The Connected Intensive Care Unit Patient: Exploratory Analyses and Cohort Discovery From a Critical Care Telemedicine Database

Association between white blood cell count to hemoglobin ratio and risk of in-hospital mortality in patients with lung cancer

BACKGROUND

The objective of this study was to investigate the association between white blood cell count to hemoglobin ratio (WHR) and risk of in-hospital mortality in patients with lung cancer.

METHODS

In this retrospective cohort study, the medical records of patients with lung cancer were retrieved from the electronic ICU (eICU) Collaborative Research Database between 2014 and 2015. The primary outcome was in-hospital mortality. The secondary outcome was the length of stay in intensive care unit (ICU). The cut-off value for the WHR was calculated by the X-tile software. The Cox model was applied to assess the association between WHR and in-hospital mortality among patients with lung cancer and the linear regression model was used to investigate the association between WHR and length of ICU stay. Subgroup analyses of age (< 65 years or >  = 65 years), Acute Physiology and Chronic Health Evaluation (APACHE) score (< 59 or >  = 59), gender, ventilation (yes or no), and vasopressor (yes or no) in patients with lung cancer were conducted.

RESULTS

Of the 768 included patients with lung cancer, 153 patients (19.92%) died in the hospital. The median total follow-up time was 6.88 (4.17, 11.23) days. The optimal cut-off value for WHR was 1.4. ICU lung cancer patients with WHR >  = 1.4 had a significantly higher risk of in-hospital mortality [Hazard ratio: (HR): 1.65, 95% confidence interval (CI): 1.15 to 2.38, P = 0.007) and length of stay in ICU (HR: 0.63, 0.01, 95% CI: 1.24 to 0.045, P = 0.045). According to the subgroup analysis, WHR was found to be associated with in-hospital mortality in patients with higher APACHE score (HR: 1.60, 95% CI: 1.06 to 2.41, P = 0.024), in male patients (HR: 1.87, 95% CI: 1.15 to 3.04, P = 0.012), and in patients with the treatment of ventilation (HR: 2.33, 95% CI: 1.49 to 3.64, P < 0.001).

CONCLUSION

This study suggests the association between WHR and risk of in-hospital mortality in patients with lung cancer and length of stay, which indicates the importance of attention to WHR for patients with lung cancer.

Exploratory Descriptive Analysis of Smart Speaker Utilization in the Emergency Department During the COVID-19 Pandemic

BACKGROUND

In March 2020, the U.S. Department of Health and Human Services Office for Civil Rights stated that they would use discretion when enforcing the Health Insurance Portability and Accountability Act regarding remote communication technologies that promoted telehealth delivery during the COVID-19 pandemic. This was in an effort to protect patients, clinicians, and staff. More recently, smart speakers-voice-activated, hands-free devices-are being proposed as productivity tools within hospitals.

OBJECTIVE

We aimed to characterize the novel use of smart speakers in the emergency department (ED).

METHODS

A retrospective observational study of Amazon Echo Show® utilization from May 2020 to October 2020 in a large academic Northeast health system ED. Voice commands and queries were classified as either patient care-related or non-patient care-related, and then further subcategorized to explore the content of given commands.

RESULTS

Of 1232 commands analyzed, 200 (16.23%) were determined to be patient care-related. Of these commands, 155 (77.5%) were clinical in nature (i.e., "drop in on triage") and 23 (11.5%) were environment-enhancing commands (i.e., "play calming sounds"). Among non-patient care-related commands, 644 (62.4%) were for entertainment. Among all commands, 804 (65.3%) were during night-shift hours, which was statistically significant (p < 0.001).

CONCLUSIONS

Smart speakers showed notable engagement, primarily being used for patient communication and entertainment. Future studies should examine content of patient care conversations using these devices, effects on frontline staff wellbeing, productivity, patient satisfaction, and even explore opportunities for "smart" hospital rooms.

