Dataset Information

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

ABSTRACT: 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.

SUBMITTER: Essay P

PROVIDER: S-EPMC7442938 | biostudies-literature | 2020 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

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

Essay Patrick P Balkan Baran B Subbian Vignesh V

JMIR medical informatics 20200807 8

<h4>Background</h4>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 anal ...[more]

PMID: 32663162

Similar Datasets

Project description:AimsTachycardia is a reversible event that may cause hemodynamic decompensation but may not necessarily cause direct damages to the myocardium. To evaluate the clinical outcomes of patients with heart failure (HF) and atrial fibrillation (AF), whose acute decompensation was tachycardia mediated.Methods and resultsThe Korean Acute Heart Failure registry was a prospective registry that consecutively enrolled 5625 patients with acute HF. Patients were classified into three groups according to the rhythm and aggravating factor: (i) 3664 (65.1%) patients with sinus rhythm (SR), (ii) 1033 (18.4%) patients with AF whose decompensation was tachycardia-mediated, AF-TM (+), and (iii) N = 928 (16.5%) patients with AF whose decompensation was not tachycardia-mediated, AF-TM (-). The primary outcomes were in-hospital and post-discharge 1 year all-cause mortality. At admission, the mean heart rate was 90.8 ± 23.4, 86.8 ± 26.8, and 106.3 ± 29.7 beats per minute for the SR, AF-TM (-), and AF-TM (+) groups, respectively. The AF-TM (+) group had more favourable characteristics such as de novo onset HF, less diabetes, ischaemic heart disease, and higher blood pressure than the AF-TM (-) group. In-hospital mortality rates were 5.1%, 6.5%, and 1.7% for SR, AF-TM (-), and AF-TM (+) groups, respectively. In logistic regression analysis, the AF-TM (+) group had lower in-hospital mortality after adjusting the significant covariates (odds ratio, 0.49; 95% confidence interval, 0.26-0.93). The mortality rate did not differ between SR and AF-TM (-) groups. During 1 year follow-up, 990 (18.5%) patients died. In univariate and multivariate Cox proportional regression analyses, there was no difference in 1-year all-cause mortality between the three groups.ConclusionsIn patients with HF and AF, patients whose acute decompensation is tachycardia-mediated have better in-hospital, but similar post-discharge outcomes compared with those with SR or those with AF whose decompensation is not tachycardia-mediated.Clinical trial registrationClinicalTrial.gov NCT01389843.

Project description:AimsHeart failure (HF) patients have a high risk of mortality due to sudden cardiac death (SCD) and non-SCD, including pump failure death (PFD). Anaemia predicts more severe symptomatic burden and higher morbidity, as noted by markedly increased hospitalizations and readmission rates, and mortality, underscoring its importance in HF management. Herein, we aimed to determine whether haemoglobin (Hb) level at discharge affects the mode of death and influences SCD risk prediction.MethodsWe evaluated the data of 3020 consecutive acute HF patients from a Japanese prospective multicentre registry. Patients were divided into four groups based on discharge Hb levels. SCD was defined as an unexpected and otherwise unexplained death in a previously stable patient or death due to documented or presumed cardiac arrhythmia without a clear non-cardiovascular cause. The mode of death (SCD, PFD or other cause) was adjudicated by a central committee. Finally, we investigated whether adding Hb level to the Seattle Proportional Risk Model (SPRM; established risk score utilized to estimate 'proportion' of SCD among death events) would affect its performance.ResultsThe mean age of studied patients was 74.3 ± 12.9 years, and 59.8% were male. The mean Hb level was 12.0 ± 2.1 g/dL (61.3% of patients had anaemia defined by World Health Organization criteria). During the 2-year follow-up, 474 deaths (15.7%) occurred, including 93 SCDs (3.1%), 171 PFDs (5.7%) and 210 other deaths (7.0%; predominantly non-cardiac death). Absolute risk of PFD (P < 0.001) or other death (P < 0.001) increased along with the severity of anaemia, whereas the incidence of SCD was low but remained consistent across all four groups (P = 0.440). As a proportion of total deaths in each Hb level group, the contributions from non-SCD increased and from SCD decreased along with anaemia severity (P = 0.007). Adding Hb level to the SPRM improved the overall discrimination (c-index: 0.62 [95% confidence interval (CI) 0.56-0.69] to 0.65 [95% CI 0.59-0.71]), regardless of the baseline ejection fraction (EF) (c-index: 0.64 [95% CI 0.55-0.73] to 0.67 [95% CI 0.58-0.75] for reduced EF and 0.55 [95% CI 0.45-0.66] to 0.61 [95% CI 0.52-0.70] for preserved EF).ConclusionsThe discharge Hb level provides information about both absolute and proportional risks for each mode of death in acute HF patients, and the addition of Hb level improves the performance of SPRM by identifying more non-SCD cases. Future 'proportional' SCD risk models should incorporate Hb level as a covariate to meet this high performance.

