Real-Time Machine Learning Alerts to Prevent Escalation of Care: A Nonrandomized Clustered Pragmatic Clinical Trial.

OBJECTIVES: Machine learning algorithms can outperform older methods in predicting clinical deterioration, but rigorous prospective data on their real-world efficacy are limited. We hypothesized that real-time machine learning generated alerts sent directly to front-line providers would reduce escalations. DESIGN: Single-center prospective pragmatic nonrandomized clustered clinical trial. SETTING: Academic tertiary care medical center. PATIENTS: Adult patients admitted to four medical-surgical units. Assignment to intervention or control arms was determined by initial unit admission. INTERVENTIONS: Real-time alerts stratified according to predicted likelihood of deterioration sent either to the primary team or directly to the rapid response team (RRT). Clinical care and interventions were at the providers' discretion. For the control units, alerts were generated but not sent, and standard RRT activation criteria were used. MEASUREMENTS AND MAIN RESULTS: The primary outcome was the rate of escalation per 1000 patient bed days. Secondary outcomes included the frequency of orders for fluids, medications, and diagnostic tests, and combined in-hospital and 30-day mortality. Propensity score modeling with stabilized inverse probability of treatment weight (IPTW) was used to account for differences between groups. Data from 2740 patients enrolled between July 2019 and March 2020 were analyzed (1488 intervention, 1252 control). Average age was 66.3 years and 1428 participants (52%) were female. The rate of escalation was 12.3 vs. 11.3 per 1000 patient bed days (difference, 1.0; 95% CI, -2.8 to 4.7) and IPTW adjusted incidence rate ratio 1.43 (95% CI, 1.16-1.78; p < 0.001). Patients in the intervention group were more likely to receive cardiovascular medication orders (16.1% vs. 11.3%; 4.7%; 95% CI, 2.1-7.4%) and IPTW adjusted relative risk (RR) (1.74; 95% CI, 1.39-2.18; p < 0.001). Combined in-hospital and 30-day-mortality was lower in the intervention group (7% vs. 9.3%; -2.4%; 95% CI, -4.5% to -0.2%) and IPTW adjusted RR (0.76; 95% CI, 0.58-0.99; p = 0.045). CONCLUSIONS: Real-time machine learning alerts do not reduce the rate of escalation but may reduce mortality.

Obesity as a mortality risk factor in the medical ward: a case control study.

BACKGROUND: Research regarding the association between severe obesity and in-hospital mortality is inconsistent. We evaluated the impact of body mass index (BMI) levels on mortality in the medical wards. The analysis was performed separately before and during the COVID-19 pandemic. METHODS: We retrospectively retrieved data of adult patients admitted to the medical wards at the Mount Sinai Health System in New York City. The study was conducted between January 1, 2011, to March 23, 2021. Patients were divided into two sub-cohorts: pre-COVID-19 and during-COVID-19. Patients were then clustered into groups based on BMI ranges. A multivariate logistic regression analysis compared the mortality rate among the BMI groups, before and during the pandemic. RESULTS: Overall, 179,288 patients were admitted to the medical wards and had a recorded BMI measurement. 149,098 were admitted before the COVID-19 pandemic and 30,190 during the pandemic. Pre-pandemic, multivariate analysis showed a "J curve" between BMI and mortality. Severe obesity (BMI > 40) had an aOR of 0.8 (95% CI:0.7-1.0, p = 0.018) compared to the normal BMI group. In contrast, during the pandemic, the analysis showed a "U curve" between BMI and mortality. Severe obesity had an aOR of 1.7 (95% CI:1.3-2.4, p < 0.001) compared to the normal BMI group. CONCLUSIONS: Medical ward patients with severe obesity have a lower risk for mortality compared to patients with normal BMI. However, this does not apply during COVID-19, where obesity was a leading risk factor for mortality in the medical wards. It is important for the internal medicine physician to understand the intricacies of the association between obesity and medical ward mortality.

AKI in Hospitalized Patients with COVID-19.

