ABSTRACT
The goal of this work is to develop a Machine Learning model to predict the need for both invasive and non-invasive mechanical ventilation in intensive care unit (ICU) patients. Using the Philips eICU Research Institute (ERI) database, 2.6 million ICU patient data from 2010 to 2019 were analyzed. This data was randomly split into training (63%), validation (27%), and test (10%) sets. Additionally, an external test set from a single hospital from the ERI database was employed to assess the model's generalizability. Model performance was determined by comparing the model probability predictions with the actual incidence of ventilation use, either invasive or non-invasive. The model demonstrated a prediction performance with an AUC of 0.921 for overall ventilation, 0.937 for invasive, and 0.827 for non-invasive. Factors such as high Glasgow Coma Scores, younger age, lower BMI, and lower PaCO2 were highlighted as indicators of a lower likelihood for the need for ventilation. The model can serve as a retrospective benchmarking tool for hospitals to assess ICU performance concerning mechanical ventilation necessity. It also enables analysis of ventilation strategy trends and risk-adjusted comparisons, with potential for future testing as a clinical decision tool for optimizing ICU ventilation management.
ABSTRACT
Background: Intracerebral hemorrhage (ICH) is a serious condition characterized by bleeding within the brain tissue. Although the use of sildenafil, a vasodilator agent for erectile dysfunction, has been associated with rare cases of ICH, the combination of sildenafil usage and smoking as risk factors for ICH has not yet been reported. This case report describes the occurrence of ICH in a patient with a history of both sildenafil usage and heavy smoking. Case Presentation: A 53-year-old male, with a history of smoking and regular sildenafil use, was brought to the emergency department due to loss of consciousness with right-side weakness, he initially experienced with nausea, vomiting and dizziness after taking sildenafil 100mg tablet once. The Glasgow Coma Score (GCS) was 10 with side hemiparesis. Non-contrast CT revealed left thalamic acute hemorrhage with ventricular extension. Furthermore, a head CT angiography ruled out any vascular anomalies after that the patient was admitted to the intensive care unit (ICU) for conservative management. After three days on clinical and neurological improvement, the patient was transferred to the inpatient ward for further management, monitoring and physiotherapy. On day 6, the patient was discharged and planned for flow up. Conclusion: This rare case highlights the need for further research and awareness regarding the potential risks associated with the combination of sildenafil and heavy smoking. Healthcare professionals should carefully evaluate the individual risk factors of patients, educate them about potential complications, and consider alternative treatments if necessary. Additionally, patients should be encouraged to quit smoking and adopt a healthy lifestyle to minimize the risk of cerebrovascular events.
Subject(s)
Cigarette Smoking , Erectile Dysfunction , Male , Humans , Middle Aged , Sildenafil Citrate/adverse effects , Cerebral Hemorrhage/chemically induced , Cerebral Hemorrhage/diagnostic imaging , Risk FactorsABSTRACT
Predicting the duration of ventilation in the ICU helps in assessing the risk of ventilator-induced lung injury, ensuring sufficient oxygenation, and optimizing resource allocation. Prior models provided a prediction of total duration without distinguishing between invasive and non-invasive ventilation. This work proposes two independent gradient boosting regression models for predicting the duration of invasive and non-invasive ventilation based on commonly available ICU features. These models are trained on 2.6 million patient stays across 350 US hospitals between 2010 to 2019. The mean absolute error (MAE) for the prediction of duration was 2.08 days for invasive ventilation and 0.36 days for non-invasive ventilation. The total ventilation duration predicted by our model had MAE of 2.38 days, which outperformed the gold standard (APACHE) with MAE of 3.02 days. The feature importance analysis of the trained models showed that, for invasive ventilation, high average heart rate, diagnosis of respiratory infection and admissions from locations other than the operating room were associated with longer ventilation durations. For non-invasive ventilation, higher respiratory rates and having any GCS measurement were associated with longer durations.
