Cardiovascular Subphenotypes in Acute Respiratory Distress Syndrome.

OBJECTIVES: To use clustering methods on transthoracic echocardiography (TTE) findings and hemodynamic parameters to characterize circulatory failure subphenotypes and potentially elucidate underlying mechanisms in patients with acute respiratory distress syndrome (ARDS) and to describe their association with mortality compared with current definitions of right ventricular dysfunction (RVD). DESIGN: Retrospective, single-center cohort study. SETTING: University Hospital ICU, Birmingham, United Kingdom. PATIENTS: ICU patients that received TTE within 7 days of ARDS onset between April 2016 and December 2021. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Latent class analysis (LCA) of TTE/hemodynamic parameters was performed in 801 patients, 62 years old (interquartile range, 50-72 yr old), 63% male, and 40% 90-day mortality rate. Four cardiovascular subphenotypes were identified: class 1 (43%; mostly normal left and right ventricular [LV/RV] function), class 2 (24%; mostly dilated RV with preserved systolic function), class 3 (13%, mostly dilated RV with impaired systolic function), and class 4 (21%; mostly high cardiac output, with hyperdynamic LV function). The four subphenotypes differed in their characteristics and outcomes, with 90-day mortality rates of 19%, 40%, 78%, and 59% in classes 1-4, respectively ( p < 0.0001). Following multivariable logistic regression analysis, class 3 had the highest odds ratio (OR) for mortality (OR, 6.9; 95% CI, 4.0-11.8) compared with other RVD definitions. Different three-variable models had high diagnostic accuracy in identifying each of these latent subphenotypes. CONCLUSIONS: LCA of TTE parameters identified four cardiovascular subphenotypes in ARDS that more closely aligned with circulatory failure mechanisms and mortality than current RVD definitions.

Right Ventricular Dysfunction and Its Association With Mortality in Coronavirus Disease 2019 Acute Respiratory Distress Syndrome.

OBJECTIVES: To assess whether right ventricular dilation or systolic impairment is associated with mortality and/or disease severity in invasively ventilated patients with coronavirus disease 2019 acute respiratory distress syndrome. DESIGN: Retrospective cohort study. SETTING: Single-center U.K. ICU. PATIENTS: Patients with coronavirus disease 2019 acute respiratory distress syndrome undergoing invasive mechanical ventilation that received a transthoracic echocardiogram between March and December 2020. INTERVENTION: None. MEASUREMENTS AND MAIN RESULTS: Right ventricular dilation was defined as right ventricular:left ventricular end-diastolic area greater than 0.6, right ventricular systolic impairment as fractional area change less than 35%, or tricuspid annular plane systolic excursion less than 17 mm. One hundred seventy-two patients were included, 59 years old (interquartile range, 49-67), with mostly moderate acute respiratory distress syndrome (n = 101; 59%). Ninety-day mortality was 41% (n = 70): 49% in patients with right ventricular dilation, 53% in right ventricular systolic impairment, and 72% in right ventricular dilation with systolic impairment. The right ventricular dilation with systolic impairment phenotype was independently associated with mortality (odds ratio, 3.11 [95% CI, 1.15-7.60]), but either disease state alone was not. Right ventricular fractional area change correlated with Pao2:Fio2 ratio, Paco2, chest radiograph opacification, and dynamic compliance, whereas right ventricular:left ventricle end-diastolic area correlated negatively with urine output. CONCLUSIONS: Right ventricular systolic impairment correlated with pulmonary pathophysiology, whereas right ventricular dilation correlated with renal dysfunction. Right ventricular dilation with systolic impairment was the only right ventricular phenotype that was independently associated with mortality.

Revealing transparency gaps in publicly available COVID-19 datasets used for medical artificial intelligence development-a systematic review.

During the COVID-19 pandemic, artificial intelligence (AI) models were created to address health-care resource constraints. Previous research shows that health-care datasets often have limitations, leading to biased AI technologies. This systematic review assessed datasets used for AI development during the pandemic, identifying several deficiencies. Datasets were identified by screening articles from MEDLINE and using Google Dataset Search. 192 datasets were analysed for metadata completeness, composition, data accessibility, and ethical considerations. Findings revealed substantial gaps: only 48% of datasets documented individuals' country of origin, 43% reported age, and under 25% included sex, gender, race, or ethnicity. Information on data labelling, ethical review, or consent was frequently missing. Many datasets reused data with inadequate traceability. Notably, historical paediatric chest x-rays appeared in some datasets without acknowledgment. These deficiencies highlight the need for better data quality and transparent documentation to lessen the risk that biased AI models are developed in future health emergencies.

The value of standards for health datasets in artificial intelligence-based applications.

Artificial intelligence as a medical device is increasingly being applied to healthcare for diagnosis, risk stratification and resource allocation. However, a growing body of evidence has highlighted the risk of algorithmic bias, which may perpetuate existing health inequity. This problem arises in part because of systemic inequalities in dataset curation, unequal opportunity to participate in research and inequalities of access. This study aims to explore existing standards, frameworks and best practices for ensuring adequate data diversity in health datasets. Exploring the body of existing literature and expert views is an important step towards the development of consensus-based guidelines. The study comprises two parts: a systematic review of existing standards, frameworks and best practices for healthcare datasets; and a survey and thematic analysis of stakeholder views of bias, health equity and best practices for artificial intelligence as a medical device. We found that the need for dataset diversity was well described in literature, and experts generally favored the development of a robust set of guidelines, but there were mixed views about how these could be implemented practically. The outputs of this study will be used to inform the development of standards for transparency of data diversity in health datasets (the STANDING Together initiative).