OBJECTIVE:The aim of this study was to develop and externally validate a machine-learning model that retrospectively identifies patients with acute respiratory distress syndrome (acute respiratory distress syndrome [ARDS]) using electronic health record (EHR) data.DESIGN:In this retrospective cohort study, ARDS was identified via physician-adjudication in three cohorts of patients with hypoxemic respiratory failure (training, internal validation, and external validation). Machine-learning models were trained to classify ARDS using vital signs, respiratory support, laboratory data, medications, chest radiology reports, and clinical notes. The best-performing models were assessed and internally and externally validated using the area under receiver-operating curve (AUROC), area under precision-recall curve, integrated calibration index (ICI), sensitivity, specificity, positive predictive value (PPV), and ARDS timing.PATIENTS:Patients with hypoxemic respiratory failure undergoing mechanical ventilation within two distinct health systemsINTERVENTIONS:None.MEASUREMENTS AND MAIN RESULTS:There were 1,845 patients in the training cohort, 556 in the internal validation cohort, and 199 in the external validation cohort. ARDS prevalence was 19%, 17%, and 31%, respectively. Regularized logistic regression models analyzing structured data (EHR model) and structured data and radiology reports (EHR-radiology model) had the best performance. During internal and external validation, the EHR-radiology model had AUROC of 0.91 (95% CI, 0.88-0.93) and 0.88 (95% CI, 0.87-0.93), respectively. Externally, the ICI was 0.13 (95% CI, 0.08-0.18). At a specified model threshold, sensitivity and specificity were 80% (95% CI, 75%-98%), PPV was 64% (95% CI, 58%-71%), and the model identified patients with a median of 2.2 hours (interquartile range 0.2-18.6) after meeting Berlin ARDS criteria.CONCLUSIONS:Machine-learning models analyzing EHR data can retrospectively identify patients with ARDS across different institutions.

BACKGROUND: Traditional approaches for learning anatomy for curvilinear endobronchial ultrasound (EBUS) require learners to mentally visualize structures relative to the position of the bronchoscope. Virtual reality (VR) can show anatomy from the perspective of bronchoscopic tools. RESEARCH QUESTION: Does the use of a VR anatomy trainer for teaching EBUS-associated anatomy improve procedural performance compared with traditional methods? STUDY DESIGN AND METHODS: In this randomized, crossover study design, participants studied EBUS-related anatomy during 2 sequential sessions using a VR trainer and a traditional modality (2-dimensional pictures or a 3-dimensional model). An EBUS simulator was used to test performance at baseline and following each training session. User experience and preferences were evaluated by using a mixed-methods approach of surveys and interviews. Spatial reasoning ability was measured by using the Mental Rotation Test. RESULTS: Sixty-eight fellows and residents at 3 institutions completed the study. All 3 learning methods improved EBUS performance significantly following the first, but not second, learning session. Learners spent more time (1.37 minutes) with VR, but no training method produced a greater improvement. Spatial reasoning ability was associated with improved EBUS performance. This impact was modified by training method: the VR approach leveled the impact of baseline spatial reasoning. The VR approach was preferred by 96% of learners. Qualitative data revealed a positive VR user experience with focused anatomy learning, ease of use, acceptable realism, and tolerance. This novel "inside-looking-out" perspective helped learners understand anatomy from the vantage of procedural tools and to create a mental map, but interpreting ultrasound remained challenging. INTERPRETATION: A VR anatomy trainer was preferred by learners because it provided visualization that aligned best with the procedural perspective. This approach helped learners of all spatial reasoning ability improve their procedural performance. CHEST 2025; 167(5):1440-1450

"A 66-Year-Old Man with Persistent Altered Mental Status and New Rash." Annals of the American Thoracic Society, 20(3), pp. 466–469

