Abstract:Background Between 2013 and 2015, the UK Biobank collected accelerometer traces from 103,712 volunteers aged between 40 and 69 years using wrist-worn triaxial accelerometers for 1 week. This data set has been used in the past to verify that individuals with chronic diseases exhibit reduced activity levels compared with healthy populations. However, the data set is likely to be noisy, as the devices were allocated to participants without a set of inclusion criteria, and the traces reflect free-living conditions. Objective This study aims to determine the extent to which accelerometer traces can be used to distinguish individuals with type 2 diabetes (T2D) from normoglycemic controls and to quantify their limitations. Methods Machine learning classifiers were trained using different feature sets to segregate individuals with T2D from normoglycemic individuals. Multiple criteria, based on a combination of self-assessment UK Biobank variables and primary care health records linked to UK Biobank participants, were used to identify 3103 individuals with T2D in this population. The remaining nondiabetic 19,852 participants were further scored on their physical activity impairment severity based on other conditions found in their primary care data, and those deemed likely physically impaired at the time were excluded. Physical activity features were first extracted from the raw accelerometer traces data set for each participant using an algorithm that extends the previously developed Biobank Accelerometry Analysis toolkit from Oxford University. These features were complemented by a selected collection of sociodemographic and lifestyle features available from UK Biobank. Results We tested 3 types of classifiers, with an area under the receiver operating characteristic curve (AUC) close to 0.86 (95% CI 0.85-0.87) for all 3 classifiers and F1 scores in the range of 0.80-0.82 for T2D-positive individuals and 0.73-0.74 for T2D-negative controls. Results obtained using nonphysically impaired controls were compared with highly physically impaired controls to test the hypothesis that nondiabetic conditions reduce classifier performance. Models built using a training set that included highly impaired controls with other conditions had worse performance (AUC 0.75-0.77; 95% CI 0.74-0.78; F1 scores in the range of 0.76-0.77 for T2D positives and 0.63-0.65 for controls). Conclusions Granular measures of free-living physical activity can be used to successfully train machine learning models that are able to discriminate between individuals with T2D and normoglycemic controls, although with limitations because of the intrinsic noise in the data sets. From a broader clinical perspective, these findings motivate further research into the use of physical activity traces as a means of screening individuals at risk of diabetes and for early detection, in conjunction with routinely used risk scores, provided that appropriate quality control is enforced on the data collection protocol.

Potential predictors of type-2 diabetes risk: machine learning, synthetic data and wearable health devices

Potential predictors of type-2 diabetes risk: machine learning, synthetic data and wearable health devices

Machine Learning as a Support for the Diagnosis of Type 2 Diabetes

Predicting the Development of Type 2 Diabetes in a Large Australian Cohort Using Machine-Learning Techniques: Longitudinal Survey Study

Type 2 Diabetes with Artificial Intelligence Machine Learning: Methods and Evaluation

Prediction of Type 2 Diabetes Based on Machine Learning Algorithm

Using Wearable Activity Trackers to Predict Type 2 Diabetes: Machine Learning–Based Cross-sectional Study of the UK Biobank Accelerometer Cohort

1233-P: Prediction of Type 2 Diabetes Occurrence Using Machine Learning Model

Prediction of type 1 diabetes using a genetic risk model in the Diabetes Autoimmunity Study in the Young.

Social Determinants of Health to Predict Type 2 Diabetes

Machine Learning Methods to Predict Diabetes Complications

Identifying top ten predictors of type 2 diabetes through machine learning analysis of UK Biobank data

Population-Level Prediction of Type 2 Diabetes From Claims Data and Analysis of Risk Factors

A novel mathematical model for predicting the benefits of physical activity on type 2 diabetes progression

Exploring Prediabetes Pathways Using Explainable AI on Data from Electronic Medical Records

Performance analysis and prediction of type 2 diabetes mellitus based on lifestyle data using machine learning approaches

Predicting Type 2 Diabetes Metabolic Phenotypes Using Continuous Glucose Monitoring and a Machine Learning Framework

Prediction of Incident Diabetic Retinopathy in Adults With Type 1 Diabetes Using Machine Learning Approach: An Exploratory Study

Prediction of type 2 diabetes mellitus using hematological factors based on machine learning approaches: a cohort study analysis

Prediction of Diabetes using Machine Learning

Machine Learning Reveals Time-Varying Microbial Predictors with Complex Effects on Glucose Regulation