Using the ROX Index to Predict Treatment Outcome for High-Flow Nasal Cannula and/or Noninvasive Ventilation in Patients With COPD Exacerbations.

BACKGROUND: The ratio of oxygen saturation index (ROX index; or SpO2 /FIO2 /breathing frequency) has been shown to predict risk of intubation after high-flow nasal cannula (HFNC) support among adults with acute hypoxemic respiratory failure primarily due to pneumonia. However, its predictive value for other subtypes of respiratory failure is unknown. This study investigated whether the ROX index predicts liberation from HFNC or noninvasive ventilation (NIV), intubation with mechanical ventilation, or death in adults admitted for respiratory failure due to an exacerbation of COPD. METHODS: We performed a retrospective study of 260 adults hospitalized with a COPD exacerbation and treated with HFNC and/or NIV (continuous or bi-level). ROX index scores were collected at treatment initiation and predefined time intervals throughout HFNC and/or NIV treatment or until the subject was intubated or died. A ROX index score of ≥ 4.88 was applied to the cohort to determine if the same score would perform similarly in this different cohort. Accuracy of the ROX index was determined by calculating the area under the receiver operator curve. RESULTS: A total of 47 subjects (18%) required invasive mechanical ventilation or died while on HFNC/NIV. The ROX index at treatment initiation, 1 h, and 6 h demonstrated the best prediction accuracy for avoidance of invasive mechanical ventilation or death (area under the receiver operator curve 0.73 [95% CI 0.66-0.80], 0.72 [95% CI 0.65-0.79], and 0.72 [95% CI 0.63-0.82], respectively). The optimal cutoff value for sensitivity (Sn) and specificity (Sp) was a ROX index score > 6.88 (sensitivity 62%, specificity 57%). CONCLUSIONS: The ROX index applied to adults with COPD exacerbations treated with HFNC and/or NIV required higher scores to achieve similar prediction of low risk of treatment failure when compared to subjects with hypoxemic respiratory failure/pneumonia. ROX scores < 4.88 did not accurately predict intubation or death.

Early Mobility Index and Patient Outcomes: A Retrospective Study in Multiple Intensive Care Units.

BACKGROUND: Early mobility interventions in intensive care units (ICUs) are safe and improve outcomes in subsets of critically ill adults. However, implementation varies, and the optimal mobility dose remains unclear. OBJECTIVE: To test for associations between daily dose of out-of-bed mobility and patient outcomes in different ICUs. METHODS: In this retrospective cohort study of electronic records from 7 adult ICUs in an academic quarternary hospital, multivariable linear regression was used to examine the effects of out-of-bed events per mobility-eligible day on mechanical ventilation duration and length of ICU and hospital stays. RESULTS: In total, 8609 adults hospitalized in ICUs from 2015 through 2018 were included. Patients were mobilized out of bed on 46.5% of ICU days and were eligible for mobility interventions on a median (IQR) of 2.0 (1-3) of 2.7 (2-9) ICU days. Median (IQR) out-of-bed events per mobility-eligible day were 0.5 (0-1.2) among all patients. For every unit increase in out-of-bed events per mobility-eligible day before extubation, mechanical ventilation duration decreased by 10% (adjusted coefficient [95% CI], -0.10 [-0.18 to -0.01]). Daily mobility increased ICU stays by 4% (adjusted coefficient [95% CI], 0.04 [0.03-0.06]) and decreased hospital stays by 5% (adjusted coefficient [95% CI], -0.05 [-0.07 to -0.03]). Effect sizes differed among ICUs. CONCLUSIONS: More daily out-of-bed mobility for ICU patients was associated with shorter mechanical ventilation duration and hospital stays, suggesting a dose-response relationship between daily mobility and patient outcomes. However, relationships differed across ICU subpopulations.

MAN VERSUS MACHINE: PROVIDER DIRECTED VERSUS PRECISION AUTOMATED CRITICAL CARE MANAGEMENT IN A PORCINE MODEL OF DISTRIBUTIVE SHOCK.

ABSTRACT: Background: Critical care management of shock is a labor-intensive process. Precision Automated Critical Care Management (PACC-MAN) is an automated closed-loop system incorporating physiologic and hemodynamic inputs to deliver interventions while avoiding excessive fluid or vasopressor administration. To understand PACC-MAN efficacy, we compared PACC-MAN to provider-directed management (PDM). We hypothesized that PACC-MAN would achieve equivalent resuscitation outcomes to PDM while maintaining normotension with lower fluid and vasopressor requirements. Methods : Twelve swine underwent 30% controlled hemorrhage over 30 min, followed by 45 min of aortic occlusion to generate a vasoplegic shock state, transfusion to euvolemia, and randomization to PACC-MAN or PDM for 4.25 h. Primary outcomes were total crystalloid volume, vasopressor administration, total time spent at hypotension (mean arterial blood pressure <60 mm Hg), and total number of interventions. Results : Weight-based fluid volumes were similar between PACC-MAN and PDM; median and IQR are reported (73.1 mL/kg [59.0-78.7] vs. 87.1 mL/kg [79.4-91.8], P = 0.07). There was no statistical difference in cumulative norepinephrine (PACC-MAN: 33.4 µg/kg [27.1-44.6] vs. PDM: 7.5 [3.3-24.2] µg/kg, P = 0.09). The median percentage of time spent at hypotension was equivalent (PACC-MAN: 6.2% [3.6-7.4] and PDM: 3.1% [1.3-6.6], P = 0.23). Urine outputs were similar between PACC-MAN and PDM (14.0 mL/kg vs. 21.5 mL/kg, P = 0.13). Conclusion : Automated resuscitation achieves equivalent resuscitation outcomes to direct human intervention in this shock model. This study provides the first translational experience with the PACC-MAN system versus PDM.

Improving the precision of shock resuscitation by predicting fluid responsiveness with machine learning and arterial blood pressure waveform data.

Fluid bolus therapy (FBT) is fundamental to the management of circulatory shock in critical care but balancing the benefits and toxicities of FBT has proven challenging in individual patients. Improved predictors of the hemodynamic response to a fluid bolus, commonly referred to as a fluid challenge, are needed to limit non-beneficial fluid administration and to enable automated clinical decision support and patient-specific precision critical care management. In this study we retrospectively analyzed data from 394 fluid boluses from 58 pigs subjected to either hemorrhagic or distributive shock. All animals had continuous blood pressure and cardiac output monitored throughout the study. Using this data, we developed a machine learning (ML) model to predict the hemodynamic response to a fluid challenge using only arterial blood pressure waveform data as the input. A Random Forest binary classifier referred to as the ML fluid responsiveness algorithm (MLFRA) was trained to detect fluid responsiveness (FR), defined as a ≥ 15% change in cardiac stroke volume after a fluid challenge. We then compared its performance to pulse pressure variation, a commonly used metric of FR. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC), confusion matrix metrics, and calibration curves plotting predicted probabilities against observed outcomes. Across multiple train/test splits and feature selection methods designed to assess performance in the setting of small sample size conditions typical of large animal experiments, the MLFRA achieved an average AUROC, recall (sensitivity), specificity, and precision of 0.82, 0.86, 0.62. and 0.76, respectively. In the same datasets, pulse pressure variation had an AUROC, recall, specificity, and precision of 0.73, 0.91, 0.49, and 0.71, respectively. The MLFRA was generally well-calibrated across its range of predicted probabilities and appeared to perform equally well across physiologic conditions. These results suggest that ML, using only inputs from arterial blood pressure monitoring, may substantially improve the accuracy of predicting FR compared to the use of pulse pressure variation. If generalizable, these methods may enable more effective, automated precision management of critically ill patients with circulatory shock.

