Frontiers in Digital Health · Nov 13, 2026

The Digital Twin: A New Blueprint for Precision Nutrition-A Scoping Review

By Yutika Rajan Shirgaonkar, Devaki Gokhale

Introduction: Digital twins (DT) are artificial intelligence-driven virtual replicas that simulate real-time metabolic responses, integrating wearables, Continuous Glucose Monitoring (CGM), and different types of physiological trackers for precision nutrition. DT has evolved from NASA's 1960s model application to an intelligent, multi-omics model for nutrition and healthcare. Objectives: To evaluate the current evidence on the DT role in personalised nutrition, focusing on the real-time IoT (Internet of Things) models as well as mechanistic in silico approaches for predicting metabolic outcomes and optimising interventions. Methods: An extensive search of three databases, Scopus, PubMed, and Web of Science, was narrowed down to 155 records. After screening for inclusion and exclusion criteria, 15 peer-reviewed studies met the PICOT criteria. Data extraction covered the study design, DT typology, artificial intelligence (AI) /machine learning (ML) algorithms, IoT integrations, and outcomes. Results: DT mostly used multimodal IoT devices, such as CGM, smart scales, BP monitoring devices, gut microbiome data, sleep, activity trackers, and food databases, with ML to predict glucose spike responses, achieving better outcomes than standard care (SC). Significant drops in HbA1c, weight and hepatic fat were observed. Type 2 diabetes and hypertension remission were also noted. Mechanistic DTs validated fasting/diet responses via Ordinary Differential Equations (ODEs). Whereas food DT helped improve the quality of the food product. Conclusion: Digital twin (DT) approaches show promise for personalised nutrition and metabolic health management. However, evidence varies across study designs, and their comparative effectiveness remains uncertain. Limitations include retrospective bias, short follow-up periods, Indian-centric data, and black-box AI. Future research should focus on multi-ethnic randomised controlled trials, hybrid ML-ODE models, explainable AI, and ethical frameworks.

The paperJournal siteInterventions

AgingResearch.News