BACKGROUND: Early reports indicate that AKI is common among patients with coronavirus disease 2019 (COVID-19) and associated with worse outcomes. However, AKI among hospitalized patients with COVID-19 in the United States is not well described. METHODS: This retrospective, observational study involved a review of data from electronic health records of patients aged ≥18 years with laboratory-confirmed COVID-19 admitted to the Mount Sinai Health System from February 27 to May 30, 2020. We describe the frequency of AKI and dialysis requirement, AKI recovery, and adjusted odds ratios (aORs) with mortality. RESULTS: Of 3993 hospitalized patients with COVID-19, AKI occurred in 1835 (46%) patients; 347 (19%) of the patients with AKI required dialysis. The proportions with stages 1, 2, or 3 AKI were 39%, 19%, and 42%, respectively. A total of 976 (24%) patients were admitted to intensive care, and 745 (76%) experienced AKI. Of the 435 patients with AKI and urine studies, 84% had proteinuria, 81% had hematuria, and 60% had leukocyturia. Independent predictors of severe AKI were CKD, men, and higher serum potassium at admission. In-hospital mortality was 50% among patients with AKI versus 8% among those without AKI (aOR, 9.2; 95% confidence interval, 7.5 to 11.3). Of survivors with AKI who were discharged, 35% had not recovered to baseline kidney function by the time of discharge. An additional 28 of 77 (36%) patients who had not recovered kidney function at discharge did so on posthospital follow-up. CONCLUSIONS: AKI is common among patients hospitalized with COVID-19 and is associated with high mortality. Of all patients with AKI, only 30% survived with recovery of kidney function by the time of discharge.

RT-PCR/MALDI-TOF mass spectrometry-based detection of SARS-CoV-2 in saliva specimens.

As severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infections continue, there is a substantial need for cost-effective and large-scale testing that utilizes specimens that can be readily collected from both symptomatic and asymptomatic individuals in various community settings. Although multiple diagnostic methods utilize nasopharyngeal specimens, saliva specimens represent an attractive alternative as they can rapidly and safely be collected from different populations. While saliva has been described as an acceptable clinical matrix for the detection of SARS-CoV-2, evaluations of analytic performance across platforms for this specimen type are limited. Here, we used a novel sensitive RT-PCR/MALDI-TOF mass spectrometry-based assay (Agena MassARRAY®) to detect SARS-CoV-2 in saliva specimens. The platform demonstrated high diagnostic sensitivity and specificity when compared to matched patient upper respiratory specimens. We also evaluated the analytical sensitivity of the platform and determined the limit of detection of the assay to be 1562.5 copies/ml. Furthermore, across the five individual target components of this assay, there was a range in analytic sensitivities for each target with the N2 target being the most sensitive. Overall, this system also demonstrated comparable performance when compared to the detection of SARS-CoV-2 RNA in saliva by the cobas® 6800/8800 SARS-CoV-2 real-time RT-PCR Test (Roche). Together, we demonstrate that saliva represents an appropriate matrix for SARS-CoV-2 detection on the novel Agena system as well as on a conventional real-time RT-PCR assay. We conclude that the MassARRAY® system is a sensitive and reliable platform for SARS-CoV-2 detection in saliva, offering scalable throughput in a large variety of clinical laboratory settings.

Blood Donation and COVID-19: Reconsidering the 3-Month Deferral Policy for Gay, Bisexual, Transgender, and Other Men Who Have Sex With Men.

In April 2020, in light of COVID-19-related blood shortages, the US Food and Drug Administration (FDA) reduced the deferral period for men who have sex with men (MSM) from its previous duration of 1 year to 3 months.Although originally born out of necessity, the decades-old restrictions on MSM donors have been mitigated by significant advancements in HIV screening, treatment, and public education. The severity of the ongoing COVID-19 pandemic-and the urgent need for safe blood products to respond to such crises-demands an immediate reconsideration of the 3-month deferral policy for MSM.We review historical HIV testing and transmission evidence, discuss the ethical ramifications of the current deferral period, and examine the issue of noncompliance with donor deferral rules. We also propose an eligibility screening format that involves an individual risk-based screening protocol and, unlike current FDA guidelines, does not effectively exclude donors on the basis of gender identity or sexual orientation. Our policy proposal would allow historically marginalized community members to participate with dignity in the blood donation process without compromising blood donation and transfusion safety outcomes.