Precision Automated Critical Care Management: Closed-loop critical care for the treatment of distributive shock in a swine model of ischemia-reperfusion.

BACKGROUND: Goal-directed blood pressure management in the intensive care unit can improve trauma outcomes but is labor-intensive. Automated critical care systems can deliver scaled interventions to avoid excessive fluid or vasopressor administration. We compared a first-generation automated drug and fluid delivery platform, Precision Automated Critical Care Management (PACC-MAN), to a more refined algorithm, incorporating additional physiologic inputs and therapeutics. We hypothesized that the enhanced algorithm would achieve equivalent resuscitation endpoints with less crystalloid utilization in the setting of distributive shock. METHODS: Twelve swine underwent 30% hemorrhage and 30 minutes of aortic occlusion to induce an ischemia-reperfusion injury and distributive shock state. Next, animals were transfused to euvolemia and randomized into a standardized critical care (SCC) of PACC-MAN or an enhanced version (SCC+) for 4.25 hours. SCC+ incorporated lactate and urine output to assess global response to resuscitation and added vasopressin as an adjunct to norepinephrine at certain thresholds. Primary and secondary outcomes were decreased crystalloid administration and time at goal blood pressure, respectively. RESULTS: Weight-based fluid bolus volume was lower in SCC+ compared with SCC (26.9 mL/kg vs. 67.5 mL/kg, p = 0.02). Cumulative norepinephrine dose required was not significantly different (SCC+: 26.9 µg/kg vs. SCC: 13.76 µg/kg, p = 0.24). Three of 6 animals (50%) in SCC+ triggered vasopressin as an adjunct. Percent time spent between 60 mm Hg and 70 mm Hg, terminal creatinine and lactate, and weight-adjusted cumulative urine output were equivalent. CONCLUSION: Refinement of the PACC-MAN algorithm decreased crystalloid administration without sacrificing time in normotension, reducing urine output, increasing vasopressor support, or elevating biomarkers of organ damage. Iterative improvements in automated critical care systems to achieve target hemodynamics in a distributive-shock model are feasible.

Work of Breathing During Proportional Assist Ventilation as a Predictor of Extubation Failure.

BACKGROUND: Despite decades of research on predictors of extubation success, use of ventilatory support after extubation is common and 10-20% of patients require re-intubation. Proportional assist ventilation (PAV) mode automatically calculates estimated total work of breathing (total WOB). Here, we assessed the performance of total WOB to predict extubation failure in invasively ventilated subjects. METHODS: This prospective observational study was conducted in 6 adult ICUs at an academic medical center. We enrolled intubated subjects who successfully completed a spontaneous breathing trial, had a rapid shallow breathing index < 105 breaths/min/L, and were deemed ready for extubation by the primary team. Total WOB values were recorded at the end of a 30-min PAV trial. Extubation failure was defined as any respiratory support and/or re-intubation within 72 h of extubation. We compared total WOB scores between groups and performance of total WOB for predicting extubation failure with receiver operating characteristic curves. RESULTS: Of 61 subjects enrolled, 9.8% (n = 6) required re-intubation, and 50.8% (n = 31) required any respiratory support within 72 h of extubation. Median total WOB at 30 min on PAV was 0.9 J/L (interquartile range 0.7-1.3 J/L). Total WOB was significantly different between subjects who failed or were successfully extubated (median 1.1 J/L vs 0.7 J/L, P = .004). The area under the curve was 0.71 [95% CI 0.58-0.85] for predicting any requirement of respiratory support and 0.85 [95% CI 0.69-1.00] for predicting re-intubation alone within 72 h of extubation. Total WOB cutoff values maximizing sensitivity and specificity equally were 1.0 J/L for any respiratory support (positive predictive value [PPV] 70.0%, negative predictive value [NPV] 67.7%) and 1.3 J/L for re-intubation (PPV 26.3%, NPV 97.6%). CONCLUSIONS: The discriminative performance of a PAV-derived total WOB value to predict extubation failure was good, indicating total WOB may represent an adjunctive tool for assessing extubation readiness. However, these results should be interpreted as preliminary, with specific thresholds of PAV-derived total WOB requiring further investigation in a large multi-center study.

