Data science and machine learning in anesthesiology

Abstract

Machine learning (ML) is revolutionizing anesthesiology research. Unlike classical research methods that are largely inference-based, ML is geared more towards making accurate predictions. ML is a field of artificial intelligence concerned with developing algorithms and models to perform prediction tasks in the absence of explicit instructions. Most ML applications, despite being highly variable in the topics that they deal with, generally follow a common workflow. For classification tasks, a researcher typically tests various ML models and compares the predictive performance with the reference logistic regression model. The main advantage of ML lies in its ability to deal with many features with complex interactions and its specific focus on maximizing predictive performance. However, emphasis on data-driven prediction can sometimes neglect mechanistic understanding. This article mainly focuses on the application of supervised ML to electronic health record (EHR) data. The main limitation of EHR-based studies is in the difficulty of establishing causal relationships. However, the associated low cost and rich information content provide great potential to uncover hitherto unknown correlations. In this review, the basic concepts of ML are introduced along with important terms that any ML researcher should know. Practical tips regarding the choice of software and computing devices are also provided. Towards the end, several examples of successful ML applications in anesthesiology are discussed. The goal of this article is to provide a basic roadmap to novice ML researchers working in the field of anesthesiology.

Keywords: Artificial intelligence; Data science; Electronic health record; Machine learning; Predictive analytics; Risk score system.

Fig. 1.

Machine learning (ML) Landscape. ML algorithms can roughly be divided into three categories – Supervised learning, unsupervised learning, and reinforcement learning. Our focus will primarily be on supervised ML algorithms.

Fig. 2.

A decision tree to predict the presence of diabetes based on the Pima Indians Diabetes dataset (glucose: serum glucose [mg/dl], age [yr], mass: body mass index [kg/m²], pedigree: diabetes pedigree function). neg: negative, pos: positive.

Fig. 3.

A hypothetical landscape illustrating the relationship between the loss function and the parameters.

Fig. 4.

Splitting of the dataset into training, validation, and datasets. Model comparison is done using the validation dataset. Predictive performance of the final model is assessed using the test dataset.

Fig. 5.

The stepwise selection procedure. Each time, the most significant variable is incorporated into the prediction model. Such a selection process is repeated over all candidate variables until no more candidates exist or the algorithm hits a predefined stopping rule.

Fig. 6.

The receiver operating characteristic (ROC) curve and the area under the curve (AUC). (A) The yellow shaded area corresponds to the AUC. The dotted red line corresponds to the ROC curve of a random guess algorithm. (B) A model associated with a higher AUC is generally taken to be a better model.