Abstract Objectives To evaluate measurement discrepancies by race between pulse oximetry and arterial oxygen saturation (as measured in arterial blood gas) among inpatients not in intensive care. Design Multicenter, retrospective cohort study using electronic medical records from general care medical and surgical inpatients. Setting Veteran Health Administration, a national and racially diverse integrated health system in the United States, from 2013 to 2019. Participants Adult inpatients in general care (medical and surgical), in Veteran Health Administration medical centers. Main outcomes measures Occult hypoxemia (defined as arterial blood oxygen saturation (SaO2) of <88% despite a pulse oximetry (SpO2) reading of ≥92%), and whether rates of occult hypoxemia varied by race and ethnic origin. Results A total of 30 039 pairs of SpO2-SaO2 readings made within 10 minutes of each other were identified during the study. These pairs were predominantly among non-Hispanic white (21 918 (73.0%)) patients; non-Hispanic black patients and Hispanic or Latino patients accounted for 6498 (21.6%) and 1623 (5.4%) pairs in the sample, respectively. Among SpO2 values greater or equal to 92%, unadjusted probabilities of occult hypoxemia were 15.6% (95% confidence interval 15.0% to 16.1%) in white patients, 19.6% (18.6% to 20.6%) in black patients (P<0.001 v white patients, with similar P values in adjusted models), and 16.2% (14.4% to 18.1%) in Hispanic or Latino patients (P=0.53 v white patients, P<0.05 in adjusted models). This result was consistent in SpO2-SaO2 pairs restricted to occur within 5 minutes and 2 minutes. In white patients, an initial SpO2-SaO2 pair with little difference in saturation was associated with a 2.7% (95% confidence interval −0.1% to 5.5%) probability of SaO2 <88% on a later paired SpO2-SaO2 reading showing an SpO2 of 92%, but black patients had a higher probability (12.9% (−3.3% to 29.0%)). Conclusions In general care inpatient settings across the Veterans Health Administration where paired readings of arterial blood gas (SaO2) and pulse oximetry (SpO2) were obtained, black patients had higher odds than white patients of having occult hypoxemia noted on arterial blood gas but not detected by pulse oximetry. This difference could limit access to supplemental oxygen and other more intensive support and treatments for black patients.

"Hospital Policies on Intravenous Vasopressor Administration and Monitoring: A Survey of Michigan Hospitals." Annals of the American Thoracic Society, 19(10), pp. 1769–1772

BACKGROUND: Pulse oximeters may produce less accurate results in non-White patients. RESEARCH QUESTION: Do pulse oximeters detect arterial hypoxemia less effectively in Black, Hispanic, and/or Asian patients than in White patients in respiratory failure and about to undergo extracorporeal membrane oxygenation (ECMO)? STUDY DESIGN AND METHODS: Data on adult patients with respiratory failure readings 6 h before ECMO were provided by the Extracorporeal Life Support Organization registry. Data was collected from 324 centers between January 2019 and July 2020. Our primary analysis was of rates of occult hypoxemia-low arterial oxygen saturation (Sao(2) <= 88%) on arterial blood gas measurement despite a pulse oximetry reading in the range of 92% to 96%. RESULTS: The rate of pre-ECMO occult hypoxemia, that is, arterial oxygen saturation (Sao(2)) <= 88%, was 10.2% (95% CI, 6.2%-15.3%) for 186 White patients with peripheral oxygen saturation (Sao(2)) of 92% to 96%; 21.5% (95% CI, 11.3%-35.3%) for 51 Black patients (P = .031 vs White); 8.6% (95% CI, 3.2%-17.7%) for 70 Hispanic patients (P = .693 vs White); and 9.2% (95% CI, 3.5%-19.0%) for 65 Asian patients (P = .820 vs White). Black patients with respiratory failure had a statistically significantly higher risk of occult hypoxemia with an OR of 2.57 (95% CI, 1.12-5.92) compared with White patients (P = .026). The risk of occult hypoxemia for Hispanic and Asian patients was equivalent to that of White patients. In a secondary analysis of patients with Sao(2 )<= 88% despite Sao(2 )> 96%, Black patients had more than three times the risk compared with White patients (OR, 3.52; 95% CI, 1.12-11.10; P = .032). INTERPRETATION: Compared with White patients, the prevalence of occult hypoxemia was higher in Black patients than in White patients about to undergo ECMO for respiratory failure, but it was comparable in Hispanic and Asian patients compared with White patients.

Rationale: Quantifying acute respiratory disease syndrome (ARDS) severity is essential for prognostic enrichment to stratify patients for invasive or higher-risk treatments; however, the comparative performance of many ARDS severity measures is unknown. Objectives: To validate ARDS severity measures for their ability to predict hospital mortality and an ARDS-specific outcome (defined as death from pulmonary dysfunction or the need for extracorporeal membrane oxygenation [ECMO] therapy). Methods: We compared five individual ARDS severity measures including the ratio of arterial oxygen tension/pressure to fraction of inspired oxygen (Pa-O2/FIO2 ratio), oxygenation index, ventilatory ratio, lung compliance, and radiologic assessment of lung edema (RALE); two ARDS composite severity scores including the Murray Lung Injury Score, and a novel score combining RALE, Pa-O2/FIO2 ratio, and ventilatory ratio; and the Acute Physiology and Chronic Health Evaluation IV score, using data collected at ARDS onset in patients hospitalized at a single center in 2016 and 2017. Discrimination of hospital mortality and the ARDS-specific outcome was evaluated using the area under the receiver operator characteristic curve (AUROC). Measure calibration was also evaluated. Results: Among 340 patients with ARDS, 125 (37%) died during hospitalization and 36 (10.6%) had the ARDS-specific outcome, including one who received ECMO. Among the five individual ARDS severity measures, the RALE score had the highest discrimination of the ARDS-specific outcome (AUROC = 0.67; 95% confidence interval [CI], 0.58-0.77), although other ARDS severity measures had similar performances. However, their ability to discriminate overall mortality was low. In contrast, the Acute Physiology and Chronic Health Evaluation IV score best discriminated overall mortality (AUROC = 0.73; 95% CI, 0.67-0.79) but was unable to discriminate the ARDS-specific outcome (AUROC = 0.54; 95% CI, 0.44-0.65). Among ARDS composite severity scores, the lung injury score had an AUROC= 0.67 (95% CI, 0.58-0.75) for the ARDS-specific outcome whereas the novel score had an AUROC= 0.70 (95% CI, 0.61-0.79). Patients grouped by quartile of the novel score had a 6%, 2%, 10%, and 24% rate of the ARDS-specific outcome. Conclusions: Although most ARDS severity measures had poor discrimination of hospital mortality, they performed better at predicting death from severe pulmonary dysfunction or ECMO needs. A novel composite score had the highest discrimination of this outcome.

Background: Invasive procedures are a core aspect of pulmonary and critical care practice. Procedures performed in the intensive care unit can be divided into high-risk, low-volume (HRLV) procedures and low-risk, high-volume (LRHV) procedures. HRLV procedures include cricothyroidotomy, pericardiocentesis, Blakemore tube placement, and bronchial blocker placement. LRHV procedures include arterial line placement, central venous catheter placement, thoracentesis, and flexible bronchoscopy. Despite the frequency and importance of procedures in critical care medicine, little is known about the similarities and differences in procedural training between different Pulmonary and Critical Care Medicine (PCCM) and Critical Care Medicine (CCM) training programs. Furthermore, differences in procedural training practices for HRLV and LRHV procedures have not previously been described. Objective: To assess procedural training practices in PCCM and CCM fellowship programs in the United States, and compare differences in training between HRLV and LRHV procedures. Methods: A novel survey instrument was developed and disseminated to PCCM and CCM program directors and associate program directors at PCCM and CCM fellowship programs in the United States to assess procedural teaching practices for HRLV and LRHV procedures. Results: The survey was sent to 221 fellowship programs, 168 PCCM and 34 CCM, with 70 unique respondents (31.7% response rate). Of the procedural educational strategies assessed, each strategy was used significantly more frequently for LRHV versus HRLV procedures. The majority of respondents (51.1%) report having no dedicated training for HRLV procedures versus 6.9% reporting no dedicated training for any LRHV procedure (P < 0.001). For HRLV procedures, 76.9% of respondents indicated that there was no set number of procedures required to determine competency, versus 25.3% for LRHV procedures (P < 0.001). For LRHV procedures, fellows were allowed to perform procedures independently without supervision 21.7% of the time versus 3.9% for HRLV procedures (P = 0.004). Program directors' confidence in their ability to determine fellows' competence in performing procedures was significantly lower for HRLV versus LRHV versus HRLV procedures (P < 0.001). Conclusion: Significant differences exist in procedural training education for PCCM and CCM fellows for LRHV versus HRLV procedures, and awareness of this discrepancy presents an opportunity to address this educational gap in PCCM and CCM fellowship training.

Background Acute respiratory distress syndrome (ARDS) is a common, but under-recognised, critical illness syndrome associated with high mortality. An important factor in its under-recognition is the variability in chest radiograph interpretation for ARDS. We sought to train a deep convolutional neural network (CNN) to detect ARDS findings on chest radiographs. Methods CNNs were pretrained on 595 506 radiographs from two centres to identify common chest findings (eg, opacity and effusion), and then trained on 8072 radiographs annotated for ARDS by multiple physicians using various transfer learning approaches. The best performing CNN was tested on chest radiographs in an internal and external cohort, including a subset reviewed by six physicians, including a chest radiologist and physicians trained in intensive care medicine. Chest radiograph data were acquired from four US hospitals. Findings In an internal test set of 1560 chest radiographs from 455 patients with acute hypoxaemic respiratory failure, a CNN could detect ARDS with an area under the receiver operator characteristics curve (AUROC) of 0.92 (95% CI 0.89-0.94). In the subgroup of 413 images reviewed by at least six physicians, its AUROC was 0.93 (95% CI 0.88-0.96), sensitivity 83.0% (95% CI 74.0-91.1), and specificity 88.3% (95% CI 83.1-92.8). Among images with zero of six ARDS annotations (n=155), the median CNN probability was 11%, with six (4%) assigned a probability above 50%. Among images with six of six ARDS annotations (n=27), the median CNN probability was 91%, with two (7%) assigned a probability below 50%. In an external cohort of 958 chest radiographs from 431 patients with sepsis, the AUROC was 0.88 (95% CI 0.85-0.91). When radiographs annotated as equivocal were excluded, the AUROC was 0.93 (0.92-0.95). Interpretation A CNN can be trained to achieve expert physician-level performance in ARDS detection on chest radiographs. Further research is needed to evaluate the use of these algorithms to support real-time identification of ARDS patients to ensure fidelity with evidence-based care or to support ongoing ARDS research. (C) 2021 The Author(s). Published by Elsevier Ltd.

A 47-year-old man in Michigan presented to the ED with a 1-day history of presyncope, swelling of the eyelids, and diffuse pruritic rash. He reported a 3-week history of postprandial watery bowel movements and a 5-day history of diffuse, crampy abdominal pain.