A Least Absolute Shrinkage and Selection Operator-Derived Predictive Model for Postoperative Respiratory Failure in a Heterogeneous Adult Elective Surgery Patient Population.

BACKGROUND: Postoperative respiratory failure (PRF) is associated with increased hospital charges and worse patient outcomes. Reliable prediction models can help to guide postoperative planning to optimize care, to guide resource allocation, and to foster shared decision-making with patients. RESEARCH QUESTION: Can a predictive model be developed to accurately identify patients at high risk of PRF? STUDY DESIGN AND METHODS: In this single-site proof-of-concept study, we used structured query language to extract, transform, and load electronic health record data from 23,999 consecutive adult patients admitted for elective surgery (2014-2021). Our primary outcome was PRF, defined as mechanical ventilation after surgery of > 48 h. Predictors of interest included demographics, comorbidities, and intraoperative factors. We used logistic regression to build a predictive model and the least absolute shrinkage and selection operator procedure to select variables and to estimate model coefficients. We evaluated model performance using optimism-corrected area under the receiver operating curve and area under the precision-recall curve and calculated sensitivity, specificity, positive and negative predictive values, and Brier scores. RESULTS: Two hundred twenty-five patients (0.94%) demonstrated PRF. The 18-variable predictive model included: operations on the cardiovascular, nervous, digestive, urinary, or musculoskeletal system; surgical specialty orthopedic (nonspine); Medicare or Medicaid (as the primary payer); race unknown; American Society of Anesthesiologists class ≥ III; BMI of 30 to 34.9 kg/m2; anesthesia duration (per hour); net fluid at end of the operation (per liter); median intraoperative FIO2, end title CO2, heart rate, and tidal volume; and intraoperative vasopressor medications. The optimism-corrected area under the receiver operating curve was 0.835 (95% CI,0.808-0.862) and the area under the precision-recall curve was 0.156 (95% CI, 0.105-0.203). INTERPRETATION: This single-center proof-of-concept study demonstrated that a structured query language extract, transform, and load process, based on readily available patient and intraoperative variables, can be used to develop a prediction model for PRF. This PRF prediction model is scalable for multicenter research. Clinical applications include decision support to guide postoperative level of care admission and treatment decisions.

Physiowise: A Physics-aware Approach to Dicrotic Notch Identification.

Dicrotic Notch (DN), one of the most significant and indicative features of the arterial blood pressure (ABP) waveform, becomes less pronounced and thus harder to identify as a matter of aging and pathological vascular stiffness. Generalizable and automatic DN identification for such edge cases is even more challenging in the presence of unexpected ABP waveform deformations that happen due to internal and external noise sources or pathological conditions that cause hemodynamic instability. We propose a physics-aware approach, named Physiowise (PW), that first employs a cardiovascular model to augment the original ABP waveform and reduce unexpected deformations, then apply a set of predefined rules on the augmented signal to find DN locations. We have tested the proposed method on in-vivo data gathered from 14 pigs under hemorrhage and sepsis study. Our result indicates 52% overall mean error improvement with 16% higher detection accuracy within the lowest permitted error range of 30ms. An additional hybrid methodology is also proposed to allow combining augmentation with any application-specific user-defined rule set.

Epigenetics may characterize asymptomatic COVID-19 infection.

RT-PCR is the foremost clinical test for diagnosis of COVID-19. Unfortunately, PCR-based testing has limitations and may not result in a positive test early in the course of infection before symptoms develop. Enveloped RNA viruses, such as coronaviruses, alter peripheral blood methylation and DNA methylation signatures may characterize asymptomatic versus symptomatic infection. We used Illumina's Infinium MethylationEPIC BeadChip array to profile peripheral blood samples from 164 patients who tested positive for SARS-CoV-2 by RT-PCR, of whom 8 had no symptoms. Epigenome-wide association analysis identified 10 methylation sites associated with infection and a quantile-quantile plot showed little inflation. These preliminary results suggest that differences in methylation patterns may distinguish asymptomatic from symptomatic infection.

Real-world impact of vaccination on coronavirus disease 2019 (COVID-19) incidence in healthcare personnel at an academic medical center.

OBJECTIVE: Coronavirus disease 2019 (COVID-19) vaccination effectiveness in healthcare personnel (HCP) has been established. However, questions remain regarding its performance in high-risk healthcare occupations and work locations. We describe the effect of a COVID-19 HCP vaccination campaign on SARS-CoV-2 infection by timing of vaccination, job type, and work location. METHODS: We conducted a retrospective review of COVID-19 vaccination acceptance, incidence of postvaccination COVID-19, hospitalization, and mortality among 16,156 faculty, students, and staff at a large academic medical center. Data were collected 8 weeks prior to the start of phase 1a vaccination of frontline employees and ended 11 weeks after campaign onset. RESULTS: The COVID-19 incidence rate among HCP at our institution decreased from 3.2% during the 8 weeks prior to the start of vaccinations to 0.38% by 4 weeks after campaign initiation. COVID-19 risk was reduced among individuals who received a single vaccination (hazard ratio [HR], 0.52; 95% confidence interval [CI], 0.40-0.68; P < .0001) and was further reduced with 2 doses of vaccine (HR, 0.17; 95% CI, 0.09-0.32; P < .0001). By 2 weeks after the second dose, the observed case positivity rate was 0.04%. Among phase 1a HCP, we observed a lower risk of COVID-19 among physicians and a trend toward higher risk for respiratory therapists independent of vaccination status. Rates of infection were similar in a subgroup of nurses when examined by work location. CONCLUSIONS: Our findings show the real-world effectiveness of COVID-19 vaccination in HCP. Despite these encouraging results, unvaccinated HCP remain at an elevated risk of infection, highlighting the need for targeted outreach to combat vaccine hesitancy.

Dicrotic Notch Identification: a Generalizable Hybrid Approach under Arterial Blood Pressure (ABP) Curve Deformations.

Dicrotic Notch (DN) is a distinctive and clinically significant feature of the arterial blood pressure curve. Its automatic identification has been the focus of many kinds of research using either model-based or rule-based methodologies. However, since DN morphology is quite variant following the patient-specific underlying physiological and pathological conditions, its automatic identification with these methods is challenging. This work proposes a hybrid approach that employs both model-based and rule-based approaches to enhance DN detection's generalizability. We have tested our approach on ABP data gathered from 14 pigs. Our result strongly indicates 36% overall mean error improvement with maximum 52% and -11% accuracy enhancement and degradation in extreme cases.

Use of Machine Learning to Screen for Acute Respiratory Distress Syndrome Using Raw Ventilator Waveform Data.

To develop and characterize a machine learning algorithm to discriminate acute respiratory distress syndrome from other causes of respiratory failure using only ventilator waveform data. DESIGN: Retrospective, observational cohort study. SETTING: Academic medical center ICU. PATIENTS: Adults admitted to the ICU requiring invasive mechanical ventilation, including 50 patients with acute respiratory distress syndrome and 50 patients with primary indications for mechanical ventilation other than hypoxemic respiratory failure. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Pressure and flow time series data from mechanical ventilation during the first 24-hours after meeting acute respiratory distress syndrome criteria (or first 24-hr of mechanical ventilation for non-acute respiratory distress syndrome patients) were processed to extract nine physiologic features. A random forest machine learning algorithm was trained to discriminate between the patients with and without acute respiratory distress syndrome. Model performance was assessed using the area under the receiver operating characteristic curve, sensitivity, specificity, positive predictive value, and negative predictive value. Analyses examined performance when the model was trained using data from the first 24 hours and tested using withheld data from either the first 24 hours (24/24 model) or 6 hours (24/6 model). Area under the receiver operating characteristic curve, sensitivity, specificity, positive predictive value, and negative predictive value were 0.88, 0.90, 0.71, 0.77, and 0.90 (24/24); and 0.89, 0.90, 0.75, 0.83, and 0.83 (24/6). CONCLUSIONS: Use of machine learning and physiologic information derived from raw ventilator waveform data may enable acute respiratory distress syndrome screening at early time points after intubation. This approach, combined with traditional diagnostic criteria, could improve timely acute respiratory distress syndrome recognition and enable automated clinical decision support, especially in settings with limited availability of conventional diagnostic tests and electronic health records.

Standardisation, multi-measure, data quality and trending: A qualitative study on multidisciplinary perspectives to improve intensive care early mobility monitoring.

OBJECTIVE: To explore multi-clinician perspectives on intensive care early mobility, monitoring and to assess the perceived value of technology-generated mobility metrics to provide user feedback to inform research, practice improvement, and technology development. METHODS: We performed a qualitative descriptive study. Three focus groups were conducted with critical care clinicians, including nurses (n = 10), physical therapists (n = 8) and physicians (n = 8) at an academic medical centre that implemented an intensive care early mobility programme in 2012. Qualitative thematic analysis was used to code transcripts and identify overarching themes. FINDINGS: Along with reaffirming the value of performing early mobility interventions, four themes for improving mobility monitoring emerged, including the need for: 1) standardised indicators for documenting mobility; 2) inclusion of both quantitative and qualitative metrics to measure mobility 3) a balance between quantity and quality of data; and 4) trending mobility metrics over time. CONCLUSION: Intensive care mobility monitoring should be standardised and data generated should be high quality, capable of supporting trend analysis, and meaningful. By improving measurement and monitoring of mobility, future researchers can examine the arc of activity that patients in the intensive care unit undergo and develop models to understand factors that influence successful implementation.

Leveraging IoTs and Machine Learning for Patient Diagnosis and Ventilation Management in the Intensive Care Unit.

Future healthcare systems will rely heavily on clinical decision support systems (CDSS) to improve the decision-making processes of clinicians. To explore the design of future CDSS, we developed a research-focused CDSS for the management of patients in the intensive care unit that leverages Internet of Things (IoT) devices capable of collecting streaming physiologic data from ventilators and other medical devices. We then created machine learning (ML) models that could analyze the collected physiologic data to determine if the ventilator was delivering potentially harmful therapy and if a deadly respiratory condition, acute respiratory distress syndrome (ARDS), was present. We also present work to aggregate these models into a mobile application that can provide responsive, real-time alerts of changes in ventilation to providers. As illustrated in the recent COVID-19 pandemic, being able to accurately predict ARDS in newly infected patients can assist in prioritizing care. We show that CDSS may be used to analyze physiologic data for clinical event recognition and automated diagnosis, and we also highlight future research avenues for hospital CDSS.

Quantifying Mobility in the ICU: Comparison of Electronic Health Record Documentation and Accelerometer-Based Sensors to Clinician-Annotated Video.

To compare the accuracy of electronic health record clinician documentation and accelerometer-based sensors with a gold standard dataset derived from clinician-annotated video to quantify early mobility activities in adult ICU patients. DESIGN: Prospective, observational study. SETTING: Medical ICU at an academic hospital. PATIENTS: Adult ICU patients (n = 30) were each continuously monitored over a median of 24.4 hours, yielding 711.5 hours of video, electronic health record, and sensor data. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Electronic health record documentation estimated ambulation (intraclass correlation coefficient, 0.89; 95% CI, 0.78-0.95), sitting out-of-bed (intraclass correlation coefficient, 0.85; 95% CI, 0.72-0.93), and turning events (intraclass correlation coefficient, 0.87; 95% CI, 0.75-0.94) with excellent agreement but underestimated the number of standing, transferring, and pregait activities performed per patient. The accelerometer-based sensor had excellent agreement with video annotation for estimating duration of time spent supine (intraclass correlation coefficient, 0.99; CI, 0.97-0.99) and sitting/standing upright (intraclass correlation coefficient, 0.92; CI, 0.82-0.96) but overestimated ambulation time. CONCLUSIONS: Our results show that electronic health record documentation and sensor-based technologies accurately capture distinct but complimentary metrics for ICU mobility measurement. Innovations in artifact detection, standardization of clinically relevant mobility definitions, and electronic health record documentation enhancements may enable further use of these technologies to drive critical care research and technology leveraged data-driven ICU models of care.

Association Between Peripheral Blood Oxygen Saturation (SpO2)/Fraction of Inspired Oxygen (FiO2) Ratio Time at Risk and Hospital Mortality in Mechanically Ventilated Patients.

INTRODUCTION: Acute respiratory failure requiring mechanical ventilation is a leading cause of mortality in the intensive care unit. Although single peripheral blood oxygen saturation/fraction of inspired oxygen (SpO2/FiO2) ratios of hypoxemia have been evaluated to risk-stratify patients with acute respiratory distress syndrome, the utility of longitudinal SpO2/FiO2 ratios is unknown. OBJECTIVE: To assess time-based SpO2/FiO2 ratios ≤ 150-SpO2/FiO2 time at risk (SF-TAR)-for predicting mortality in mechanically ventilated patients. METHODS: Retrospective, observational cohort study of mechanically ventilated patients at 21 community and 2 academic hospitals. Association between the SF-TAR in the first 24 hours of ventilation and mortality was examined using multivariable logistic regression and compared with the worst recorded isolated partial pressure of arterial oxygen/fraction of inspired oxygen (P/F) ratio. RESULTS: In 28,758 derivation cohort admissions, every 10% increase in SF-TAR was associated with a 24% increase in adjusted odds of hospital mortality (adjusted odds ratio = 1.24; 95% confidence interval [CI] = 1.23-1.26); a similar association was observed in validation cohorts. Discrimination for mortality modestly improved with SF-TAR (area under the receiver operating characteristic curve [AUROC] = 0.81; 95% CI = 0.81-0.82) vs the worst P/F ratio (AUROC = 0.78; 95% CI = 0.78-0.79) and worst SpO2/FiO2 ratio (AUROC = 0.79; 95% CI = 0.79-0.80). The SF-TAR in the first 6 hours offered comparable discrimination for hospital mortality (AUROC = 0.80; 95% CI = 0.79-0.80) to the 24-hour SF-TAR. CONCLUSION: The SF-TAR can identify ventilated patients at increased risk of death, offering modest improvements compared with single SpO2/FiO2 and P/F ratios. This longitudinal, noninvasive, and broadly generalizable tool may have particular utility for early phenotyping and risk stratification using electronic health record data in ventilated patients.

How much do hospitalized adults move? A systematic review and meta-analysis.

AIM: To quantify the type and duration of physical activity performed by hospitalized adults. BACKGROUND: Inactivity is pervasive among hospitalized patients and is associated with increased mortality, functional decline, and cognitive impairment. Objective measurement of activity is necessary to examine associations with clinical outcomes and quantify optimal inpatient mobility interventions. METHODS: We used PRISMA guidelines to search three databases in December 2017 to retrieve original research evaluating activity type and duration among adult acute-care inpatients. We abstracted data on inpatient population, measurement method, monitoring time, activity duration, and study quality. RESULTS: Thirty-eight articles were included in the review and 7 articles were included in the meta-analysis. Study populations included geriatric (n = 5), surgical (n = 5), medical (n = 12), post-stroke (n = 10), psychiatric (n = 2), and critical care inpatients (n = 4). To measure activity, 29% of studies used human observation and 71% used activity monitors. Among inpatient populations, 87-100% of time was spent sitting or lying in-bed. Among medical inpatients monitored over a continuous 24-hour period (n = 7), 70 min per day was spent standing/walking (95% CI 57-83 min). CONCLUSIONS: This review provides a baseline assessment and benchmark of inpatient activity, which can be used to compare inpatient mobility practices. While there is substantial heterogeneity in how researchers measure and define how much inpatients move, there is consistent evidence that patients are mostly inactive and in-bed during hospitalization. Future research is needed to establish standardized methods to accurately and consistently measure inpatient mobility over time.

