day 28 and between day 1 and 90 through competing risk regression. RESULTS: Data from 1953 patients were included. After propensity score matching, 210 cases from each group were well matched. In the PS-matched cohort, mean (± SD) age was 60.3 ± 13.2 years and 296 (70.5%) were male and the most common comorbidities were hypertension (56.9%), obesity (41.1%), and diabetes (30.0%). The unadjusted hazard ratio (HR) for death at 90 days in the NMBA treatment vs control group was 1.12 (95% CI 0.79, 1.59, p = 0.534). After adjustment for smoking habit and critical therapeutic covariates, the HR was 1.07 (95% CI 0.72, 1.61, p = 0.729). At 28 days, VFD were 16 (IQR 0-25) and 25 (IQR 7-26) in the NMBA treatment and control groups, respectively (sub-hazard ratio 0.82, 95% CI 0.67, 1.00, p = 0.055). At 90 days, VFD were 77 (IQR 0-87) and 87 (IQR 0-88) (sub-hazard ratio 0.86 (95% CI 0.69, 1.07; p = 0.177). CONCLUSIONS: In patients with COVID-19 and moderate-to-severe ARDS, short course of NMBA treatment, applied early, did not significantly improve 90-day mortality and VFD. In the absence of definitive data from clinical trials, NMBAs should be indicated cautiously in this setting.

Use of an extended KDIGO definition to diagnose acute kidney injury in patients with COVID-19: A multinational study using the ISARIC-WHO clinical characterisation protocol.

BACKGROUND: Acute kidney injury (AKI) is one of the most common and significant problems in patients with Coronavirus Disease 2019 (COVID-19). However, little is known about the incidence and impact of AKI occurring in the community or early in the hospital admission. The traditional Kidney Disease Improving Global Outcomes (KDIGO) definition can fail to identify patients for whom hospitalisation coincides with recovery of AKI as manifested by a decrease in serum creatinine (sCr). We hypothesised that an extended KDIGO (eKDIGO) definition, adapted from the International Society of Nephrology (ISN) 0by25 studies, would identify more cases of AKI in patients with COVID-19 and that these may correspond to community-acquired AKI (CA-AKI) with similarly poor outcomes as previously reported in this population. METHODS AND FINDINGS: All individuals recruited using the International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC)-World Health Organization (WHO) Clinical Characterisation Protocol (CCP) and admitted to 1,609 hospitals in 54 countries with Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) infection from February 15, 2020 to February 1, 2021 were included in the study. Data were collected and analysed for the duration of a patient's admission. Incidence, staging, and timing of AKI were evaluated using a traditional and eKDIGO definition, which incorporated a commensurate decrease in sCr. Patients within eKDIGO diagnosed with AKI by a decrease in sCr were labelled as deKDIGO. Clinical characteristics and outcomes-intensive care unit (ICU) admission, invasive mechanical ventilation, and in-hospital death-were compared for all 3 groups of patients. The relationship between eKDIGO AKI and in-hospital death was assessed using survival curves and logistic regression, adjusting for disease severity and AKI susceptibility. A total of 75,670 patients were included in the final analysis cohort. Median length of admission was 12 days (interquartile range [IQR] 7, 20). There were twice as many patients with AKI identified by eKDIGO than KDIGO (31.7% versus 16.8%). Those in the eKDIGO group had a greater proportion of stage 1 AKI (58% versus 36% in KDIGO patients). Peak AKI occurred early in the admission more frequently among eKDIGO than KDIGO patients. Compared to those without AKI, patients in the eKDIGO group had worse renal function on admission, more in-hospital complications, higher rates of ICU admission (54% versus 23%) invasive ventilation (45% versus 15%), and increased mortality (38% versus 19%). Patients in the eKDIGO group had a higher risk of in-hospital death than those without AKI (adjusted odds ratio: 1.78, 95% confidence interval: 1.71 to 1.80, p-value < 0.001). Mortality and rate of ICU admission were lower among deKDIGO than KDIGO patients (25% versus 50% death and 35% versus 70% ICU admission) but significantly higher when compared to patients with no AKI (25% versus 19% death and 35% versus 23% ICU admission) (all p-values <5 × 10-5). Limitations include ad hoc sCr sampling, exclusion of patients with less than two sCr measurements, and limited availability of sCr measurements prior to initiation of acute dialysis. CONCLUSIONS: An extended KDIGO definition of AKI resulted in a significantly higher detection rate in this population. These additional cases of AKI occurred early in the hospital admission and were associated with worse outcomes compared to patients without AKI.