MUST-Plus: A Machine Learning Classifier That Improves Malnutrition Screening in Acute Care Facilities.

OBJECTIVE: Malnutrition among hospital patients, a frequent, yet under-diagnosed problem is associated with adverse impact on patient outcome and health care costs. Development of highly accurate malnutrition screening tools is, therefore, essential for its timely detection, for providing nutritional care, and for addressing the concerns related to the suboptimal predictive value of the conventional screening tools, such as the Malnutrition Universal Screening Tool (MUST). We aimed to develop a machine learning (ML) based classifier (MUST-Plus) for more accurate prediction of malnutrition. METHOD: A retrospective cohort with inpatient data consisting of anthropometric, lab biochemistry, clinical data, and demographics from adult (≥ 18 years) admissions at a large tertiary health care system between January 2017 and July 2018 was used. The registered dietitian (RD) nutritional assessments were used as the gold standard outcome label. The cohort was randomly split (70:30) into training and test sets. A random forest model was trained using 10-fold cross-validation on training set, and its predictive performance on test set was compared to MUST. RESULTS: In all, 13.3% of admissions were associated with malnutrition in the test cohort. MUST-Plus provided 73.07% (95% confidence interval [CI]: 69.61%-76.33%) sensitivity, 76.89% (95% CI: 75.64%-78.11%) specificity, and 83.5% (95% CI: 82.0%-85.0%) area under the receiver operating curve (AUC). Compared to classic MUST, MUST-Plus demonstrated 30% higher sensitivity, 6% higher specificity, and 17% increased AUC. CONCLUSIONS: ML-based MUST-Plus provided superior performance in identifying malnutrition compared to the classic MUST. The tool can be used for improving the operational efficiency of RDs by timely referrals of high-risk patients.

Developing an Institute for Health Care Delivery Science: successes, challenges, and solutions in the first five years.

Medical knowledge is increasing at an exponential rate. At the same time, unexplained variations in practice and patient outcomes and unacceptable rates of medical errors and inefficiencies in health care delivery have emerged. Our Institute for Health Care Delivery Science (I-HDS) began in 2014 as a novel platform to conduct multidisciplinary healthcare delivery research. We followed ten strategies to develop a successful institute with excellence in methodology and strong understanding of the value of team science. Our work was organized around five hubs: 1) Quality/Process Improvement and Systematic Review, 2) Comparative Effectiveness Research, Pragmatic Clinical Trials, and Predictive Analytics, 3) Health Economics and Decision Modeling, 4) Qualitative, Survey, and Mixed Methods, and 5) Training and Mentoring. In the first 5 years of the I-HDS, we have identified opportunities for change in clinical practice through research using our health system's electronic health record (EHR) data, and designed programs to educate clinicians in the value of research to improve patient care and recognize efficiencies in processes. Testing the value of several model interventions has guided prioritization of evidence-based quality improvements. Some of the changes in practice have already been embedded in the EHR workflow successfully. Development and sustainability of the I-HDS has been fostered by a mix of internal and external funding, including philanthropic foundations. Challenges remain due to the highly competitive funding environment and changes needed to adapt the EHR to healthcare delivery research. Further stakeholder engagement and culture change working with hospital leadership and I-HDS core and affiliate members continues.

Association between COVID-19 diagnosis and presenting chief complaint from New York City triage data.

BACKGROUND AND AIM: New York City (NYC) is an epicenter of the COVID-19 pandemic in the United States. Proper triage of patients with possible COVID-19 via chief complaint is critical but not fully optimized. This study aimed to investigate the association between presentation by chief complaints and COVID-19 status. METHODS: We retrospectively analyzed adult emergency department (ED) patient visits from five different NYC hospital campuses from March 1, 2020 to May 13, 2020 of patients who underwent nasopharyngeal COVID-19 RT-PCR testing. The positive and negative COVID-19 cohorts were then assessed for different chief complaints obtained from structured triage data. Sub-analysis was performed for patients older than 65 and within chief complaints with high mortality. RESULTS: Of 11,992 ED patient visits who received COVID-19 testing, 6524/11992 (54.4%) were COVID-19 positive. 73.5% of fever, 67.7% of shortness of breath, and 65% of cough had COVID-19, but others included 57.5% of weakness/fall/altered mental status, 55.5% of glycemic control, and 51.4% of gastrointestinal symptoms. In patients over 65, 76.7% of diarrhea, 73.7% of fatigue, and 69.3% of weakness had COVID-19. 45.5% of dehydration, 40.5% of altered mental status, 27% of fall, and 24.6% of hyperglycemia patients experienced mortality. CONCLUSION: A novel high risk COVID-19 patient population was identified from chief complaint data, which is different from current suggested CDC guidelines, and may help triage systems to better isolate COVID-19 patients. Older patients with COVID-19 infection presented with more atypical complaints warranting special consideration. COVID-19 was associated with higher mortality in a unique group of complaints also warranting special consideration.

Moderate or Severe Impairment in Pulmonary Function is Associated with Mortality in Sarcoidosis Patients Infected with SARS­CoV­2.

PURPOSE: To investigate whether sarcoidosis patients infected with SARS-CoV-2 are at risk for adverse disease outcomes. STUDY DESIGN AND METHODS: This retrospective study was conducted in five hospitals within the Mount Sinai Health System during March 1, 2020 to July 29, 2020. All patients diagnosed with COVID-19 were included in the study. We identified sarcoidosis patients who met diagnostic criteria for sarcoidosis according to accepted guidelines. An adverse disease outcome was defined as the presence of intubation and mechanical ventilation or in-hospital mortality. In sarcoidosis patients, we reported (when available) the results of pulmonary function testing measured within 3 years prior to the time of SARS­CoV­2 infection. A multivariable logistic regression model was used to generate an adjusted odds ratio (aOR) to evaluate sarcoidosis as a risk factor for an adverse outcome. The same model was used to analyze sarcoidosis patients with moderate and/or severe impairment in pulmonary function. RESULTS: The study included 7337 patients, 37 of whom (0.5%) had sarcoidosis. The crude rate of developing an adverse outcome was significantly higher in patients with moderately and/or severely impaired pulmonary function (9/14 vs. 3/23, p = 0.003). While the diagnosis of sarcoidosis was not independently associated with risk of an adverse event, (aOR 1.8, 95% CI 0.9-3.6), the diagnosis of sarcoidosis in patients with moderately and/or severely impaired pulmonary function was associated with an adverse outcome (aOR 7.8, 95% CI 2.4-25.8). CONCLUSION: Moderate or severe impairment in pulmonary function is associated with mortality in sarcoidosis patients infected with SARS­CoV­2.

Machine Learning to Predict Mortality and Critical Events in a Cohort of Patients With COVID-19 in New York City: Model Development and Validation.

