Use of Machine Learning Approaches in Clinical Epidemiological Research of Diabetes

Abstract

Purpose of Review

Machine learning approaches—which seek to predict outcomes or classify patient features by recognizing patterns in large datasets—are increasingly applied to clinical epidemiology research on diabetes. Given its novelty and emergence in fields outside of biomedical research, machine learning terminology, techniques, and research findings may be unfamiliar to diabetes researchers. Our aim was to present the use of machine learning approaches in an approachable way, drawing from clinical epidemiological research in diabetes published from 1 Jan 2017 to 1 June 2020.

Recent Findings

Machine learning approaches using tree-based learners—which produce decision trees to help guide clinical interventions—frequently have higher sensitivity and specificity than traditional regression models for risk prediction. Machine learning approaches using neural networking and “deep learning” can be applied to medical image data, particularly for the identification and staging of diabetic retinopathy and skin ulcers. Among the machine learning approaches reviewed, researchers identified new strategies to develop standard datasets for rigorous comparisons across older and newer approaches, methods to illustrate how a machine learner was treating underlying data, and approaches to improve the transparency of the machine learning process.

Summary

Machine learning approaches have the potential to improve risk stratification and outcome prediction for clinical epidemiology applications. Achieving this potential would be facilitated by use of universal open-source datasets for fair comparisons. More work remains in the application of strategies to communicate how the machine learners are generating their predictions.

Appendix 1

Article Database Search Strategies

Arxiv.org (47 total results)

• •“Diabetes” AND “Machine Learning” in abstract, 2017–2020.

• “Diabetes” AND “Machine Learning” in title, 2017–2020.

• “Diabetes” AND “Random Forest*” in abstract, 2017–2020.

• “Diabetes” AND “Decision tree*” in abstract, 2017–2020.

• “Diabetes” AND “Support vector machine*” in abstract, 2017–2020.

• “Diabetes” AND “Naive Bayes*” in abstract, 2017–2020.

• “Diabetes” AND “Naïve Bayes*” in abstract, 2017–2020.

• “Diabetes” AND “K-means clustering*” in abstract, 2017–2020.

• “Diabetes” AND “K-nearest neighbor*” in abstract, 2017–2020.

• “Diabetes” AND “regression tree*” in abstract, 2017–2020.

• “Diabetes” AND “Artificial neural network*” in abstract, 2017–2020.

PubMed (70 total results).

(Search: (“Machine Learning”[tiab] OR “neural network”[tiab] OR “Deep Learning”[tiab] OR “Computer-aided”[tiab] OR “Prediction Model”[tiab] OR “Data Mining”[tiab] OR “Cross-validation” [tiab] OR “Cross validation”[tiab] OR “Regularized logistic”[tiab] OR “Linear discriminant analysis”[tiab] OR “LDA”[tiab] OR “Random forest”[tiab] OR “Naïve Bayes” [tiab] OR “Least Absolute selection shrinkage operator”[tiab] OR “elastic net”[tiab] OR “LASSO”[tiab] OR “RVM”[tiab] OR “relevance vector machine”[tiab] OR “pattern recognition”[tiab] OR “Computational Intelligence”[tiab] OR “Computational Intelligences” [tiab] OR “Machine Intelligence”[tiab] OR “Knowledge Representation”[tiab] OR “Knowledge Representations”[tiab] OR “support vector”[tiab] OR “SVM”[tiab] OR “Pattern classification” [tiab] OR “Decision Tree”[tiab] OR “Nearest neighbor”[tiab] OR “Classification Tree”[tiab] OR “Regression Tree”[tiab] OR “Iterative Learning”[tiab] OR “recursive partitioning”[tiab] OR “causal forest”[tiab] OR “generalized random forest”[tiab] OR “X-learner”[tiab] OR “R-learner”[tiab] OR “S-learner”[tiab] OR “Neural Networks, Computer”[Mesh] OR “Data Mining”[Mesh] OR “Deep Learning”[Mesh] OR “Pattern Recognition, Automated”[Mesh] OR “Artificial Intelligence”[Mesh] OR “Computational Intelligence”[Mesh] OR “Intelligence, Computational”[Mesh] OR “Machine Intelligence”[Mesh] OR “Intelligence, Machine”[Mesh] OR “Computer Reasoning”[Mesh] OR “Reasoning, Computer”[Mesh] OR “Knowledge Acquisition (Computer)”[Mesh] OR “Acquisition, Knowledge (Computer)”[Mesh] OR “Knowledge Representation (Computer)”[Mesh] OR “Knowledge Representations (Computer)”[Mesh] OR “Support Vector Machine”[Mesh] OR “Representation, Knowledge (Computer)”[Mesh] OR “Decision Trees”[Mesh])) AND (“diabetes mellitus”[mesh] OR “diabetes”[tiab]) Filters: English, from 2017 to 2020.

Scopus (497 total results).

(TITLE-ABS-KEY (“Machine Learning”)) AND (TITLE-ABS-KEY (“diabetes”)) AND (LIMIT-TO (PUBSTAGE, “final”)) AND (LIMIT-TO (DOCTYPE, “ar”)) AND (LIMIT-TO (SUBJAREA, “MEDI”) OR LIMIT-TO (SUBJAREA, “COMP”)) AND (LIMIT-TO (PUBYEAR, 2020) OR LIMIT-TO (PUBYEAR, 2019) OR LIMIT-TO (PUBYEAR, 2018) OR LIMIT-TO (PUBYEAR, 2017)) AND (LIMIT-TO (LANGUAGE, “English”))

Web of Science (161 total results).

TI = ((“Machine Learning” OR “neural network” OR “Deep Learning” OR “Computer-aided” OR “Prediction Model” OR “Data Mining” OR “Cross-validation” OR “Cross validation” OR “Regularized logistic” OR “Linear discriminant analysis” OR “LDA” OR “Random forest” OR “Naïve Bayes” OR “Least Absolute selection shrinkage operator” OR “elastic net” OR “LASSO” OR “RVM” OR “relevance vector machine” OR “pattern recognition” OR “Computational Intelligence” OR “Computational Intelligences” OR “Machine Intelligence” OR “Knowledge Representation” OR “Knowledge Representations” OR “support vector” OR “SVM” OR “Pattern classification” OR “Decision Tree” OR “Nearest neighbor” OR “Classification Tree” OR “Regression Tree” OR “Iterative Learning” OR “recursive partitioning” OR “causal forest” OR “generalized random forest” OR “X-learner” OR “R-learner” OR “S-learner” OR “Neural Networks, Computer” OR “Data Mining” OR “Deep Learning” OR “Pattern Recognition, Automated” OR “Artificial Intelligence” OR “Computational Intelligence” OR “Intelligence, Computational” OR “Machine Intelligence” OR “Intelligence, Machine” OR “Computer Reasoning” OR “Reasoning, Computer” OR “Knowledge Acquisition (Computer)” OR “Acquisition, Knowledge (Computer)” OR “Knowledge Representation (Computer)” OR “Knowledge Representations (Computer)” OR “Support Vector Machine” OR “Representation, Knowledge (Computer)” OR “Decision Trees”) AND (Diabetes)).