A Survey of Tele-Critical Care State and Needs in 2019 and 2020 Conducted among the Members of the Society of Critical Care Medicine

The study’s objective was to assess facilitators and barriers of Tele-Critical Care (TCC) perceived by SCCM members. By utilizing a survey distributed to SCCM members, a cross-sectional study was developed to analyze survey results from December 2019 and July 2020. SCCM members responded to the survey (n = 15,502) with a 1.9% response rate for the first distribution and a 2.54% response rate for the second survey (n = 9985). Participants (n = 286 and n = 254) were almost equally distributed between non-users, providers, users, and potential users of TCC services. The care delivery models for TCC were similar across most participants. Some consumers of TCC services preferred algorithmic coverage and scheduled rounds, while reactive and on-demand models were less utilized. The surveys revealed that outcome-driven measures were the principal form of TCC performance evaluation. A 1:100 (provider: patients) ratio was reported to be optimal. Factors related to costs, perceived lack of need for services, and workflow challenges were described by those who terminated TCC services. Barriers to implementation revolved around lack of reimbursement and adequate training. Interpersonal communication was identified as an essential TCC provider skill. The second survey introduced after the onset pandemic demonstrated more frequent use of advanced practice providers and focus on performance measures. Priorities for effective TCC deployment include communication, knowledge, optimal operationalization, and outcomes measurement at the organizational level. The potential effect of COVID-19 during the early stages of the pandemic on survey responses was limited and focused on the need to demonstrate TCC value.

Telehealth in emergency medicine: A consensus conference to map the intersection of telehealth and emergency medicine

INTRODUCTION

Telehealth has the potential to significantly change the specialty of emergency medicine (EM) and has rapidly expanded in EM during the COVID pandemic; however, it is unclear how EM should intersect with telehealth. The field lacks a unified research agenda with priorities for scientific questions on telehealth in EM.

METHODS

Through the 2020 Society for Academic Emergency Medicine's annual consensus conference, experts in EM and telehealth created a research agenda for the topic. The multiyear process used a modified Delphi technique to develop research questions related to telehealth in EM. Research questions were excluded from the final research agenda if they did not meet a threshold of at least 80% of votes indicating "important" or "very important."

RESULTS

Round 1 of voting included 94 research questions, expanded to 103 questions in round 2 and refined to 36 questions for the final vote. Consensus occurred with a final set of 24 important research questions spanning five breakout group topics. Each breakout group domain was represented in the final set of questions. Examples of the questions include: "Among underserved populations, what are mechanisms by which disparities in emergency care delivery may be exacerbated or ameliorated by telehealth" (health care access) and "In what situations should the quality and safety of telehealth be compared to in-person care and in what situations should it be compared to no care" (quality and safety).

CONCLUSION

The primary finding from the process was the breadth of gaps in the evidence for telehealth in EM and telehealth in general. Our consensus process identified priority research questions for the use of and evaluation of telehealth in EM to fill the current knowledge gaps. Support should be provided to answer the research questions to guide the evidenced-based development of telehealth in EM.

Characterizing the Patients, Hospitals, and Data Quality of the eICU Collaborative Research Database

OBJECTIVES

The eICU Collaborative Research Database is a publicly available repository of granular data from more than 200,000 ICU admissions. The quantity and variety of its entries hold promise for observational critical care research. We sought to understand better the data available within this resource to guide its future use.

DESIGN

We conducted a descriptive analysis of the eICU Collaborative Research Database, including patient, practitioner, and hospital characteristics. We investigated the completeness of demographic and hospital data, as well as those values required to calculate an Acute Physiology and Chronic Health Evaluation score. We also assessed the rates of ventilation, intubation, and dialysis, and looked for potential errors in the vital sign data.

SETTING

American ICUs that participated in the Philips Healthcare eICU program between 2014 and 2015.

PATIENTS

A total of 139,367 individuals who were admitted to one of the 335 participating ICUs between 2014 and 2015.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Most encounters were from small- and medium-sized hospitals, and managed by nonintensivists. The median ICU length of stay was 1.57 days (interquartile range, 0.82-2.97 d). The median Acute Physiology and Chronic Health Evaluation IV-predicted ICU mortality was 2.2%, with an observed mortality of 5.4%. Rates of ventilation (20-33%), intubation (15-24%), and dialysis (3-5%) varied according to the query method used. Most vital sign readings fell into realistic ranges, with manually curated data less likely to contain implausible results than automatically entered data.

CONCLUSIONS

Data in the eICU Collaborative Research Database are for the most part complete and plausible. Some ambiguity exists in determining which encounters are associated with various interventions, most notably mechanical ventilation. Caution is warranted in extrapolating findings from the eICU Collaborative Research Database to larger ICUs with higher acuity.

Incidence, predictors, and prognosis of thrombocytopenia among patients undergoing intra-aortic balloon pumping in the intensive care unit: a propensity score analysis

OBJECTIVE

To explore the incidence, predictors, and prognosis of intra-aortic balloon pumping (IABP)-related thrombocytopenia in critically ill patients.

METHODS

This multi-center study used the eICU Collaborative Research Database V1.2, comprising data on > 130,000 patients from multiple intensive care units (ICUs) in America between 2014 and 2015. A total of 710 patients undergoing IABP were included. Thrombocytopenia was defined as a drop in platelet count > 50% from baseline. From the cohort, 167 patients who developed thrombocytopenia were matched 1:1 with 167 patients who did not, after propensity score (PS) matching. The associations between IABP-related thrombocytopenia and clinical outcomes were examined by multivariable logistic regression.

RESULTS

Among 710 patients undergoing IABP, 249 patients (35.07%) developed thrombocytopenia. The APACHE IVa score was a predictor of thrombocytopenia [adjusted odds ratio (OR) = 1.09, 95% confidence interval (CI): 1.02-1.15]. After 1:1 PS matching, in-hospital mortality (adjusted OR = 0.76, 95% CI: 0.37-1.56) and in-ICU mortality (adjusted OR = 0.74, 95% CI: 0.34-1.63) were similar between the thrombocytopenia and non-thrombocytopenia groups. However, major bleeding occurred more frequently in the thrombocytopenia group (adjusted OR = 2.54, 95% CI: 1.54-4.17). In-hospital length of stay (LOS) and in-ICU LOS were significantly longer in patients who developed thrombocytopenia than in those who did not (9.71vs. 7.36, P < 0.001; 5.13 vs. 2.83, P < 0.001).

CONCLUSIONS

Among patients undergoing IABP in the ICUs, thrombocytopenia was not associated with a difference in in-hospital mortality or in-ICU mortality; however, thrombocytopenia was significantly associated with a greater risk of major bleeding and increased in-ICU and in-hospital LOS.

Predicting mortality of patients with acute kidney injury in the ICU using XGBoost model

PURPOSE

The goal of this study is to construct a mortality prediction model using the XGBoot (eXtreme Gradient Boosting) decision tree model for AKI (acute kidney injury) patients in the ICU (intensive care unit), and to compare its performance with that of three other machine learning models.

METHODS

We used the eICU Collaborative Research Database (eICU-CRD) for model development and performance comparison. The prediction performance of the XGBoot model was compared with the other three machine learning models. These models included LR (logistic regression), SVM (support vector machines), and RF (random forest). In the model comparison, the AUROC (area under receiver operating curve), accuracy, precision, recall, and F1 score were used to evaluate the predictive performance of each model.

RESULTS

A total of 7548 AKI patients were analyzed in this study. The overall in-hospital mortality of AKI patients was 16.35%. The best performing algorithm in this study was XGBoost with the highest AUROC (0.796, p < 0.01), F1(0.922, p < 0.01) and accuracy (0.860). The precision (0.860) and recall (0.994) of the XGBoost model rank second among the four models.

