Data processing pipeline for cardiogenic shock prediction using machine learning

Abstract

IntroductionRecent advances in machine learning provide new possibilities to process and analyse observational patient data to predict patient outcomes. In this paper, we introduce a data processing pipeline for cardiogenic shock (CS) prediction from the MIMIC III database of intensive cardiac care unit patients with acute coronary syndrome. The ability to identify high-risk patients could possibly allow taking pre-emptive measures and thus prevent the development of CS.MethodsWe mainly focus on techniques for the imputation of missing data by generating a pipeline for imputation and comparing the performance of various multivariate imputation algorithms, including k-nearest neighbours, two singular value decomposition (SVD)—based methods, and Multiple Imputation by Chained Equations. After imputation, we select the final subjects and variables from the imputed dataset and showcase the performance of the gradient-boosted framework that uses a tree-based classifier for cardiogenic shock prediction.ResultsWe achieved good classification performance thanks to data cleaning and imputation (cross-validated mean area under the curve 0.805) without hyperparameter optimization.ConclusionWe believe our pre-processing pipeline would prove helpful also for other classification and regression experiments.

Artificial Intelligence in the Early Prediction of Cardiogenic Shock in Acute Heart Failure or Myocardial Infarction Patients: A Systematic Review and Meta-Analysis

“…All studies were conducted in the USA except for Bai et al 2021 [ 21 ], which was conducted in China. Rahman et al 2022 [ 19 ] included patients hospitalized with acute decompensated heart failure, whereas two studies [ 16 , 18 ] included patients having cardiac catheterizations for acute coronary syndrome. The remaining three studies [ 17 , 20 , 21 ] included patients from various backgrounds, including those with heart failure and myocardial infarction.…”

“…Area under the curve (AUC): Three studies reported [ 16 , 18 , 20 ] the AUC in a way to be combined in a meta-analysis with a total sample size of 10,869. The pooled mean of AUC was 0.808 (95% CI: 0.727, 0.890) (Figure 2 ).…”

“… This plot represents the combined average Area Under the Curve (AUC) values of machine learning predictive models as reported in three studies: Pinevich et al 2021 [ 20 ], Jajcay et al 2023 [ 18 ], and Bohm et al 2022 [ 16 ]. …”

“…Bohm et al 2022 [ 16 ] revealed that the AUC of their selected two best-performing algorithms on 3,232 patients hospitalized with AMI was 0.77 and 0.76. Bai et al 2021 [ 21 ] and Jajcay et al 2023 [ 18 ] applied their ML models to 684 and 2,253 AMI patients, respectively, and the AUC was 0.82 and 0.805, respectively. Pinevich et al (2021) found that the AUC of their ML model applied to 5,384 patients with various backgrounds, including HF and AMI, was 0.86.…”

“…The accuracy of the ML models in predicting CS was reported in four studies [ 18 - 21 ]. It ranged between 0.88 and 0.93.…”