Explainable machine learning prediction of ICU mortality

Chia, Alvin Har Teck; Khoo, May Sze; Lim, Andy; Ong, Kian Eng; Sun, Yixuan; Nguyen, Binh P.; Chua, Matthew Chin Heng; Pang, Junxiong

doi:10.1016/j.imu.2021.100674

Cited by 12 publications

(5 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, Veith and Steele[ 23 ] developed a LazyKStar model to predict mortality in ICU patients at the time of hospital admission, obtaining a 10-fold validation AUC value of 0.75.A recurrent neural network inputted with 44 clinical and laboratory features from the first 24 h of ICU patient admission proposed by Thorsen-Meyer et al [ 24 ] achieved an AUC of 0.82. The extreme gradient boosted trees classifier developed by Chia et al [ 25 ] reached an AUC of 0.83 using 42 predictive variables. The formats and results of these last two studies are comparable to ours, since we reached an AUC of 0.85 using a random forest fed by 50 features.…”

Section: Discussionmentioning

confidence: 99%

Prediction of hospital mortality in intensive care unit patients from clinical and laboratory data: A machine learning approach

Silveira¹,

Pretti²,

Santos³

et al. 2022

WJCCM

View full text Add to dashboard Cite

BACKGROUND Intensive care unit (ICU) patients demand continuous monitoring of several clinical and laboratory parameters that directly influence their medical progress and the staff’s decision-making. Those data are vital in the assistance of these patients, being already used by several scoring systems. In this context, machine learning approaches have been used for medical predictions based on clinical data, which includes patient outcomes. AIM To develop a binary classifier for the outcome of death in ICU patients based on clinical and laboratory parameters, a set formed by 1087 instances and 50 variables from ICU patients admitted to the emergency department was obtained in the “WiDS (Women in Data Science) Datathon 2020: ICU Mortality Prediction” dataset. METHODS For categorical variables, frequencies and risk ratios were calculated. Numerical variables were computed as means and standard deviations and Mann-Whitney U tests were performed. We then divided the data into a training (80%) and test (20%) set. The training set was used to train a predictive model based on the Random Forest algorithm and the test set was used to evaluate the predictive effectiveness of the model. RESULTS A statistically significant association was identified between need for intubation, as well predominant systemic cardiovascular involvement, and hospital death. A number of the numerical variables analyzed (for instance Glasgow Coma Score punctuations, mean arterial pressure, temperature, pH, and lactate, creatinine, albumin and bilirubin values) were also significantly associated with death outcome. The proposed binary Random Forest classifier obtained on the test set ( n = 218) had an accuracy of 80.28%, sensitivity of 81.82%, specificity of 79.43%, positive predictive value of 73.26%, negative predictive value of 84.85%, F1 score of 0.74, and area under the curve score of 0.85. The predictive variables of the greatest importance were the maximum and minimum lactate values, adding up to a predictive importance of 15.54%. CONCLUSION We demonstrated the efficacy of a Random Forest machine learning algorithm for handling clinical and laboratory data from patients under intensive monitoring. Therefore, we endorse the emerging notion that machine learning has great potential to provide us support to critically question existing methodologies, allowing improvements that reduce mortality.

show abstract

Section: Discussionmentioning

confidence: 99%

Prediction of hospital mortality in intensive care unit patients from clinical and laboratory data: A machine learning approach

Silveira¹,

Pretti²,

Santos³

et al. 2022

WJCCM

View full text Add to dashboard Cite

show abstract

“…To tackle the first part of the question, the studies reviewed cover a diverse range of healthcare-focused datasets used for developing transparent and interpretable AI models. These datasets span various medical domains, including cancer [6], [15], [91], [1], [106], medical imaging [107], [108], [10], [32], clinical and physiological data [109], [110], [111], [112], [41], [84], [113], [114], [115], and mobility and activity data [116]. In the cancer domain, the studies utilized datasets such as The Cancer Genome Atlas (TCGA-COAD) and an Asian colorectal cancer (Asian-CRC) cohort [117] to create a pathomics-based model capable of forecasting microsatellite instability occurrences in colorectal cancer.…”

Section: ) Rq1mentioning

confidence: 99%

“…Endoscopic images of Barrett's esophagus and early-stage adenocarcinoma from the MICCAI 2015 EndoVis Challenge [32]. The clinical and physiological datasets encompass a wide variety of healthcare scenarios, such as the PhysioNet Challenge 2012 dataset [109] for ICU mortality prediction, the Electrocardiogram Vigilance with Electronic Data Warehouse (ECG-ViEW II) dataset [110] for acute myocardial infarction prediction, the multi-center eICU Collaborative Research Database [111], the MIMIC-IV dataset [112] for interpretability and fairness evaluation of DL models, the MIT-BIH Arrhythmia Database [84] for heartbeat classification, the ImmuneCODE database [113] for immune repertoire-based medical condition identification, and various datasets for cardiovascular risk assessment [114] and red blood cell transfusion prediction [115].…”

