An explainable machine learning algorithm for risk factor analysis of in-hospital mortality in sepsis survivors with ICU readmission

Jiang, Zhengyu; Bo, Lulong; Xu, Zhenhua; Song, Yubing; Wang, Jiafeng; Wen, Ping-shan; Wan, Xiaojian; Yang, Tao; Deng, Xiaoming; Bian, Jinjun

doi:10.1016/j.cmpb.2021.106040

Cited by 45 publications

(29 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Guo et al [ 36 ] constructed a nomogram to predict in-hospital mortality for myocardial infarction patients (AUROC 0.803). Jiang et al [ 37 ] used machine learning to predict in-hospital mortality in sepsis survivors (sepsis: AUROC 0.732; nonsepsis: AUROC 0.830). Suresh et al [ 38 ] developed a multitask model (AUROC 0.869) for mortality prediction that outperformed global and separate models.…”

Section: Discussionmentioning

confidence: 99%

Sequential Data–Based Patient Similarity Framework for Patient Outcome Prediction: Algorithm Development

Wang¹,

Wang²,

Zhou³

et al. 2022

J Med Internet Res

View full text Add to dashboard Cite

Background Sequential information in electronic medical records is valuable and helpful for patient outcome prediction but is rarely used for patient similarity measurement because of its unevenness, irregularity, and heterogeneity. Objective We aimed to develop a patient similarity framework for patient outcome prediction that makes use of sequential and cross-sectional information in electronic medical record systems. Methods Sequence similarity was calculated from timestamped event sequences using edit distance, and trend similarity was calculated from time series using dynamic time warping and Haar decomposition. We also extracted cross-sectional information, namely, demographic, laboratory test, and radiological report data, for additional similarity calculations. We validated the effectiveness of the framework by constructing k–nearest neighbors classifiers to predict mortality and readmission for acute myocardial infarction patients, using data from (1) a public data set and (2) a private data set, at 3 time points—at admission, on Day 7, and at discharge—to provide early warning patient outcomes. We also constructed state-of-the-art Euclidean-distance k–nearest neighbor, logistic regression, random forest, long short-term memory network, and recurrent neural network models, which were used for comparison. Results With all available information during a hospitalization episode, predictive models using the similarity model outperformed baseline models based on both public and private data sets. For mortality predictions, all models except for the logistic regression model showed improved performances over time. There were no such increasing trends in predictive performances for readmission predictions. The random forest and logistic regression models performed best for mortality and readmission predictions, respectively, when using information from the first week after admission. Conclusions For patient outcome predictions, the patient similarity framework facilitated sequential similarity calculations for uneven electronic medical record data and helped improve predictive performance.

show abstract

Section: Discussionmentioning

confidence: 99%

Sequential Data–Based Patient Similarity Framework for Patient Outcome Prediction: Algorithm Development

Wang¹,

Wang²,

Zhou³

et al. 2022

J Med Internet Res

View full text Add to dashboard Cite

show abstract

“…As sepsis has gradually become an important difficulty and subject content in clinical medical research, clinical research on sepsis has also begun to increase, and many clinical treatment results have been obtained, and valuable diagnosis and treatment experience has been accumulated. Comprehensive progress and updates have also been made in terms of the definition of the nature of the disease, diagnosis and treatment, and treatment standards of sepsis [6][7][8][9][10].…”

Section: Introductionmentioning

confidence: 99%

Analysis of Prognostic Risk Factors of Sepsis Patients in Intensive Care Unit Based on Data Analysis