ABSTRACT
BACKGROUND: Institutions struggle with successful use of sepsis alerts within electronic health records. OBJECTIVE: Test the association of sepsis screening measurement criteria in discrimination of mortality and detection of sepsis in a large dataset. DESIGN: Retrospective, cohort study using a large United States (U.S.) intensive care database. The Institutional Review Board exempt status was obtained from Kansas University Medical Center Human Research Protection Program (10-1-2015). SETTING: 334 U.S. hospitals participating in the eICU Research Institute. PARTICIPANTS: Nine hundred twelve thousand five hundred and nine adult intensive care admissions from 183 hospitals. METHODS: Exposures included: systemic inflammatory response syndrome criteriaâ¯≥â¯2 (Sepsis-1); systemic inflammatory response syndrome criteria with organ failure criteriaâ¯≥â¯3.5 points (Sepsis-2); and sepsis-related organ failure assessment scoreâ¯≥â¯2 and quick scoreâ¯≥â¯2 (Sepsis-3). Discrimination of outcomes was determined with/without (adjusted/unadjusted) baseline risk exposure to a model. The receiver operating characteristic curve (AUROC) and odds ratios (ORs) for each decile of baseline risk of sepsis or death were assessed. RESULTS: Within the eligible cohort of 912,509, a total of 86,219 (9.4â¯%) patients did not survive their hospital stay and 186,870 (20.5â¯%) met the definition of suspected sepsis. For suspected sepsis discrimination, Sepsis-2 (unadjusted AUROC 0.67, 99â¯% CI: 0.66-0.67 and adjusted AUROC 0.77, 99â¯% CI: 0.77-0.77) outperformed Sepsis-3 (SOFA unadjusted AUROC 0.61, 99â¯% CI: 0.61-0.61 and adjusted AUROC 0.74, 99â¯% CI: 0.74-0.74) (qSOFA unadjusted AUROC 0.59, 99â¯% CI: 0.59-0.60 and adjusted AUROC 0.73, 99â¯% CI: 0.73-0.73). Sepsis-2 also outperformed Sepsis-1 (unadjusted AUROC 0.58, 99â¯% CI: 0.58-0.58 and adjusted AUROC 0.73, 99â¯% CI: 0.73-0.73). In between differences of AUROCs were statistically significantly different. Sepsis-2 ORs were higher for the outcome of suspected sepsis when considering deciles of risk than the other measurement systems. CONCLUSIONS AND RELEVANCE: Sepsis-2 outperformed other systems in suspected sepsis detection and was comparable to SOFA in prognostic accuracy of mortality in adult intensive care patients.
Subject(s)
Sepsis , Humans , Adult , Cohort Studies , Retrospective Studies , Hospital Mortality , Sepsis/diagnosis , Intensive Care Units , Prognosis , ROC CurveABSTRACT
Introduction: The Society of Critical Care Medicine Tele-Critical Care (TCC) Committee has identified the need for rigorous comparative research of different TCC delivery models to support the development of best practices for staffing, application, and approaches to workflow. Our objective was to describe and compare outcomes between two TCC delivery models, TCC with 24/7 Bedside Intensivist (BI) compared with TCC with Private Daytime Attending Intensivist (PI) in relation to intensive care unit (ICU) and hospital mortality, ICU and hospital length of stay (LOS), cost, and complications across the spectrum of routine ICU standards of care. Methods: Observational cohort study at large health care system in 12 ICUs and included patients, ≥18, with Acute Physiology and Chronic Health Evaluation (APACHE) IVa scores and predictions (October 2016-June 2019). Results: Of the 19,519 ICU patients, 71.7% (n = 13,993) received TCC with 24/7 BI while 28.3% (n = 5,526) received TCC with PI. ICU and Hospital mortality (4.8% vs. 3.1%, p < 0.0001; 12.6% vs. 8.1%, p < 0.001); and ICU and Hospital LOS (3.2 vs. 2.4 days, p < 0.001; 9.8 vs. 7.2 days, p < 0.001) were significantly higher among 24/7 BI compared with PI. The APACHE observed/expected ratios (odds ratio [OR]; 95% confidence interval [CI]) for ICU mortality (0.62; 0.58-0.67) vs. (0.53; 0.46-0.61) and Hospital mortality (0.95; 0.57-1.48) vs. (0.77; 0.70-0.84) were significantly different for 24/7 BI compared with PI. Multivariate mixed models that adjusted for confounders demonstrated significantly greater odds of (OR; 95% CI) ICU mortality (1.58; 1.28-1.93), Hospital mortality (1.52; 1.33-1.73), complications (1.55; 1.18-2.04), ICU LOS [3.14 vs. 2.59 (1.25; 1.19-1.51)], and Hospital LOS [9.05 vs. 7.31 (1.23; 1.21-1.25)] among 24/7 BI when compared with PI. Sensitivity analyses adjusting for ICU admission within 24 h of hospital admission, receiving active ICU treatments, nighttime admission, sepsis, and highest third acute physiology score indicated significantly higher odds for 24/7 BI compared with PI. Conclusion: Our comparison demonstrated that TCC delivery model with PI provided high-quality care with significant positive effects on outcomes. This suggests that TCC delivery models have broad-ranging applicability and benefits in routine critical care, thus necessitating progressive research in this direction.