Toward a Learning Health Care System: A Systematic Review and Evidence-Based Conceptual Framework for Implementation of Clinical Analytics in a Digital Hospital.

OBJECTIVE: A learning health care system (LHS) uses routinely collected data to continuously monitor and improve health care outcomes. Little is reported on the challenges and methods used to implement the analytics underpinning an LHS. Our aim was to systematically review the literature for reports of real-time clinical analytics implementation in digital hospitals and to use these findings to synthesize a conceptual framework for LHS implementation. METHODS: Embase, PubMed, and Web of Science databases were searched for clinical analytics derived from electronic health records in adult inpatient and emergency department settings between 2015 and 2021. Evidence was coded from the final study selection that related to (1) dashboard implementation challenges, (2) methods to overcome implementation challenges, and (3) dashboard assessment and impact. The evidences obtained, together with evidence extracted from relevant prior reviews, were mapped to an existing digital health transformation model to derive a conceptual framework for LHS analytics implementation. RESULTS: A total of 238 candidate articles were reviewed and 14 met inclusion criteria. From the selected studies, we extracted 37 implementation challenges and 64 methods employed to overcome such challenges. We identified common approaches for evaluating the implementation of clinical dashboards. Six studies assessed clinical process outcomes and only four studies evaluated patient health outcomes. A conceptual framework for implementing the analytics of an LHS was developed. CONCLUSION: Health care organizations face diverse challenges when trying to implement real-time data analytics. These challenges have shifted over the past decade. While prior reviews identified fundamental information problems, such as data size and complexity, our review uncovered more postpilot challenges, such as supporting diverse users, workflows, and user-interface screens. Our review identified practical methods to overcome these challenges which have been incorporated into a conceptual framework. It is hoped this framework will support health care organizations deploying near-real-time clinical dashboards and progress toward an LHS.

Machine learning models for diabetes management in acute care using electronic medical records: A systematic review.

BACKGROUND: Machine learning (ML) is a subset of Artificial Intelligence (AI) that is used to predict and potentially prevent adverse patient outcomes. There is increasing interest in the application of these models in digital hospitals to improve clinical decision-making and chronic disease management, particularly for patients with diabetes. The potential of ML models using electronic medical records (EMR) to improve the clinical care of hospitalised patients with diabetes is currently unknown. OBJECTIVE: The aim was to systematically identify and critically review the published literature examining the development and validation of ML models using EMR data for improving the care of hospitalised adult patients with diabetes. METHODS: The Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) guidelines were followed. Four databases were searched (Embase, PubMed, IEEE and Web of Science) for studies published between January 2010 to January 2022. The reference lists of the eligible articles were manually searched. Articles that examined adults and both developed and validated ML models using EMR data were included. Studies conducted in primary care and community care settings were excluded. Studies were independently screened and data was extracted using Covidence® systematic review software. For data extraction and critical appraisal, the Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) was followed. Risk of bias was assessed using the Prediction model Risk Of Bias Assessment Tool (PROBAST). Quality of reporting was assessed by adherence to the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) guideline. The IJMEDI checklist was followed to assess quality of ML models and the reproducibility of their outcomes. The external validation methodology of the studies was appraised. RESULTS: Of the 1317 studies screened, twelve met inclusion criteria. Eight studies developed ML models to predict disglycaemic episodes for hospitalized patients with diabetes, one study developed a ML model to predict total insulin dosage, two studies predicted risk of readmission, and one study improved the prediction of hospital readmission for inpatients with diabetes. All included studies were heterogeneous with regard to ML types, cohort, input predictors, sample size, performance and validation metrics and clinical outcomes. Two studies adhered to the TRIPOD guideline. The methodological reporting of all the studies was evaluated to be at high risk of bias. The quality of ML models in all studies was assessed as poor. Robust external validation was not performed on any of the studies. No models were implemented or evaluated in routine clinical care. CONCLUSIONS: This review identified a limited number of ML models which were developed to improve inpatient management of diabetes. No ML models were implemented in real hospital settings. Future research needs to enhance the development, reporting and validation steps to enable ML models for integration into routine clinical care.