BACKGROUND: COVID-19 has infected millions of people worldwide and is responsible for several hundred thousand fatalities. The COVID-19 pandemic has necessitated thoughtful resource allocation and early identification of high-risk patients. However, effective methods to meet these needs are lacking. OBJECTIVE: The aims of this study were to analyze the electronic health records (EHRs) of patients who tested positive for COVID-19 and were admitted to hospitals in the Mount Sinai Health System in New York City; to develop machine learning models for making predictions about the hospital course of the patients over clinically meaningful time horizons based on patient characteristics at admission; and to assess the performance of these models at multiple hospitals and time points. METHODS: We used Extreme Gradient Boosting (XGBoost) and baseline comparator models to predict in-hospital mortality and critical events at time windows of 3, 5, 7, and 10 days from admission. Our study population included harmonized EHR data from five hospitals in New York City for 4098 COVID-19-positive patients admitted from March 15 to May 22, 2020. The models were first trained on patients from a single hospital (n=1514) before or on May 1, externally validated on patients from four other hospitals (n=2201) before or on May 1, and prospectively validated on all patients after May 1 (n=383). Finally, we established model interpretability to identify and rank variables that drive model predictions. RESULTS: Upon cross-validation, the XGBoost classifier outperformed baseline models, with an area under the receiver operating characteristic curve (AUC-ROC) for mortality of 0.89 at 3 days, 0.85 at 5 and 7 days, and 0.84 at 10 days. XGBoost also performed well for critical event prediction, with an AUC-ROC of 0.80 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. In external validation, XGBoost achieved an AUC-ROC of 0.88 at 3 days, 0.86 at 5 days, 0.86 at 7 days, and 0.84 at 10 days for mortality prediction. Similarly, the unimputed XGBoost model achieved an AUC-ROC of 0.78 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. Trends in performance on prospective validation sets were similar. At 7 days, acute kidney injury on admission, elevated LDH, tachypnea, and hyperglycemia were the strongest drivers of critical event prediction, while higher age, anion gap, and C-reactive protein were the strongest drivers of mortality prediction. CONCLUSIONS: We externally and prospectively trained and validated machine learning models for mortality and critical events for patients with COVID-19 at different time horizons. These models identified at-risk patients and uncovered underlying relationships that predicted outcomes.

Bedside medication delivery programs: suggestions for systematic evaluation and reporting.

PURPOSE: Several factors lead to medication non-adherence after hospital discharge. Hospitals and pharmacies have implemented bedside medication delivery (BMD) programs for patients, in an attempt to reduce barriers and improve medication adherence. Here, we provide a critical review of the literature on these programs. DATA SOURCES: We conducted a literature search on BMD programs in PubMed, Google Scholar, Scopus and a general Google search using these keywords: 'medication delivery bedside', 'discharge medication delivery', 'meds to bedside' and 'meds to beds'. STUDY SELECTION: We identified 10 reports and include data from all reports. DATA EXTRACTION: Data on study characteristics and settings were extracted along with four outcomes: medication error, patient satisfaction, 30-day hospital readmission and visits to the emergency department. RESULTS OF DATA SYNTHESIS: Of the 10 reports, only 4 were peer-reviewed publications; others were reported in the lay press. Outcomes were reported in both qualitative and quantitative terms. Less than half of reports provided quantitative data on 30-day readmission and patient satisfaction. Others suggested qualitative improvement in these outcomes but did not provide data or specific details. None reported outcomes of their programs beyond 30 days. CONCLUSION: We highlight the need for increased use of optimal program design and more rigorous evaluations of the impact of BMD programs. We also provide guidelines on the types of evaluations that are likely needed and encourage improved reporting.

Pharmacological risk factors associated with hospital readmission rates in a psychiatric cohort identified using prescriptome data mining.

