Using data available in the ICU in real-time, Artificial Intelligence Sepsis Expert can accurately predict the onset of sepsis in an ICU patient 4-12 hours prior to clinical recognition. A prospective study is necessary to determine the clinical utility of the proposed sepsis prediction model.

Sepsis, a life-threatening organ dysfunction, is a clinical syndrome triggered by acute infection and affects over 1 million Americans every year. Untreated sepsis can progress to septic shock and organ failure, making sepsis one of the leading causes of morbidity and mortality in hospitals. Early detection of sepsis and timely antibiotics administration is known to save lives. In this work, we design a sepsis prediction algorithm based on data from electronic health records (EHR) using a deep learning approach. While most existing EHR-based sepsis prediction models utilize structured data including vitals, labs, and clinical information, we show that incorporation of features based on clinical texts, using a pre-trained neural language representation model, allows for incorporation of unstructured data without an explicit need for ontology-based named-entity recognition and classification. The proposed model is trained on a large critical care database of over 40,000 patients, including 2805 septic patients, and is compared against competing baseline models. In comparison to a baseline model based on structured data alone, incorporation of clinical texts improved AUC from 0.81 to 0.84. Our findings indicate that incorporation of clinical text features via a pre-trained language representation model can improve early prediction of sepsis and reduce false alarms.

OBJECTIVE To identify circumstances in which repeated measures of organ failure would improve mortality prediction in ICU patients. DESIGN Retrospective cohort study, with external validation in a de-identified ICU database. SETTING Eleven ICUs in three university hospitals within an academic healthcare system in 2014. PATIENTS Adults (18 years or older) who satisfied the following criteria: (1) 2 of 4 systemic inflammatory response syndrome criteria plus an ordered blood culture, all within 24 hours of hospital admission; and (2) ICU admission for at least 2 calendar days, within 72 hours of emergency department presentation. INTERVENTION None MEASUREMENTS AND MAIN RESULTS Data were collected until death, ICU discharge, or the seventh ICU day, whichever came first. The highest SOFA score from the ICU admission day (ICU day 1) was included in a multivariable model controlling for other covariates. The worst SOFA scores from the first 7 days after ICU admission were incrementally added and retained if they obtained statistical significance (p<0.05). The cohort was divided into seven subcohorts to facilitate statistical comparison using the integrated discriminatory index (IDI). Of the 1290 derivation cohort patients, 83 patients (6.4%) died in the ICU, compared with 949 of the 8441 patients (11.2%) in the validation cohort. Incremental addition of SOFA data up to ICU day 5 improved the IDI in the validation cohort. Adding ICU day 6 or 7 SOFA data did not further improve model performance. CONCLUSIONS Serial organ failure data improves prediction of ICU mortality, but a point exists after which further data no longer improves ICU mortality prediction of early sepsis.

No abstract

Poisoning attacks are a category of adversarial machine learning threats in which an adversary attempts to subvert the outcome of the machine learning systems by injecting crafted data into training data set, thus increasing the machine learning model's test error. The adversary can tamper with the data feature space, data labels, or both, each leading to a different attack strategy with different strengths. Various detection approaches have recently emerged, each focusing on one attack strategy. The Achilles heel of many of these detection approaches is their dependence on having access to a clean, untampered data set. In this paper, we propose CAE, a Classification Auto-Encoder based detector against diverse poisoned data. CAE can detect all forms of poisoning attacks using a combination of reconstruction and classification errors without having any prior knowledge of the attack strategy. We show that an enhanced version of CAE (called CAE+) does not have to employ a clean data set to train the defense model. Our experimental results on three real datasets MNIST, Fashion-MNIST and CIFAR demonstrate that our proposed method can maintain its functionality under up to 30% contaminated data and help the defended SVM classifier to regain its best accuracy.