Exploration of Machine Learning and Statistical Techniques in Development of a Low-Cost Screening Method Featuring the Global Diet Quality Score for Detecting Prediabetes in Rural India

Abstract

Background The prevalence of type 2 diabetes has increased substantially in India over the past 3 decades. Undiagnosed diabetes presents a public health challenge, especially in rural areas, where access to laboratory testing for diagnosis may not be readily available. Objectives The present work explores the use of several machine learning and statistical methods in the development of a predictive tool to screen for prediabetes using survey data from an FFQ to compute the Global Diet Quality Score (GDQS). Methods The outcome variable prediabetes status (yes/no) used throughout this study was determined based upon a fasting blood glucose measurement ≥100 mg/dL. The algorithms utilized included the generalized linear model (GLM), random forest, least absolute shrinkage and selection operator (LASSO), elastic net (EN), and generalized linear mixed model (GLMM) with family unit as a (cluster) random (intercept) effect to account for intrafamily correlation. Model performance was assessed on held-out test data, and comparisons made with respect to area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Results The GLMM, GLM, LASSO, and random forest modeling techniques each performed quite well (AUCs >0.70) and included the GDQS food groups and age, among other predictors. The fully adjusted GLMM, which included a random intercept for family unit, achieved slightly superior results (AUC of 0.72) in classifying the prediabetes outcome in these cluster-correlated data. Conclusions The models presented in the current work show promise in identifying individuals at risk of developing diabetes, although further studies are necessary to assess other potentially impactful predictors, as well as the consistency and generalizability of model performance. In addition, future studies to examine the utility of the GDQS in screening for other noncommunicable diseases are recommended.

Identifying Glucose Metabolism Status in Nondiabetic Japanese Adults Using Machine Learning Model with Simple Questionnaire

“…However, prediabetes presents overlapping pathophysiology of impaired insulin sensitivity and secretion [25,26]. Although screening tools for prediabetes have been developed [14,[16][17][18][19][20], this is the first study to develop a model to identify the glucose metabolism status of individuals without diabetes. This model encourages individuals to understand their glucose metabolism status and learn how they should change their lifestyle to prevent diabetes.…”

“…The combination of variable selection and machine learning model resulted in high classification accuracy. Birk et al [16] developed a tool for screening individuals with high FPG using global diet quality score (GDQS) and lifestyle questionnaire responses. In this study, RF, generalized linear mixed model (GLMM), least absolute shrinkage and selection operator (LASSO), and elastic net (EN) were used.…”

Value of machine learning algorithms for predicting diabetes risk: A subset analysis from a real‐world retrospective cohort study

“…The findings suggested that the artificial neural network classifier based on demographic, lifestyle and anthropometric data was an effective predictive model 39 . Birk et al 40 . applied several ML and statistical methods, combined with survey data from the Food Frequency Questionnaire to calculate a global diet quality score, and developed a predictive tool for screening for prediabetes.…”

“…Birk et al . 40 applied several ML and statistical methods, combined with survey data from the Food Frequency Questionnaire to calculate a global diet quality score, and developed a predictive tool for screening for prediabetes. Findings showed that the model had a positive predictive effect.…”

Food Frequency Questionnaire Personalisation Using Multi-Target Regression

“…The use of machine learning seems an excellent option to work with existing, already validated questionnaires and adapt them to different needs and not have an expert involved to design the questionnaires by hand, but rather use expert knowledge to supervise and validate the machine-learning outcomes. However, to the best of our knowledge, machine learning has mainly been used just to estimate nutrient intake or to detect dietary patterns [ 8 , 9 , 10 ]. Dimensionality reduction methods such as Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) have been used in order to detect correlations between different food groups [ 11 ] also involving data gathered from FFQs [ 12 ].…”