BACKGROUND: Worldwide, over 14% of individuals hospitalized for psychiatric reasons have readmissions to hospitals within 30 days after discharge. Predicting patients at risk and leveraging accelerated interventions can reduce the rates of early readmission, a negative clinical outcome (i.e., a treatment failure) that affects the quality of life of patient. To implement individualized interventions, it is necessary to predict those individuals at highest risk for 30-day readmission. In this study, our aim was to conduct a data-driven investigation to find the pharmacological factors influencing 30-day all-cause, intra- and interdepartmental readmissions after an index psychiatric admission, using the compendium of prescription data (prescriptome) from electronic medical records (EMR). METHODS: The data scientists in the project received a deidentified database from the Mount Sinai Data Warehouse, which was used to perform all analyses. Data was stored in a secured MySQL database, normalized and indexed using a unique hexadecimal identifier associated with the data for psychiatric illness visits. We used Bayesian logistic regression models to evaluate the association of prescription data with 30-day readmission risk. We constructed individual models and compiled results after adjusting for covariates, including drug exposure, age, and gender. We also performed digital comorbidity survey using EMR data combined with the estimation of shared genetic architecture using genomic annotations to disease phenotypes. RESULTS: Using an automated, data-driven approach, we identified prescription medications, side effects (primary side effects), and drug-drug interaction-induced side effects (secondary side effects) associated with readmission risk in a cohort of 1275 patients using prescriptome analytics. In our study, we identified 28 drugs associated with risk for readmission among psychiatric patients. Based on prescription data, Pravastatin had the highest risk of readmission (OR = 13.10; 95% CI (2.82, 60.8)). We also identified enrichment of primary side effects (n = 4006) and secondary side effects (n = 36) induced by prescription drugs in the subset of readmitted patients (n = 89) compared to the non-readmitted subgroup (n = 1186). Digital comorbidity analyses and shared genetic analyses further reveals that cardiovascular disease and psychiatric conditions are comorbid and share functional gene modules (cardiomyopathy and anxiety disorder: shared genes (n = 37; P = 1.06815E-06)). CONCLUSIONS: Large scale prescriptome data is now available from EMRs and accessible for analytics that could improve healthcare outcomes. Such analyses could also drive hypothesis and data-driven research. In this study, we explored the utility of prescriptome data to identify factors driving readmission in a psychiatric cohort. Converging digital health data from EMRs and systems biology investigations reveal a subset of patient populations that have significant comorbidities with cardiovascular diseases are more likely to be readmitted. Further, the genetic architecture of psychiatric illness also suggests overlap with cardiovascular diseases. In summary, assessment of medications, side effects, and drug-drug interactions in a clinical setting as well as genomic information using a data mining approach could help to find factors that could help to lower readmission rates in patients with mental illness.

Effectiveness of an Electronic Alert for Hypotension and Low Bispectral Index on 90-day Postoperative Mortality: A Prospective, Randomized Trial.

BACKGROUND: We tested the hypothesis that an electronic alert for a "double low" of mean arterial pressure less than 75 mmHg and a bispectral index less than 45 reduces the primary outcome of 90-day mortality. METHODS: Adults having noncardiac surgery were randomized to receive either intraoperative alerts for double-low events or no alerts. Anesthesiologists were not blinded and not required to alter care based upon the alerts. The primary outcome was all-cause 90-day mortality. RESULTS: Patients (20,239) were randomized over 33 months, and 19,092 were analyzed. After adjusting for age, comorbidities, and perioperative factors, patients with more than 60 min of cumulative double-low time were twice as likely to die (hazard ratio, 1.99; 95% CI, 1.2 to 3.2; P = 0.005). The median number of double-low minutes (quartiles) was only slightly lower in the alert arm: 10 (2 to 30) versus 12 (2 to 34) min. Ninety-day mortality was 135 (1.4%) in the alert arm and 123 (1.3%) in the control arm. The difference in percent mortality was 0.18% (99% CI, -0.25 to 0.61). CONCLUSIONS: Ninety-day mortality was not significantly lower in patients cared for by anesthesiologists who received automated alerts to double-low states. Prolonged cumulative double-low conditions were strongly associated with mortality.

Association between anesthesiologist experience and mortality after orthotopic liver transplantation.