Yao

Zhang

Zhou

2022

Journal of Healthcare Engineering

View full text Add to dashboard Cite

In this paper, a data-enabled analysis of the prognostic risk factors of sepsis patients in the intensive care unit is presented. For this purpose, we have selected 220 sepsis patients, preferably those admitted to the intensive care unit for treatment in a tertiary a hospital in Tianjin from June 2018 to June 2019 and received complete data as the research objects, to explore the prognostic risk factors of sepsis patients in the intensive care unit. All patients met the SSC sepsis diagnosis guidelines and recorded the patients’ age, gender, underlying disease, and infection site. Laboratory indicators, such as blood routine, electrolytes, arterial blood gas, liver function, and renal function, were collected within 24 hours of admission. Furthermore, the corresponding specimens were cultured for pathogenic microorganisms according to the site of infection. The LAC value was measured at admission and 24 h after admission, and the 24 h lactate clearance rate was calculated. The Acute Physiological and Chronic Health Status Score II (APACHE-II) and SOFA score were calculated, which were based on the worst value of the index within 24 hours after admission. According to the prognosis of patients during hospitalization, they are divided into two groups: (i) survival group and (ii) death group. We entered all the data into Excel and used SPSS21.0 statistical software for data analysis and processing. Quantitative data are tested for normality. Quantitative data for normal distribution are expressed as mean ± standard deviation, and normal distribution and uniform variance are measured. The factors affecting the prognosis of patients with sepsis were first subjected to a single-factor logistic regression analysis, and a multiple logistic regression analysis was performed on the basis of the significance of the single-factor analysis. The results found that the prognosis of patients with sepsis in the ICU is affected by multiple factors such as underlying diseases, infectious microorganisms, comorbidities, and interventional therapy. APACHE-II score, 24 h lactate clearance rate, ARDS, and DIC are independent risk factors that affect the prognosis of ICU patients.

show abstract

“…As mentioned above, no studies were found that analyzed the problem of mortality within the ICU by age group. There are studies focused on specific diseases [15,23], but the features that most affect mortality and the thresholds involving a specific disease may be different from those that affect mortality in a general way. In addition, there are other factors that affect the results, such as the database, the selected variables, the predictor model, the data collection time window, or the defined age groups, among others.…”

Section: Discussionmentioning

confidence: 99%

“…Once the cohort was selected, the next step was feature extraction. In this work, 33 clinical variables were considered, specifically those that had less than 20% empty data and were frequently used in ICU analysis [15]. Data from the first 24 h of each of the selected patient's first ICU stay were considered.…”

Section: Feature Extractionmentioning

confidence: 99%

Using Explainable Machine Learning to Improve Intensive Care Unit Alarm Systems

González-Nóvoa

Busto

Rodríguez-Andina

et al. 2021

Sensors

View full text Add to dashboard Cite

Due to the continuous monitoring process of critical patients, Intensive Care Units (ICU) generate large amounts of data, which are difficult for healthcare personnel to analyze manually, especially in overloaded situations such as those present during the COVID-19 pandemic. Therefore, the automatic analysis of these data has many practical applications in patient monitoring, including the optimization of alarm systems for alerting healthcare personnel. In this paper, explainable machine learning techniques are used for this purpose, with a methodology based on age-stratification, boosting classifiers, and Shapley Additive Explanations (SHAP) proposed. The methodology is evaluated using MIMIC-III, an ICU patient research database. The results show that the proposed model can predict mortality within the ICU with AUROC values of 0.961, 0.936, 0.898, and 0.883 for age groups 18–45, 45–65, 65–85 and 85+, respectively. By using SHAP, the features with the highest impact in predicting mortality for different age groups and the threshold from which the value of a clinical feature has a negative impact on the patient’s health can be identified. This allows ICU alarms to be improved by identifying the most important variables to be sensed and the threshold values at which the health personnel must be warned.

show abstract

An explainable machine learning algorithm for risk factor analysis of in-hospital mortality in sepsis survivors with ICU readmission

Cited by 45 publications

References 24 publications

Sequential Data–Based Patient Similarity Framework for Patient Outcome Prediction: Algorithm Development

Sequential Data–Based Patient Similarity Framework for Patient Outcome Prediction: Algorithm Development

Analysis of Prognostic Risk Factors of Sepsis Patients in Intensive Care Unit Based on Data Analysis

Using Explainable Machine Learning to Improve Intensive Care Unit Alarm Systems

Contact Info

Product

Resources

About