Section: ) Rq1mentioning

confidence: 99%

“…Similarly, the laboratory and diagnosis data used in the knowledge-driven AI system [41] were not thoroughly examined for data quality and potential biases. Furthermore, a significant limitation identified in the reviewed studies is the lack of external validation [117], [15], [91], [1], [106], [107], [108], [10], [32], [109], [110], [111], [112], [41], [84], [113], [114], [115], [116]. Many of the studies did not report on the performance and generalizability of their models when applied to external, independent datasets, which is crucial for assessing the real-world applicability of the developed AI systems.…”

Section: ) Rq1mentioning

confidence: 99%

See 1 more Smart Citation

A Systematic Literature Review on Transparencyand Interpretability of AI models in Healthcare:Taxonomies, Tools, Techniques, Datasets, OpenResearch Challenges, and Future Trends

Shafik,

Hidayatullah,

Kalinaki

et al. 2024

Preprint

View full text Add to dashboard Cite

The increased utilization of disruptive health and biomedical informatics technologies, such as artificial intelligence (AI), has accelerated medical operations from patient-centered medical experience data management to simplified medical procedures in this generative era. As these technologies integrate into traditional approaches, they raise critical medical concerns, entailing transparency and interpretability of these AI models. This study conducts a systematic literature review (SLR) and presents an exhaustive review of the studies using data collection procedures and publicly available academic databases. 1837 articles published between 2014 and 2024 were obtained from eight popular academic databases: PubMed, ACM Library, Springer, Scopus, IEEE Xplore, ScienceDirect, Google Scholar, and Web of Science. A comprehensive screening process was used, and 148 articles were considered based on the relevance of the AI method to healthcare and biomedical. The studied studies demonstrate that the majority of medical people still find it complex to effectively explain the reasoning behind the decisions AI models make during biomedical experiments, leading to limited trust, biased decision-making, and unknown patient data safety. Model-agnostic strategies and explainable AI (XAI) frameworks are inspected, together with crucial datasets for training and assessment. The main challenges are AI model intricacy and regulatory compliance, while future trends highlight fairness and predisposition mitigation. Limited studies are focusing on improving AI openness, trust, and interpretability. Towards the end, it reveals that there is still a big research gap in descriptive explainable AI models in healthcare when integrating AI into clinical practice while maintaining ethical standards and patient-centric care.

show abstract

“…Several ML-based models have been already developed for mortality prediction in ICU, especially for adult population (30)(31)(32)(33)(34), based on large public available ICU datasets such as Medical Information Mart for Intensive Care (MIMIC-III) (35) and eICU Collaborative Research Database (eICU) (36). However, few MLbased models have been specifically developed for pediatric ICU.…”

Section: Machine Learning To Predict Mortalitymentioning

confidence: 99%

Machine Learning-Based Systems for the Anticipation of Adverse Events After Pediatric Cardiac Surgery

et al. 2022

View full text Add to dashboard Cite

Pediatric congenital heart disease (CHD) patients are at higher risk of postoperative complications and clinical deterioration either due to their underlying pathology or due to the cardiac surgery, contributing significantly to mortality, morbidity, hospital and family costs, and poor quality of life. In current clinical practice, clinical deterioration is detected, in most of the cases, when it has already occurred. Several early warning scores (EWS) have been proposed to assess children at risk of clinical deterioration using vital signs and risk indicators, in order to intervene in a timely manner to reduce the impact of deterioration and risk of death among children. However, EWS are based on measurements performed at a single time point without incorporating trends nor providing information about patient's risk trajectory. Moreover, some of these measurements rely on subjective assessment making them susceptible to different interpretations. All these limitations could explain why the implementation of EWS in high-resource settings failed to show a significant decrease in hospital mortality. By means of machine learning (ML) based algorithms we could integrate heterogeneous and complex data to predict patient's risk of deterioration. In this perspective article, we provide a brief overview of the potential of ML technologies to improve the identification of pediatric CHD patients at high-risk for clinical deterioration after cardiac surgery, and present the CORTEX traffic light, a ML-based predictive system that Sant Joan de Déu Barcelona Children's Hospital is implementing, as an illustration of the application of an ML-based risk stratification system in a relevant hospital setting.

show abstract

Explainable machine learning prediction of ICU mortality

Cited by 12 publications

References 10 publications

Prediction of hospital mortality in intensive care unit patients from clinical and laboratory data: A machine learning approach

Prediction of hospital mortality in intensive care unit patients from clinical and laboratory data: A machine learning approach

A Systematic Literature Review on Transparencyand Interpretability of AI models in Healthcare:Taxonomies, Tools, Techniques, Datasets, OpenResearch Challenges, and Future Trends

Machine Learning-Based Systems for the Anticipation of Adverse Events After Pediatric Cardiac Surgery

Contact Info

Product

Resources

About