Improving Mechanical Ventilator Clinical Decision Support Systems with a Machine Learning Classifier for Determining Ventilator Mode.

Clinical decision support systems (CDSS) will play increasing role in improving quality of medical care for critically ill patients. However, due to limitations in current informatics infrastructure, CDSS do not always have complete information on state of supporting physiologic monitoring devices, which can limit input data available to CDSS. This is especially true in use case of mechanical ventilation (MV), where current CDSS have no knowledge of critical ventilation settings, such as ventilation mode. To enable MV CDSS make accurate recommendations related to ventilator mode, we developed a highly performant machine learning model that is able to perform per-breath classification of five of most widely used ventilation modes in USA with average F1-score of 97.52%. We also show how our approach makes methodologic improvements over previous work and is highly robust to missing data caused by software/sensor error.

A 51-Year-Old Man With Unresolved Pulmonary Infiltrates Following Streptococcus pneumoniae Pneumonia.

CASE PRESENTATION: A 51-year-old man presented to the clinic 8 weeks after a 6-day hospital admission for severe multilobar pneumonia caused by Streptococcus pneumoniae. His productive cough resolved after antibiotics, but he reported persistent dyspnea. He recounted a lifelong history of recurrent sinusitis but no previous episodes of pneumonia. The patient denied fever, weight loss, or tobacco, alcohol, or drug use. He worked as an upholstery craftsman with no work-related exposures. He had no bird or exotic animal exposures, and no history of travel outside Sacramento, California, where he lived. Aside from the recently completed 2-week course of levofloxacin, he was not taking any medications.

Dexmedetomidine for Refractory Intracranial Hypertension.

Dexmedetomidine (DEX) is a selective α2 adrenergic agonist that is commonly used for sedation in the intensive care unit (ICU). The role of DEX for adjunctive treatment of refractory intracranial hypertension is poorly defined. The primary objective of this study was to determine the effect of DEX on the need for rescue therapy (ie, hyperosmolar boluses, extraventricular drain [EVD] drainages) for refractory intracranial hypertension. Secondary objectives included the number of intracranial pressure (ICP) excursions, bradycardic, hypotensive, and compromised cerebral perfusion pressure episodes. This retrospective cohort study evaluated patients admitted to the neurosurgical ICU from August 1, 2009, to July 29, 2015, and who received DEX for refractory intracranial hypertension. The objectives were compared between the 2 time periods-before (pre-DEX) and during therapy (DEX). Twenty-three patients with 26 episodes of refractory intracranial hypertension met the inclusion criteria. The number of hyperosmolar boluses was decreased after DEX therapy was initiated. Mannitol boluses required were statistically reduced (1 vs 0.5, P = .03); however, reduction in hypertonic boluses was not statistically significant (1.3 vs 0.9, P = .2). The mean number of EVD drainages per 24 hours was not significantly different between the time periods (15.7 vs 14.0, P = .35). The rate of ICP excursions did not differ between the 2 groups (24.3 vs 22.5, P = .62). When compared to pre-DEX data, there was no difference in the median number of hypotensive (0 vs 0), bradycardic (0 vs 0), or compromised cerebral perfusion pressure episodes (0.5 vs 1.0). Dexmedetomidine may avoid increases in the need for rescue therapy when used as an adjunctive treatment of refractory intracranial hypertension without compromising hemodynamics.