The anesthesiologist has been recognized as an integral member of the liver transplant team, and previous studies have demonstrated that inter-anesthesiologist variability can be a driver of outcomes for high-risk patients. We hypothesized that anesthesiologist experience, defined as the number of previous liver transplants performed at our institution, the Icahn School of Medicine at Mount Sinai, would be independently associated with outcomes for liver transplant patients. Eight hundred forty-nine liver transplants performed between January 2003 and January 2013 with a total of 22 anesthesiologists were analyzed. Each transplant was assigned an incremental case number that corresponded to the number of transplants that the attending anesthesiologist had already performed at our institution. Several perioperative covariates were controlled for in the context of a generalized linear mixed effects model to detail the influence of threshold levels of the incremental case number on the primary outcome, 30-day mortality, and a secondary outcome, 30-day graft failure. Sensitivity analyses were conducted to confirm the robustness of these findings. An incremental case number ≤ 5 was associated with a significantly greater risk of 30-day mortality (odds ratio = 2.24, 95% confidence interval = 1.11-4.54, P = 0.025), and there was evidence suggestive of a greater risk of 30-day graft failure (odds ratio = 1.93, 95% confidence interval = 0.95-3.93, P = 0.071). Sensitivity analyses ruled out threats to the validity of these findings, including dropout effects and time trends in the overall performance of the transplantation unit. In conclusion, this study shows that an anesthesiologist's level of experience has a significant effect on outcomes for liver transplant recipients, with increased mortality and possibly graft failure during a provider's first 5 cases. These findings may indicate the need for increased training and supervision for anesthesiologists joining the liver transplant team.

Development and Validation of a Deep Learning Classifier Using Chest Radiographs to Predict Extubation Success in Patients Undergoing Invasive Mechanical Ventilation.

The decision to extubate patients on invasive mechanical ventilation is critical; however, clinician performance in identifying patients to liberate from the ventilator is poor. Machine Learning-based predictors using tabular data have been developed; however, these fail to capture the wide spectrum of data available. Here, we develop and validate a deep learning-based model using routinely collected chest X-rays to predict the outcome of attempted extubation. We included 2288 serial patients admitted to the Medical ICU at an urban academic medical center, who underwent invasive mechanical ventilation, with at least one intubated CXR, and a documented extubation attempt. The last CXR before extubation for each patient was taken and split 79/21 for training/testing sets, then transfer learning with k-fold cross-validation was used on a pre-trained ResNet50 deep learning architecture. The top three models were ensembled to form a final classifier. The Grad-CAM technique was used to visualize image regions driving predictions. The model achieved an AUC of 0.66, AUPRC of 0.94, sensitivity of 0.62, and specificity of 0.60. The model performance was improved compared to the Rapid Shallow Breathing Index (AUC 0.61) and the only identified previous study in this domain (AUC 0.55), but significant room for improvement and experimentation remains.

Increased left ventricular myocardial mass is associated with arrhythmias after cardiac surgery.

OBJECTIVES: The purpose of this study was to investigate whether patients with an elevated left ventricular mass index undergoing cardiac surgery were more likely to experience postoperative atrial and ventricular arrhythmias. DESIGN: A retrospective analysis. SETTING: A single tertiary care university hospital. PARTICIPANTS: One thousand consecutive patients undergoing all types of adult cardiac surgery. INTERVENTIONS: With institutional review board approval, intraoperative transesophageal echocardiographic images were reviewed by a single reviewer. The left ventricular mass index was calculated using the American Society of Echocardiography-recommended formula. Medical charts were reviewed for the occurrence and type of clinically significant postoperative arrhythmias. MEASUREMENTS AND RESULTS: Of the patients who had an elevated left ventricular mass index, 47.6% (225/473) developed clinically significant postoperative arrhythmias compared with 38.3% (142/371) of patients with a normal left ventricular mass index (odds ratio [OR] = 1.46; 95% confidence interval [CI], 1.11-1.93; p = 0.007). In the multivariate analysis, this finding remained statistically significant, controlling for the effects of age, weight, sex, surgery type, left ventricular function, functional status, left atrial dimensions, and a history of atrial fibrillation (OR = 1.40; 95% CI, 1.03-1.90 per 100-g/m(2) increase in the left ventricular mass index). An increased left ventricular mass index was also an independent predictor of the separate or combined occurrence of atrial or ventricular arrhythmias. CONCLUSIONS: An elevated left ventricular mass index was a strong independent predictor of clinically significant postoperative atrial and ventricular arrhythmias after adult cardiac surgery. Although prospective validation is required, targeting patients for arrhythmia prophylaxis therapy may be justified in patients with a left ventricular mass index >188 g/m(2).