Rationale: Patients with severe coronavirus disease (COVID-19) meet clinical criteria for the acute respiratory distress syndrome (ARDS), yet early reports suggested they differ physiologically and clinically from patients with non-COVID-19 ARDS, prompting treatment recommendations that deviate from standard evidence-based practices for ARDS. Objectives: To compare respiratory physiology, clinical outcomes, and extrapulmonary clinical features of severe COVID-19 with non-COVID-19 ARDS. Methods: We performed a retrospective cohort study, comparing 130 consecutive mechanically ventilated patients with severe COVID-19 with 382 consecutive mechanically ventilated patients with non-COVID-19 ARDS. Initial respiratory physiology and 28-day outcomes were compared. Extrapulmonary manifestations (inflammation, extrapulmonary organ injury, and coagulation) were compared in an exploratory analysis. Results: Comparison of patients with COVID-19 and non-COVID-19 ARDS suggested small differences in respiratory compliance, ventilatory efficiency, and oxygenation. The 28-day mortality was 30% in patients with COVID-19 and 38% in patients with non-COVID-19 ARDS. In adjusted analysis, point estimates of differences in time to breathing unassisted at 28 days (adjusted subdistributional hazards ratio, 0.98 [95% confidence interval (CI), 0.77-1.26]) and 28-day mortality (risk ratio, 1.01 [95% CI, 0.72-1.42]) were small for COVID-19 versus non-COVID-19 ARDS, although the confidence intervals for these estimates include moderate differences. Patients with COVID-19 had lower neutrophil counts but did not differ in lymphocyte count or other measures of systemic inflammation. Conclusions: In this single-center cohort, we found no evidence for large differences between COVID-19 and non-COVID-19 ARDS. Many key clinical features of severe COVID-19 were similar to those of non-COVID-19 ARDS, including respiratory physiology and clinical outcomes, although our sample size precludes definitive conclusions. Further studies are needed to define COVID-19-specific pathophysiology before a deviation from evidence-based treatment practices can be recommended.

OBJECTIVES:To prepare for the predicted surge of patients with COVID-19 in Southeast Michigan, the US Department of Veterans Affairs Ann Arbor Healthcare System engineered, built, and staffed a 12-bed intensive care unit (ICU) from the existing postanesthesia care unit (PACU).OBSERVATIONS:Considerations including floor planning, reversal of airflow, strict airborne precautions, sealing off a dedicated nursing station, and developing an infection control plan in an open care unit. A staffing model was created that included anesthesiologist intensivists, advanced practice providers, residents, certified registered nurse anesthetists, and perioperative nurses working alongside ICU trained nurses. Challenges arose in infection control, communication, mechanical ventilation using anesthesia machines, providing renal replacement therapy, and maintaining patient privacy in an open unit.CONCLUSIONS:This article describes the setup, challenges, and solutions that allowed the creation of the PACU-ICU to help serve veterans and civilians during a time of unprecedented strain on the health care system due to COVID-19.

Background Acute hypoxemic respiratory failure (AHRF) and acute respiratory distress syndrome (ARDS) are associated with high in-hospital mortality. However, in cohorts of ARDS patients from the 1990s, patients more commonly died from sepsis or multi-organ failure rather than refractory hypoxemia. Given increased attention to lung-protective ventilation and sepsis treatment in the past 25 years, we hypothesized that causes of death may be different among contemporary cohorts. These differences may provide clinicians with insight into targets for future therapeutic interventions. Methods We identified adult patients hospitalized at a single tertiary care center (2016–2017) with AHRF, defined as PaO 2 /FiO 2 ≤ 300 while receiving invasive mechanical ventilation for > 12 h, who died during hospitalization. ARDS was adjudicated by multiple physicians using the Berlin definition. Separate abstractors blinded to ARDS status collected data on organ dysfunction and withdrawal of life support using a standardized tool. The primary cause of death was defined as the organ system that most directly contributed to death or withdrawal of life support. Results We identified 385 decedents with AHRF, of whom 127 (33%) had ARDS. The most common primary causes of death were sepsis (26%), pulmonary dysfunction (22%), and neurologic dysfunction (19%). Multi-organ failure was present in 70% at time of death, most commonly due to sepsis (50% of all patients), and 70% were on significant respiratory support at the time of death. Only 2% of patients had insupportable oxygenation or ventilation. Eighty-five percent died following withdrawal of life support. Patients with ARDS more often had pulmonary dysfunction as the primary cause of death (28% vs 19%; p = 0.04) and were also more likely to die while requiring significant respiratory support (82% vs 64%; p < 0.01). Conclusions In this contemporary cohort of patients with AHRF, the most common primary causes of death were sepsis and pulmonary dysfunction, but few patients had insupportable oxygenation or ventilation. The vast majority of deaths occurred after withdrawal of life support. ARDS patients were more likely to have pulmonary dysfunction as the primary cause of death and die while requiring significant respiratory support compared to patients without ARDS.