Subject(s)
Critical Care , Intensive Care Units , Humans , Cohort Studies , Length of Stay , Hospital Mortality , Delivery of Health Care , Hospitals , Retrospective StudiesABSTRACT
OBJECTIVES: Electronic health records enable automated data capture for risk models but may introduce bias. We present the Philips Critical Care Outcome Prediction Model (CCOPM) focused on addressing model features sensitive to data drift to improve benchmarking ICUs on mortality performance. DESIGN: Retrospective, multicenter study of ICU patients randomized in 3:2 fashion into development and validation cohorts. Generalized additive models (GAM) with features designed to mitigate biases introduced from documentation of admission diagnosis, Glasgow Coma Scale (GCS), and extreme vital signs were developed using clinical features representing the first 24 hours of ICU admission. SETTING: eICU Research Institute database derived from ICUs participating in the Philips eICU telecritical care program. PATIENTS: A total of 572,985 adult ICU stays discharged from the hospital between January 1, 2017, and December 31, 2018, were included, yielding 509,586 stays in the final cohort; 305,590 and 203,996 in development and validation cohorts, respectively. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Model discrimination was compared against Acute Physiology and Chronic Health Evaluation (APACHE) IVa/IVb models on the validation cohort using the area under the receiver operating characteristic (AUROC) curve. Calibration assessed by actual/predicted ratios, calibration-in-the-large statistics, and visual analysis. Performance metrics were further stratified by subgroups of admission diagnosis and ICU characteristics. Historic data from two health systems with abrupt changes in Glasgow Coma Scale (GCS) documentation were assessed in the year prior to and after data shift. CCOPM outperformed APACHE IVa/IVb for ICU mortality (AUROC, 0.925 vs 0.88) and hospital mortality (AUROC, 0.90 vs 0.86). Better calibration performance was also attained among subgroups of different admission diagnoses, ICU types, and over unique ICU-years. The CCOPM provided more stable predictions compared with APACHE IVa within an external cohort of greater than 120,000 patients from two health systems with known changes in GCS documentation. CONCLUSIONS: These mortality risk models demonstrated excellent performance compared with APACHE while appearing to mitigate bias introduced through major shifts in GCS documentation at two large health systems. This provides evidence to support using automated capture rather than trained personnel for capture of GCS data used in benchmarking ICUs on mortality performance.
Subject(s)
Intensive Care Units , Adult , Humans , Retrospective Studies , APACHE , Hospital Mortality , Bias , AutomationABSTRACT
BACKGROUND: As more plastic surgery clinicians pursue advanced degrees and strive to become stronger physician-scientists, an objective understanding of how such degrees influence careers becomes important. We hypothesized that having a master's degree is associated with higher scholarly activity, research funding, academic progression, and leadership appointments. METHODS: Accreditation Council for Graduate Medical Education-accredited integrated plastic surgery residency program Web sites were queried to create a data set of current academic plastic surgeons (APSs) and plastic surgery residents (PSRs). Scholarly metrics such as publications, citations, and H-indices were extracted from the Scopus database. National Institutes of Health and Plastic Surgery Foundation funding information was collected through their respective Web sites. RESULTS: Our cohort comprised 799 APSs and 922 PSRs, of whom 8% and 7.4%, respectively, had at least one master's degree. Academic plastic surgeons with master's of public health degrees had a significantly higher median number of publications and citations than APSs without a master's of public health. There was no association between any master's degree and academic rank or being a department chairman or program director. Academic plastic surgeons with master of science degrees were more likely to receive National Institutes of Health grants. Among PSRs, master's of science graduates had a higher median number of publications. Other master's degrees did not significantly influence scholarly productivity or funding. CONCLUSIONS: Certain master's degrees had an impact on scholarly productivity, with no significant effect on academic rank or leadership positions. The value of master's degrees in programs focusing on healthcare management, leadership skills, and business acumen likely extends beyond the scope of this study.
Subject(s)
Surgeons , Surgery, Plastic , United States , Humans , National Institutes of Health (U.S.) , Efficiency , BibliometricsABSTRACT
OBJECTIVES: To develop and demonstrate the feasibility of a Global Open Source Severity of Illness Score (GOSSIS)-1 for critical care patients, which generalizes across healthcare systems and countries. DESIGN: A merger of several critical care multicenter cohorts derived from registry and electronic health record data. Data were split into training (70%) and test (30%) sets, using each set exclusively for development and evaluation, respectively. Missing data were imputed when not available. SETTING/PATIENTS: Two large multicenter datasets from Australia and New Zealand (Australian and New Zealand Intensive Care Society Adult Patient Database [ANZICS-APD]) and the United States (eICU Collaborative Research Database [eICU-CRD]) representing 249,229 and 131,051 patients, respectively. ANZICS-APD and eICU-CRD contributed data from 162 and 204 hospitals, respectively. The cohort included all ICU admissions discharged in 2014-2015, excluding patients less than 16 years old, admissions less than 6 hours, and those with a previous ICU stay. INTERVENTIONS: Not applicable. MEASUREMENTS AND MAIN RESULTS: GOSSIS-1 uses data collected during the ICU stay's first 24 hours, including extrema values for vital signs and laboratory results, admission diagnosis, the Glasgow Coma Scale, chronic comorbidities, and admission/demographic variables. The datasets showed significant variation in admission-related variables, case-mix, and average physiologic state. Despite this heterogeneity, test set discrimination of GOSSIS-1 was high (area under the receiver operator characteristic curve [AUROC], 0.918; 95% CI, 0.915-0.921) and calibration was excellent (standardized mortality ratio [SMR], 0.986; 95% CI, 0.966-1.005; Brier score, 0.050). Performance was held within ANZICS-APD (AUROC, 0.925; SMR, 0.982; Brier score, 0.047) and eICU-CRD (AUROC, 0.904; SMR, 0.992; Brier score, 0.055). Compared with GOSSIS-1, Acute Physiology and Chronic Health Evaluation (APACHE)-IIIj (ANZICS-APD) and APACHE-IVa (eICU-CRD), had worse discrimination with AUROCs of 0.904 and 0.869, and poorer calibration with SMRs of 0.594 and 0.770, and Brier scores of 0.059 and 0.063, respectively. CONCLUSIONS: GOSSIS-1 is a modern, free, open-source inhospital mortality prediction algorithm for critical care patients, achieving excellent discrimination and calibration across three countries.
Subject(s)
Critical Care , Intensive Care Units , APACHE , Adolescent , Adult , Australia , Hospital Mortality , HumansABSTRACT
OBJECTIVE: Identify opportunities to improve the interaction between clinicians and Tele-Critical Care (Tele-CC) programs through an analysis of alert occurrence and reactivation in a specific Tele-CC application. MATERIALS AND METHODS: Data were collected automatically through the Philips eCaremanager® software system used at multiple hospitals in the Avera health system. We evaluated the distribution of alerts per patient, frequency of alert types, time between consecutive alerts, and Tele-CC clinician choice of alert reactivation times. RESULTS: Each patient generated an average of 79.8 alerts during their ICU stay (median 31.0; 25th - 75th percentile 10.0-89.0) with 46.4 for blood pressure and 38.4 for oxygenation. The most frequent alerts for continuous physiological parameters were: MAP limit (28.9 %), O2/RR (26.4 %), MAP trend (16.5 %), HR trend (12.1 %), and HR limit (11.3 %). The median time between consecutive alerts for one parameter was less than 10 min for 86 % of patients. Tele-CC providers responded to all alert types with immediate reactivation 47-88 % of the time. Limit alerts had longer reactivation times than their trend alert counterparts (p-value < .001). CONCLUSIONS: The alert type specific differences in frequency, time occurrence and provider choice of reactivation time provide insight into how clinicians interact with the Tele-CC system. Systems engineering enhancements to Tele-CC software algorithms may reduce alert burden and thereby decrease clinicians' cognitive workload for alert assessment. Further study of Tele-CC alert generation, alert presentation to clinicians, and the clinicians' options to respond to these alerts may reduce provider workload, minimize alert desensitization, and optimize the ability of Tele-CC clinicians to provide efficient and timely critical care management.
Subject(s)
Critical Care/methods , Decision Support Systems, Clinical/standards , Medical Order Entry Systems/standards , Telemedicine/methods , Workload/standards , Critical Care/trends , Humans , Telemedicine/trendsABSTRACT
BACKGROUND: Admission to high-acuity ICUs has been associated with improved outcomes compared with outcomes in low-acuity ICUs, although the mechanism for these findings is unclear. RESEARCH QUESTION: The goal of this study was to determine if high-acuity ICUs more effectively implement evidence-based processes of care that have been associated with improved clinical outcomes. STUDY DESIGN AND METHODS: This retrospective cohort study was performed in adult ICU patients admitted to 322 ICUs in 199 hospitals in the Philips ICU telemedicine database between 2010 and 2015. The primary exposure was ICU acuity, defined as the mean Acute Physiology and Chronic Health Evaluation IVa score of all admitted patients in a calendar year, stratified into quartiles. Multivariable logistic regression was used to examine relations of ICU acuity with adherence to evidence-based VTE and stress ulcer prophylaxis, and with the avoidance of potentially harmful events. These events included hypoglycemia, sustained hyperglycemia, and liberal transfusion practices (defined as RBC transfusions prescribed for nonbleeding patients with preceding hemoglobin levels ≥ 7 g/dL). RESULTS: Among 1,058,510 ICU admissions, adherence to VTE and stress ulcer prophylaxis was high across acuity levels. In adjusted analyses, those admitted to low-acuity ICUs compared with the highest acuity ICUs were more likely to experience hypoglycemic events (adjusted OR [aOR], 1.12; 95% CI, 1.04-1.19), sustained hyperglycemia (aOR, 1.07; 95% CI, 1.04-1.10), and liberal transfusion practices (aOR, 1.55; 95% CI, 1.33-1.82). INTERPRETATION: High-acuity ICUs were associated with better adherence to several evidence-based practices, which may be a marker of high-quality care. Future research should investigate how high-acuity ICUs approach ICU organization to identify targets for improving the quality of critical care across all ICU acuity levels.
Subject(s)
Critical Care , Guideline Adherence , Patient Acuity , Aged , Blood Transfusion , Female , Hospital Mortality , Humans , Hyperglycemia/prevention & control , Hypoglycemia/prevention & control , Male , Middle Aged , Practice Guidelines as Topic , Pressure Ulcer/prevention & control , Retrospective StudiesABSTRACT
Rationale: Whether critical care improvements over the last 10 years extend to all hospitals has not been described.Objectives: To examine the temporal trends of critical care outcomes in minority and non-minority-serving hospitals using an inception cohort of critically ill patients.Measurements and Main Results: Using the Philips Health Care electronic ICU Research Institute Database, we identified minority-serving hospitals as those with an African American or Hispanic ICU census more than twice its regional mean. We examined almost 1.1 million critical illness admissions among 208 ICUs from across the United States admitted between 2006 and 2016. Adjusted hospital mortality (primary) and length of hospitalization (secondary) were the main outcomes. Large pluralities of African American (25%, n = 27,242) and Hispanic individuals (48%, n = 26,743) were cared for in minority-serving hospitals, compared with only 5.2% (n = 42,941) of white individuals. Over the last 10 years, although the risk of critical illness mortality steadily decreased by 2% per year (95% confidence interval [CI], 0.97-0.98) in non-minority-serving hospitals, outcomes within minority-serving hospitals did not improve comparably. This disparity in temporal trends was particularly noticeable among African American individuals, where each additional calendar year was associated with a 3% (95% CI, 0.96-0.97) lower adjusted critical illness mortality within a non-minority-serving hospital, but no change within minority-serving hospitals (hazard ratio, 0.99; 95% CI, 0.97-1.01). Similarly, although ICU and hospital lengths of stay decreased by 0.08 (95% CI, -0.08 to -0.07) and 0.16 (95% CI, -0.16 to -0.15) days per additional calendar year, respectively, in non-minority-serving hospitals, there was little temporal change for African American individuals in minority-serving hospitals.Conclusions: Critically ill African American individuals are disproportionately cared for in minority-serving hospitals, which have shown significantly less improvement than non-minority-serving hospitals over the last 10 years.
Subject(s)
Black or African American/statistics & numerical data , Critical Care/statistics & numerical data , Critical Care/trends , Hispanic or Latino/statistics & numerical data , Hospitals/statistics & numerical data , Minority Groups/statistics & numerical data , White People/statistics & numerical data , Adult , Aged , Aged, 80 and over , Critical Care Outcomes , Female , Hospitals/trends , Humans , Male , Middle Aged , United StatesABSTRACT
OBJECTIVE: To determine whether patients transfused red blood cell (RBC) products according to guideline-specified pretransfusion hemoglobin (Hb) concentrations or for other reasons were more likely to survive their intensive care unit (ICU) stay. DESIGN: An observational study of 375 478 episodes of ICU care, over 5 years, was performed with ICU survival as the primary outcome. Outcomes were analyzed as a function of pretransfusion Hb concentration for groups with distinct transfusion indications while adjusting for potential confounders. SETTING AND PATIENTS: This study included all adult patients discharged from 1 of 203 adult ICUs from 32 US health-care systems. The patients were from community hospitals, tertiary, and academic medical centers. INTERVENTION: Transfusion of allogenic packed RBCs or whole blood was prescribed at the discretion of the treating clinicians. MEASUREMENTS AND MAIN RESULTS: We found that 15% of adult ICU patients are transfused RBC products, and most transfusions for hemodynamically stable patients are administered above the guideline-specified pretransfusion Hb threshold of 7 g/dL. Hemodynamically stable patients transfused below this threshold were significantly more likely to survive their ICU stay than those not transfused (odds ratio [OR] 0.59, 95% confidence interval [CI], 0.43-0.81; P = .001), and patients transfused at thresholds above 9 g/dL were less likely to survive their ICU stay than those not transfused. Patients of the acute blood loss group who were transfused appeared to benefit or were not harmed by transfusion. CONCLUSION: Conservative RBC product transfusion practices for groups that are targeted by guidelines are justified by outcomes observed in clinical practice. This study provides evidence for the liberal administration of RBC products to critically ill adults with acute blood loss based on association with lower risk of mortality.
Subject(s)
Critical Care Outcomes , Erythrocyte Transfusion/mortality , Guideline Adherence/statistics & numerical data , Hemostatic Techniques/mortality , Intensive Care Units/statistics & numerical data , Aged , Critical Illness/therapy , Erythrocyte Transfusion/standards , Female , Hemoglobins/analysis , Hemostatic Techniques/standards , Humans , Length of Stay/statistics & numerical data , Male , Middle Aged , Patient Discharge/statistics & numerical dataABSTRACT
Objective: To compare the characteristics, use of invasive ventilation and outcomes of patients admitted with critical asthma syndrome (CAS) to ICUs in Australia and New Zealand (ANZ), and a large cohort of ICUs in the United States (US). Methods: We examined two large databases of ICU for patients admitted with CAS in 2014 and 2015. We obtained, analyzed, and compared information on demographic and physiological characteristics, use of invasive mechanical ventilation, and clinical outcome and derived predictive models. Results: Overall, 2202 and 762 patients were admitted with a primary diagnosis of CAS in the ANZ and US databases respectively (0.73% vs. 0.46% of all ICU admissions, P < 0.001). A similar percentage of patients received invasive mechanical ventilation in the first 24 h (24.7% vs. 24.4%, P = 0.87) but ANZ patients had lower respiratory rates and higher PaCO2 levels. Overall mortality was low (1.23 for ANZ and 1.71 for USA; P = 0.36) and even among invasively ventilated patients (2.4% for ANZ vs. 1.1% for USA; P = 0.38). However, ANZ patients also had longer length of stay in ICU (43 vs. 37 h, P = 0.001) and hospital (105 vs. 78 h, P = 0.003). Conclusions: Patients admitted to ANZ and USA ICU with CAS are broadly similar and have a low and similar rate of invasive ventilation and mortality. However, ANZ patients made up a greater proportion of ICU patients and had longer ICU and hospital stays. These findings provide a modern invasive ventilation and mortality rates benchmark for future studies of CAS.