Clinical characteristics, risk factors and outcomes in patients with severe COVID-19 registered in the International Severe Acute Respiratory and Emerging Infection Consortium WHO clinical characterisation protocol: a prospective, multinational, multicentre, observational study.

Due to the large number of patients with severe coronavirus disease 2019 (COVID-19), many were treated outside the traditional walls of the intensive care unit (ICU), and in many cases, by personnel who were not trained in critical care. The clinical characteristics and the relative impact of caring for severe COVID-19 patients outside the ICU is unknown. This was a multinational, multicentre, prospective cohort study embedded in the International Severe Acute Respiratory and Emerging Infection Consortium World Health Organization COVID-19 platform. Severe COVID-19 patients were identified as those admitted to an ICU and/or those treated with one of the following treatments: invasive or noninvasive mechanical ventilation, high-flow nasal cannula, inotropes or vasopressors. A logistic generalised additive model was used to compare clinical outcomes among patients admitted or not to the ICU. A total of 40 440 patients from 43 countries and six continents were included in this analysis. Severe COVID-19 patients were frequently male (62.9%), older adults (median (interquartile range (IQR), 67 (55-78) years), and with at least one comorbidity (63.2%). The overall median (IQR) length of hospital stay was 10 (5-19) days and was longer in patients admitted to an ICU than in those who were cared for outside the ICU (12 (6-23) days versus 8 (4-15) days, p<0.0001). The 28-day fatality ratio was lower in ICU-admitted patients (30.7% (5797 out of 18 831) versus 39.0% (7532 out of 19 295), p<0.0001). Patients admitted to an ICU had a significantly lower probability of death than those who were not (adjusted OR 0.70, 95% CI 0.65-0.75; p<0.0001). Patients with severe COVID-19 admitted to an ICU had significantly lower 28-day fatality ratio than those cared for outside an ICU.

Assessment of 28-Day In-Hospital Mortality in Mechanically Ventilated Patients With Coronavirus Disease 2019: An International Cohort Study.

Factors associated with mortality in coronavirus disease 2019 patients on invasive mechanical ventilation are still not fully elucidated. OBJECTIVES: To identify patient-level parameters, readily available at the bedside, associated with the risk of in-hospital mortality within 28 days from commencement of invasive mechanical ventilation or coronavirus disease 2019. DESIGN SETTING AND PARTICIPANTS: Prospective observational cohort study by the global Coronavirus Disease 2019 Critical Care Consortium. Patients with laboratory-confirmed coronavirus disease 2019 requiring invasive mechanical ventilation from February 2, 2020, to May 15, 2021. MAIN OUTCOMES AND MEASURES: Patient characteristics and clinical data were assessed upon ICU admission, the commencement of invasive mechanical ventilation and for 28 days thereafter. We primarily aimed to identify time-independent and time-dependent risk factors for 28-day invasive mechanical ventilation mortality. RESULTS: One-thousand five-hundred eighty-seven patients were included in the survival analysis; 588 patients died in hospital within 28 days of commencing invasive mechanical ventilation (37%). Cox-regression analysis identified associations between the hazard of 28-day invasive mechanical ventilation mortality with age (hazard ratio, 1.26 per 10-yr increase in age; 95% CI, 1.16-1.37; p < 0.001), positive end-expiratory pressure upon commencement of invasive mechanical ventilation (hazard ratio, 0.81 per 5 cm H2O increase; 95% CI, 0.67-0.97; p = 0.02). Time-dependent parameters associated with 28-day invasive mechanical ventilation mortality were serum creatinine (hazard ratio, 1.28 per doubling; 95% CI, 1.15-1.41; p < 0.001), lactate (hazard ratio, 1.22 per doubling; 95% CI, 1.11-1.34; p < 0.001), Paco2 (hazard ratio, 1.63 per doubling; 95% CI, 1.19-2.25; p < 0.001), pH (hazard ratio, 0.89 per 0.1 increase; 95% CI, 0.8-14; p = 0.041), Pao2/Fio2 (hazard ratio, 0.58 per doubling; 95% CI, 0.52-0.66; p < 0.001), and mean arterial pressure (hazard ratio, 0.92 per 10 mm Hg increase; 95% CI, 0.88-0.97; p = 0.003). CONCLUSIONS AND RELEVANCE: This international study suggests that in patients with coronavirus disease 2019 on invasive mechanical ventilation, older age and clinically relevant variables monitored at baseline or sequentially during the course of invasive mechanical ventilation are associated with 28-day invasive mechanical ventilation mortality hazard. Further investigation is warranted to validate any causative roles these parameters might play in influencing clinical outcomes.