CONCLUSION

XGBoot model had obvious advantages of performance compared to the other machine learning models. This will be helpful for risk identification and early intervention for AKI patients at risk of death.

Explainable Machine Learning Model for Predicting GI Bleed Mortality in the Intensive Care Unit

INTRODUCTION

Acute gastrointestinal (GI) bleed is a common reason for hospitalization with 2%-10% risk of mortality. In this study, we developed a machine learning (ML) model to calculate the risk of mortality in intensive care unit patients admitted for GI bleed and compared it with APACHE IVa risk score. We used explainable ML methods to provide insight into the model's prediction and outcome.

METHODS

We analyzed the patient data in the Electronic Intensive Care Unit Collaborative Research Database and extracted data for 5,691 patients (mean age = 67.4 years; 61% men) admitted with GI bleed. The data were used in training a ML model to identify patients who died in the intensive care unit. We compared the predictive performance of the ML model with the APACHE IVa risk score. Performance was measured by area under receiver operating characteristic curve (AUC) analysis. This study also used explainable ML methods to provide insights into the model's outcome or prediction using the SHAP (SHapley Additive exPlanations) method.

RESULTS

The ML model performed better than the APACHE IVa risk score in correctly classifying the low-risk patients. The ML model had a specificity of 27% (95% confidence interval [CI]: 25-36) at a sensitivity of 100% compared with the APACHE IVa score, which had a specificity of 4% (95% CI: 3-31) at a sensitivity of 100%. The model identified patients who died with an AUC of 0.85 (95% CI: 0.80-0.90) in the internal validation set, whereas the APACHE IVa clinical scoring systems identified patients who died with AUC values of 0.80 (95% CI: 0.73-0.86) with P value <0.001.

DISCUSSION

We developed a ML model that predicts the mortality in patients with GI bleed with a greater accuracy than the current scoring system. By making the ML model explainable, clinicians would be able to better understand the reasoning behind the outcome.

Decompensation in Critical Care: Early Prediction of Acute Heart Failure Onset

BACKGROUND

Heart failure is a leading cause of mortality and morbidity worldwide. Acute heart failure, broadly defined as rapid onset of new or worsening signs and symptoms of heart failure, often requires hospitalization and admission to the intensive care unit (ICU). This acute condition is highly heterogeneous and less well-understood as compared to chronic heart failure. The ICU, through detailed and continuously monitored patient data, provides an opportunity to retrospectively analyze decompensation and heart failure to evaluate physiological states and patient outcomes.

OBJECTIVE

The goal of this study is to examine the prevalence of cardiovascular risk factors among those admitted to ICUs and to evaluate combinations of clinical features that are predictive of decompensation events, such as the onset of acute heart failure, using machine learning techniques. To accomplish this objective, we leveraged tele-ICU data from over 200 hospitals across the United States.

METHODS

We evaluated the feasibility of predicting decompensation soon after ICU admission for 26,534 patients admitted without a history of heart failure with specific heart failure risk factors (ie, coronary artery disease, hypertension, and myocardial infarction) and 96,350 patients admitted without risk factors using remotely monitored laboratory, vital signs, and discrete physiological measurements. Multivariate logistic regression and random forest models were applied to predict decompensation and highlight important features from combinations of model inputs from dissimilar data.

RESULTS

The most prevalent risk factor in our data set was hypertension, although most patients diagnosed with heart failure were admitted to the ICU without a risk factor. The highest heart failure prediction accuracy was 0.951, and the highest area under the receiver operating characteristic curve was 0.9503 with random forest and combined vital signs, laboratory values, and discrete physiological measurements. Random forest feature importance also highlighted combinations of several discrete physiological features and laboratory measures as most indicative of decompensation. Timeline analysis of aggregate vital signs revealed a point of diminishing returns where additional vital signs data did not continue to improve results.

CONCLUSIONS

Heart failure risk factors are common in tele-ICU data, although most patients that are diagnosed with heart failure later in an ICU stay presented without risk factors making a prediction of decompensation critical. Decompensation was predicted with reasonable accuracy using tele-ICU data, and optimal data extraction for time series vital signs data was identified near a 200-minute window size. Overall, results suggest combinations of laboratory measurements and vital signs are viable for early and continuous prediction of patient decompensation.

Benchmarking machine learning models on multi-centre eICU critical care dataset

Progress of machine learning in critical care has been difficult to track, in part due to absence of public benchmarks. Other fields of research (such as computer vision and natural language processing) have established various competitions and public benchmarks. Recent availability of large clinical datasets has enabled the possibility of establishing public benchmarks. Taking advantage of this opportunity, we propose a public benchmark suite to address four areas of critical care, namely mortality prediction, estimation of length of stay, patient phenotyping and risk of decompensation. We define each task and compare the performance of both clinical models as well as baseline and deep learning models using eICU critical care dataset of around 73,000 patients. This is the first public benchmark on a multi-centre critical care dataset, comparing the performance of clinical gold standard with our predictive model. We also investigate the impact of numerical variables as well as handling of categorical variables on each of the defined tasks. The source code, detailing our methods and experiments is publicly available such that anyone can replicate our results and build upon our work.

Rule-Based Cohort Definitions for Acute Respiratory Failure: Electronic Phenotyping Algorithm

Background

Acute respiratory failure is generally treated with invasive mechanical ventilation or noninvasive respiratory support strategies. The efficacies of the various strategies are not fully understood. There is a need for accurate therapy-based phenotyping for secondary analyses of electronic health record data to answer research questions regarding respiratory management and outcomes with each strategy.

Objective

The objective of this study was to address knowledge gaps related to ventilation therapy strategies across diverse patient populations by developing an algorithm for accurate identification of patients with acute respiratory failure. To accomplish this objective, our goal was to develop rule-based computable phenotypes for patients with acute respiratory failure using remotely monitored intensive care unit (tele-ICU) data. This approach permits analyses by ventilation strategy across broad patient populations of interest with the ability to sub-phenotype as research questions require.

Methods

Tele-ICU data from ≥200 hospitals were used to create a rule-based algorithm for phenotyping patients with acute respiratory failure, defined as an adult patient requiring invasive mechanical ventilation or a noninvasive strategy. The dataset spans a wide range of hospitals and ICU types across all US regions. Structured clinical data, including ventilation therapy start and stop times, medication records, and nurse and respiratory therapy charts, were used to define clinical phenotypes. All adult patients of any diagnoses with record of ventilation therapy were included. Patients were categorized by ventilation type, and analysis of event sequences using record timestamps defined each phenotype. Manual validation was performed on 5% of patients in each phenotype.

Results

We developed 7 phenotypes: (0) invasive mechanical ventilation, (1) noninvasive positive-pressure ventilation, (2) high-flow nasal insufflation, (3) noninvasive positive-pressure ventilation subsequently requiring intubation, (4) high-flow nasal insufflation subsequently requiring intubation, (5) invasive mechanical ventilation with extubation to noninvasive positive-pressure ventilation, and (6) invasive mechanical ventilation with extubation to high-flow nasal insufflation. A total of 27,734 patients met our phenotype criteria and were categorized into these ventilation subgroups. Manual validation of a random selection of 5% of records from each phenotype resulted in a total accuracy of 88% and a precision and recall of 0.8789 and 0.8785, respectively, across all phenotypes. Individual phenotype validation showed that the algorithm categorizes patients particularly well but has challenges with patients that require ≥2 management strategies.

Conclusions

Our proposed computable phenotyping algorithm for patients with acute respiratory failure effectively identifies patients for therapy-focused research regardless of admission diagnosis or comorbidities and allows for management strategy comparisons across populations of interest.