No abstract

Care management/epidimeological, level IV.

No abstract

BACKGROUND The efficacy of vitamin C (VitC) and thiamine (THMN) in patients admitted to the intensive care unit (ICU) with sepsis is unclear. The purpose of this study was to evaluate the effect of VitC and THMN on mortality and lactate clearance in ICU patients. We hypothesized that survival and lactate clearance would be improved when treated with thiamine and/or VitC. METHODS The Philips eICU database version 2.0 was queried for patients admitted to the ICU in 2014 to 2015 for 48 hours or longer and patients with sepsis and an elevated lactate of 2.0 mmol/L or greater. Subjects were categorized according to the receipt of VitC, THMN, both, or neither. The primary outcome was in-hospital mortality. Secondary outcome was lactate clearance defined as lactate less than 2.0 mmol/L achieved after maximum lactate. Univariable comparisons included age, sex, race, Acute Physiology Score III, Acute Physiology and Chronic Health Evaluation (APACHE) IVa score, Sequential Organ Failure Assessment, surgical ICU admission status, intubation status, hospital region, liver disease, vasopressors, steroids, VitC and THMN orders. Kaplan-Meier curves, logistic regression, propensity score matching, and competing risks modeling were constructed. RESULTS Of 146,687 patients from 186 hospitals, 7.7% (n = 11,330) were included. Overall mortality was 25.9% (n = 2,930). Evidence in favor of an association between VitC and/or THMN administration, and survival was found on log rank test (all p < 0.001). After controlling for confounding factors, VitC (adjusted odds ratio [AOR], 0.69 [0.50–0.95]) and THMN (AOR, 0.71 [0.55–0.93]) were independently associated with survival and THMN was associated with lactate clearance (AOR, 1.50 [1.22–1.96]). On competing risk model VitC (AOR, 0.675 [0.463–0.983]), THMN (AOR, 0.744 [0.569–0.974]), and VitC+THMN (AOR, 0.335 [0.13–0.865]) were associated with survival but not lactate clearance. For subgroup analysis of patients on vasopressors, VitC+THMN were associated with lactate clearance (AOR, 1.85 [1.05–3.24]) and survival (AOR, 0.223 [0.0678–0.735]). CONCLUSION VitC+THMN is associated with increased survival in septic ICU patients. Randomized, multicenter trials are needed to better understand their effects on outcomes. LEVEL OF EVIDENCE Therapeutic Study, Level IV.

No abstract

BACKGROUND:There are no national studies of nonelective readmissions after emergency general surgery (EGS) diagnoses that track nonindex hospital readmission. We sought to determine the rate of overall and nonindex hospital readmissions at 30 and 90 days after discharge for EGS diagnoses, hypothesizing a significant portion would be to nonindex hospitals. METHODS:The 2013 to 2014 Nationwide Readmissions Database was queried for all patients 16 years or older admitted with an EGS primary diagnosis and survived index hospitalization. Multivariable logistic regression identified risk factors for nonelective 30-and 90-day readmission to index and nonindex hospitals. RESULTS: Of 4,171,983 patients, 13% experienced unplanned readmission at 30 days. Of these, 21% were admitted to a nonindex hospital. By 90 days, 22% experienced an unplanned readmission, of which 23% were to a nonindex hospital. The most common reason for readmission was infection. Publicly insured or uninsured patients accounted for 67% of admissions and 77% of readmissions. Readmission predictors at 30 days included leaving against medical advice (odds ratio [OR], 2.51 [2.47-2.56]), increased length of stay (4-7 days: OR, 1.42 [1.41-1.43]; >7 days: OR, 2.04 [2.02-2.06]), Charlson Comorbidity Index ≥2 (OR, 1.72 [1.71-1.73]), public insurance (Medicare: OR, 1.45 [1.44-1.46]; Medicaid: OR, 1.38 [1.37-1.40]), EGS patients who fell into the "Other" surgical category (OR, 1.42 [1.38-1.48]), and nonroutine discharge. Risk factors for readmission remained consistent at 90 days. CONCLUSION:Given that nonindex hospital EGS readmission accounts for nearly a quarter of readmissions and often related to important benchmarks such as infection, current EGS quality metrics are inaccurate. This has implications for policy, benchmarking, and readmission reduction programs.

Prior studies have used vital signs and laboratory measurements with conventional modeling techniques to predict acute kidney injury (AKI). The purpose of this study was to use the trend in vital signs and laboratory measurements with machine learning algorithms for predicting AKI in ICU patients. The eICU Collaborative Research Database was queried for five consecutive days of laboratory measurements per patient. Patients with AKI were identified and trends in vital signs and laboratory values were determined by calculating the slope of the least-squares-fit linear equation using three days for each value. Different machine learning classifiers (gradient boosted trees [GBT], logistic regression, and deep learning) were trained to predict AKI using the laboratory values, vital signs, and slopes. There were 151,098 ICU stays identified and the rate of AKI was 5.6 per cent. The best performing algorithm was GBT with an AUC of 0.834 ± 0.006 and an F-measure of 42.96 per cent ± 1.26 per cent. Logistic regression performed with an AUC of 0.827 ± 0.004 and an F-measure of 28.29 per cent ± 1.01 per cent. Deep learning performed with an AUC of 0.817 ± 0.005 and an F-measure of 42.89 per cent ± 0.91 per cent. The most important variable for GBT was the slope of the minimum creatinine (30.32%). This study identifies the best performing machine learning algorithms for predicting AKI using trends in laboratory values in ICU patients. Early identification of these patients using readily available data indicates that incorporating machine learning predictive models into electronic medical record systems is an inevitable requisite for improving patient outcomes.

The purpose of this study was to use natural language processing of physician documentation to predict mortality in patients admitted to the surgical intensive care unit (SICU). The Multiparameter Intelligent Monitoring in Intensive Care III database was used to obtain SICU stays with six different severity of illness scores. Natural language processing was performed on the physician notes. Classifiers for predicting mortality were created. One classifier used only the physician notes, one used only the severity of illness scores, and one used the physician notes with severity of injury scores. There were 3838 SICU stays identified during the study period and 5.4 per cent ended with mortality. The classifier trained with physician notes with severity of injury scores performed with the highest area under the curve (0.88 ± 0.05) and accuracy (94.6 ± 1.1%). The most important variable was the Oxford Acute Severity of Illness Score (16.0%). The most important terms were “dilated” (4.3%) and “hemorrhage” (3.7%). This study demonstrates the novel use of artificial intelligence to process physician documentation to predict mortality in the SICU. The classifiers were able to detect the subtle nuances in physician vernacular that predict mortality. These nuances provided improved performance in predicting mortality over physiologic parameters alone.