Project description:Nosocomial acute-on-chronic liver failure (nACLF) develops in at least 10% of patients with cirrhosis hospitalized for acute decompensation (AD), greatly worsening their prognosis. In this prospective observational study, we aimed to identify rapidly obtainable predictors at admission, which allow for the early recognition and stratification of patients at risk of nACLF. Methods:A total of 516 consecutive patients hospitalized for AD of cirrhosis were screened: those who did not present ACLF at admission (410) were enrolled and surveilled for the development of nACLF. Results:Fifty-nine (14%) patients developed nALCF after a median of 7 (IQR 4-18) days. At admission, they presented a more severe disease and higher degrees of systemic inflammation and anemia than those (351; 86%) who remained free from nACLF. Competing risk multivariable regression analysis showed that baseline MELD score (sub-distribution hazard ratio [sHR] 1.15; 95% CI 1.10-1.21; p ?0.001), hemoglobin level (sHR 0.81; 95% CI 0.68-0.96; p = 0.018), and leukocyte count (sHR 1.11; 95% CI 1.06-1.16; p ?0.001) independently predicted nACLF. Their optimal cut-off points, determined by receiver-operating characteristic curve analysis, were: 13 points for MELD score, 9.8 g/dl for hemoglobin, and 5.6x109/L for leukocyte count. These thresholds were used to stratify patients according to the cumulative incidence of nACLF, being 0, 6, 21 and 59% in the presence of 0, 1, 2 or 3 risk factors (p ?0.001). Nosocomial bacterial infections only increased the probability of developing nACLF in patients with at least 1 risk factor, rising from 3% to 29%, 16% to 50% and 52% to 83% in patients with 1, 2 or 3 risk factors, respectively. Conclusions:Easily available laboratory parameters, related to disease severity, systemic inflammation, and anemia, can be used to identify, at admission, hospitalized patients with AD at increased risk of developing nACLF. Lay summary:More than 10% of patients with cirrhosis hospitalized because of an acute decompensation develop acute-on-chronic liver failure, which is associated with high short-term mortality, during their hospital stay. We found that the combination of 3 easily obtainable variables (model for end-stage liver disease score, leukocyte count and hemoglobin level) help to identify and stratify patients according to their risk of developing nosocomial acute-on-chronic liver failure, from nil to 59%. Moreover, if a nosocomial bacterial infection occurs, such an incidence proportionally increases from nil to 83%. This simple approach helps to identify patients at risk of developing nosocomial acute-on-chronic liver failure at admission to hospital, enabling clinicians to put in place preventive measures.

Project description:BackgroundAcute kidney injury (AKI) in pediatric critical care patients is diagnosed using elevated serum creatinine, which occurs only after kidney impairment. There are no treatments other than supportive care for AKI once it has developed, so it is important to identify patients at risk to prevent injury. This study develops a machine learning model to learn pre-disease patterns of physiological measurements and predict pediatric AKI up to 48 h earlier than the currently established diagnostic guidelines.MethodsEHR data from 16,863 pediatric critical care patients between 1 month to 21 years of age from three independent institutions were used to develop a single machine learning model for early prediction of creatinine-based AKI using intelligently engineered predictors, such as creatinine rate of change, to automatically assess real-time AKI risk. The primary outcome is prediction of moderate to severe AKI (Stage 2/3), and secondary outcomes are prediction of any AKI (Stage 1/2/3) and requirement of renal replacement therapy (RRT). Predictions generate alerts allowing fast assessment and reduction of AKI risk, such as: "patient has 90% risk of developing AKI in the next 48 h" along with contextual information and suggested response such as "patient on aminoglycosides, suggest check level and review dose and indication".ResultsThe model was successful in predicting Stage 2/3 AKI prior to detection by conventional criteria with a median lead-time of 30 h at AUROC of 0.89. The model predicted 70% of subsequent RRT episodes, 58% of Stage 2/3 episodes, and 41% of any AKI episodes. The ratio of false to true alerts of any AKI episodes was approximately one-to-one (PPV 47%). Among patients predicted, 79% received potentially nephrotoxic medication after being identified by the model but before development of AKI.ConclusionsAs the first multi-center validated AKI prediction model for all pediatric critical care patients, the machine learning model described in this study accurately predicts moderate to severe AKI up to 48 h in advance of AKI onset. The model may improve outcome of pediatric AKI by providing early alerting and actionable feedback, potentially preventing or reducing AKI by implementing early measures such as medication adjustment.

Dataset Information

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

Publications

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

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets