Digital Twin Technology Has Potential to Redefine Patient Care

Imagine if you could create a digital clone of yourself that can be used to test various treatment options to determine which one is best for your real self. 

By John Halamka, M.D., President, Mayo Clinic Platform and Paul Cerrato, MA, senior research analyst and communications specialist, Mayo Clinic Platform

As strange as this may sound, it’s no longer in the realm of science fiction. One source states: “A digital twin is a digital replica of a physical object, person, system, or process, contextualized in a digital version of its environment. Digital twins can help many kinds of organizations simulate real situations and their outcomes, ultimately allowing them to make better decisions.” In the inanimate world, that could include a computer model of an airplane that can be used to test out new design concepts or safety features, or a digital twin that simulates the functioning of a piece of machinery that needs an update. According to consulting firm McKinsey & Co, it’s likely about $73.5 billion will be spent on this technology by 2027, including product, data and system twins.

A recent study from Scientific Reports illustrates how the technology can be used in a medical setting. Indian investigators enrolled over 1,800 patients with type 2 diabetes and created a digital twin (DT) for each patient that simulated their metabolic status, dietary intake, blood glucose levels, and lifestyle habits, enabling the twin to predict a patient’s outcomes. Shamanna et al explained that: “The DT system continuously collects and analyzes data from various sensors and inputs, allowing it to offer personalized dietary and lifestyle recommendations that are precisely calibrated to minimize PPGRs [postprandial glucose response] and improve overall glycemic control. The DT platform will suggest the right food to the right participant at the right time based on current readings. The behavioral nudges provided by the digital twin were accompanied by human coaching.” This kind of individualized advice goes way beyond what is offered diabetic patients, who typically receive recommendations based on static guidelines and the results of the blood glucose.

The study results suggest DT technology works. At one year follow-up, patients on the program saw significant improvement in hemoglobin A1c levels, with a drop of 1.8%; 89% achieved a drop to less than 7%, a goal that diabetes experts recommend. They also required significantly less anti-diabetes medication, and experienced better weight reduction and less insulin resistance.

Similarly, researchers from Yale Medical School, University of Oxford, and elsewhere have shown that digital twin technology has the potential to transform cardiovascular medicine. Thangaraj et al explain that these cardiac replicas can “enhance disease phenotyping, enrich diagnostic workflows, and optimize procedural planning. Digital twin technology is rapidly evolving in the setting of newly available data modalities and advances in generative artificial intelligence, enabling dynamic and comprehensive simulations unique to an individual. These twins fuse physiologic, environmental, and healthcare data into machine learning and generative models to build real-time patient predictions that can model interactions with the clinical environment to accelerate personalized patient care.”

In simple English, the healthcare data they refer to integrates various diagnostic procedures—e.g, ECGs, cardiac imaging, and vital signs—with several other multimodal sources, including content from an individual patient’s EHR chart, their lifestyle decisions, exposure to climate change, medications, environmental toxins, and so on. These resources enable clinicians to make predictions about what is likely to happen to the real patient being profiled. Studies suggest, for instance, that digital twins may help cardiologists estimate an individual’s risk of ventricular arrhythmias if they already have ischemic cardiomyopathy by using certain anatomical substrates, triggers, and modulators. 

Phyllis M. Thangaraj and her colleagues provide an example of how the technology might work. Ms. K, 76 years old, has heart failure, preserved ejection fraction, type 2 diabetes, obesity, and hypertension. Her electronic health record offers additional data, including an enlarged left atrium, and a note recommending her diuretic be paused because hydrochlorothiazide has lowered her potassium level and increased uric acid. Her digital twin model is created based on all her data and runs a simulation of different blood pressure and diuretic drugs, comparing it to other patients with similar profiles. The twin also takes into account the latest guidelines and RCTs, finally recommending the patient be put back on hydrochlorothiazide and several other medications.

While this scenario is not yet within reach of most healthcare providers, it has the potential to profoundly transform patient care!