An appraisal of respiratory system compliance in mechanically ventilated covid-19 patients.

BACKGROUND: Heterogeneous respiratory system static compliance (CRS) values and levels of hypoxemia in patients with novel coronavirus disease (COVID-19) requiring mechanical ventilation have been reported in previous small-case series or studies conducted at a national level. METHODS: We designed a retrospective observational cohort study with rapid data gathering from the international COVID-19 Critical Care Consortium study to comprehensively describe CRS-calculated as: tidal volume/[airway plateau pressure-positive end-expiratory pressure (PEEP)]-and its association with ventilatory management and outcomes of COVID-19 patients on mechanical ventilation (MV), admitted to intensive care units (ICU) worldwide. RESULTS: We studied 745 patients from 22 countries, who required admission to the ICU and MV from January 14 to December 31, 2020, and presented at least one value of CRS within the first seven days of MV. Median (IQR) age was 62 (52-71), patients were predominantly males (68%) and from Europe/North and South America (88%). CRS, within 48 h from endotracheal intubation, was available in 649 patients and was neither associated with the duration from onset of symptoms to commencement of MV (p = 0.417) nor with PaO2/FiO2 (p = 0.100). Females presented lower CRS than males (95% CI of CRS difference between females-males: - 11.8 to - 7.4 mL/cmH2O p < 0.001), and although females presented higher body mass index (BMI), association of BMI with CRS was marginal (p = 0.139). Ventilatory management varied across CRS range, resulting in a significant association between CRS and driving pressure (estimated decrease - 0.31 cmH2O/L per mL/cmH20 of CRS, 95% CI - 0.48 to - 0.14, p < 0.001). Overall, 28-day ICU mortality, accounting for the competing risk of being discharged within the period, was 35.6% (SE 1.7). Cox proportional hazard analysis demonstrated that CRS (+ 10 mL/cm H2O) was only associated with being discharge from the ICU within 28 days (HR 1.14, 95% CI 1.02-1.28, p = 0.018). CONCLUSIONS: This multicentre report provides a comprehensive account of CRS in COVID-19 patients on MV. CRS measured within 48 h from commencement of MV has marginal predictive value for 28-day mortality, but was associated with being discharged from ICU within the same period. Trial documentation: Available at https://www.covid-critical.com/study . TRIAL REGISTRATION: ACTRN12620000421932.

International Society of Nephrology Global Kidney Health Atlas: structures, organization, and services for the management of kidney failure in Latin America.

Latin America is a region with a widely variable socioeconomic landscape, facing a surge in noncommunicable diseases, including chronic kidney disease and kidney failure, exposing significant limitations in the delivery of care. Despite region-wide efforts to explore and address these limitations, much uncertainty remains as to the capacity, accessibility, and quality of kidney failure care in Latin America. Through this second iteration of the International Society of Nephrology Global Kidney Health Atlas, we aimed to report on these indicators to provide a comprehensive map of kidney failure care in the region. Survey responses were received from 18 (64.2%) countries, representing 93.8% of the total population in Latin America. The median prevalence and incidence of treated kidney failure in Latin America were 715 and 157 per million population, respectively, the latter being higher than the global median (142 per million population), with Puerto Rico, Mexico, and El Salvador experiencing much of this growing burden. In most countries, public and private systems collectively funded most aspects of kidney replacement therapy (dialysis and transplantation) care, with patients incurring at least 1% to 25% of out-of-pocket costs. In most countries, >90% of dialysis patients able to access kidney replacement therapy received hemodialysis (n = 11; 5 high income and 6 upper-middle income), and only a small minority began with peritoneal dialysis (1%-10% in 67% of countries; n = 12). Few countries had chronic kidney disease registries or targeted detection programs. There is a large variability in the availability, accessibility, and quality of kidney failure care in Latin America, which appears to be subject to individual countries' funding structures, underreliance on cheap kidney replacement therapy, such as peritoneal dialysis, and limited chronic kidney disease surveillance and management initiatives.

