Much of modern healthcare remains reactive: patients present with symptoms and clinicians respond. What if, instead, we could anticipate illnesses long before symptoms emerge using the ‘digital twin’: a comprehensive model that dynamically integrates, analyses and simulates a wide array of data from the electronic medical record (EHR) allowing for pattern recognition and projection, allowing providers to intervene along a clinical trajectory to prevent or treat earlier in the disease course. Unlike standard EHR analytics, which typically produce static population-level predictions from historical clinical data, a medical digital twin is a continuously synchronised, patient-specific, computational model, enabling in silico simulation of disease trajectories and therapeutic interventions for a single patient. A digital twin goes beyond simple ‘analytics’ as it allows ‘n of 1’ studies, simulates future trajectories and asks counterfactuals. Moreover, in the digital twin construct, the virtual and physical twins are bidirectionally updated, converting biological and physiological knowledge into computational models to inform clinical decision-making.1 This offers an immense opportunity to efficiently and accurately predict clinical trends proactively, moving us from a treatment mindset to a preventive paradigm.
Ahuja et al. (Mon,) studied this question.