Subject(s)
Asthma/therapy , Cross-Cultural Comparison , Intensive Care Units/statistics & numerical data , Adult , Asthma/mortality , Australia/epidemiology , Cohort Studies , Databases, Factual/statistics & numerical data , Female , Hospital Mortality , Humans , Length of Stay/statistics & numerical data , Male , Middle Aged , New Zealand/epidemiology , Respiration, Artificial/statistics & numerical data , Treatment Outcome , United States/epidemiologyABSTRACT
Illness severity scores are regularly employed for quality improvement and benchmarking in the intensive care unit, but poor generalization performance, particularly with respect to probability calibration, has limited their use for decision support. These models tend to perform worse in patients at a high risk for mortality. We hypothesized that a sequential modeling approach wherein an initial regression model assigns risk and all patients deemed high risk then have their risk quantified by a second, high-risk-specific, regression model would result in a model with superior calibration across the risk spectrum. We compared this approach to a logistic regression model and a sophisticated machine learning approach, the gradient boosting machine. The sequential approach did not have an effect on the receiver operating characteristic curve or the precision-recall curve but resulted in improved reliability curves. The gradient boosting machine achieved a small improvement in discrimination performance and was similarly calibrated to the sequential models.
ABSTRACT
This article examines the history of the telemedicine intensive care unit (tele-ICU), the current state of clinical decision support systems (CDSS) in the tele-ICU, applications of machine learning (ML) algorithms to critical care, and opportunities to integrate ML with tele-ICU CDSS. The enormous quantities of data generated by tele-ICU systems is a major driver in the development of the large, comprehensive, heterogeneous, and granular data sets necessary to develop generalizable ML CDSS algorithms, and deidentification of these data sets expands opportunities for ML CDSS research.
Subject(s)
Artificial Intelligence , Big Data , Decision Support Systems, Clinical , Intensive Care Units , Telemedicine , HumansABSTRACT
Sepsis is the third leading cause of death worldwide and the main cause of mortality in hospitals1-3, but the best treatment strategy remains uncertain. In particular, evidence suggests that current practices in the administration of intravenous fluids and vasopressors are suboptimal and likely induce harm in a proportion of patients1,4-6. To tackle this sequential decision-making problem, we developed a reinforcement learning agent, the Artificial Intelligence (AI) Clinician, which extracted implicit knowledge from an amount of patient data that exceeds by many-fold the life-time experience of human clinicians and learned optimal treatment by analyzing a myriad of (mostly suboptimal) treatment decisions. We demonstrate that the value of the AI Clinician's selected treatment is on average reliably higher than human clinicians. In a large validation cohort independent of the training data, mortality was lowest in patients for whom clinicians' actual doses matched the AI decisions. Our model provides individualized and clinically interpretable treatment decisions for sepsis that could improve patient outcomes.
Subject(s)
Artificial Intelligence/standards , Clinical Decision-Making , Sepsis/drug therapy , Software , Administration, Intravenous/adverse effects , Artificial Intelligence/trends , Cohort Studies , Critical Care/methods , Female , Humans , Learning , Male , Sepsis/mortality , Sepsis/pathology , Vasoconstrictor Agents/adverse effects , Vasoconstrictor Agents/therapeutic useABSTRACT
Critical care patients are monitored closely through the course of their illness. As a result of this monitoring, large amounts of data are routinely collected for these patients. Philips Healthcare has developed a telehealth system, the eICU Program, which leverages these data to support management of critically ill patients. Here we describe the eICU Collaborative Research Database, a multi-center intensive care unit (ICU)database with high granularity data for over 200,000 admissions to ICUs monitored by eICU Programs across the United States. The database is deidentified, and includes vital sign measurements, care plan documentation, severity of illness measures, diagnosis information, treatment information, and more. Data are publicly available after registration, including completion of a training course in research with human subjects and signing of a data use agreement mandating responsible handling of the data and adhering to the principle of collaborative research. The freely available nature of the data will support a number of applications including the development of machine learning algorithms, decision support tools, and clinical research.