Intensive care digital health response to emerging infectious disease outbreaks such as COVID-19.

The COVID-19 pandemic has required intensive care units to rapidly adjust and adapt their existing practices. Although there has a focus on expanding critical care infrastructure, equipment and workforce, plans have not emphasised the need to increase digital capabilities. The objective of this report was to recognise key areas of digital health related to the COVID-19 response. We identified and explored six focus areas relevant to intensive care, including using digital solutions to increase critical care capacity, developing surge capacity within an electronic health record, maintenance and downtime planning, training considerations and the role of data analytics. This article forms the basis of a framework for the intensive care digital health response to COVID-19 and other emerging infectious disease outbreaks.

Design and rationale of the COVID-19 Critical Care Consortium international, multicentre, observational study.

INTRODUCTION: There is a paucity of data that can be used to guide the management of critically ill patients with COVID-19. In response, a research and data-sharing collaborative-The COVID-19 Critical Care Consortium-has been assembled to harness the cumulative experience of intensive care units (ICUs) worldwide. The resulting observational study provides a platform to rapidly disseminate detailed data and insights crucial to improving outcomes. METHODS AND ANALYSIS: This is an international, multicentre, observational study of patients with confirmed or suspected SARS-CoV-2 infection admitted to ICUs. This is an evolving, open-ended study that commenced on 1 January 2020 and currently includes >350 sites in over 48 countries. The study enrols patients at the time of ICU admission and follows them to the time of death, hospital discharge or 28 days post-ICU admission, whichever occurs last. Key data, collected via an electronic case report form devised in collaboration with the International Severe Acute Respiratory and Emerging Infection Consortium/Short Period Incidence Study of Severe Acute Respiratory Illness networks, include: patient demographic data and risk factors, clinical features, severity of illness and respiratory failure, need for non-invasive and/or mechanical ventilation and/or extracorporeal membrane oxygenation and associated complications, as well as data on adjunctive therapies. ETHICS AND DISSEMINATION: Local principal investigators will ensure that the study adheres to all relevant national regulations, and that the necessary approvals are in place before a site may contribute data. In jurisdictions where a waiver of consent is deemed insufficient, prospective, representative or retrospective consent will be obtained, as appropriate. A web-based dashboard has been developed to provide relevant data and descriptive statistics to international collaborators in real-time. It is anticipated that, following study completion, all de-identified data will be made open access. TRIAL REGISTRATION NUMBER: ACTRN12620000421932 (http://anzctr.org.au/ACTRN12620000421932.aspx).

Classical and Quantum Causal Interventions.

Characterising causal structure is an activity that is ubiquitous across the sciences. Causal models are representational devices that can be used as oracles for future interventions, to predict how values of some variables will change in response to interventions on others. Recent work has generalised concepts from this field to situations involving quantum systems, resulting in a new notion of quantum causal structure. A key concept in both the classical and quantum context is that of an intervention. Interventions are the controlled operations required to identify causal structure and ultimately the feature that endows causal models with empirical meaning. Although interventions are a crucial feature of both the classical and quantum causal modelling frameworks, to date there has been no discussion of their physical basis. In this paper, we consider interventions from a physical perspective and show that, in both the classical and quantum case, they are constrained by the thermodynamics of measurement and feedback in open systems. We demonstrate that the perfect "atomic" or "surgical" interventions characterised by Pearl's famous do-calculus are physically impossible, and this is the case for both classical and quantum systems.