Medical ‘Digital Twins’ Will Lead the Way to Personalized Medicine

We face a moment of opportunity—and competition—in bringing digital twin technology to patients

Two computer generated heads - one on the right in foreground out of focus with bokeh effect, second on the left in center of frame in focus

Simon Carter Peter Crowther/Getty Images

In March, China’s Xi Jinping announced that “digital twin” technology, which uses computer simulation to manage real-world systems, is one of six “new productive forces” in which his nation will lead the world and overtake the U.S. Amid this geopolitical jousting, this simulation technology—already widely used in manufacturing—is coming to medicine. It promises to speed up a long-envisioned era of personalized medicine, which uses targeted interventions customized to each patient, to maintain or restore health.

At the heart of a digital twin is a computer simulation that captures all relevant mechanisms and features of some physical thing, whether a city or an airplane. Consider Singapore, which has a digital twin of its infrastructure. Or the Pentagon, which will use one for its new B-21 bomber. The physical twin sends operational data continuously or frequently, often using the Internet of Things, to update the digital twin so that it faithfully reflects the current state of its physical twin, for example, reflecting accurately the wear on urban roadways or mechanical parts in a specific B-21 engine. In turn, simulation studies using the digital twin inform what we do to its real-life twin.

But what about medicine? Computer simulations can help us understand biochemical processes in our body and how different substances behave and affect our health. A digital twin of you or me or even parts of us, based on such computer simulations, could help drug developers design, test and monitor, and aid doctors in applying, the safest and most effective treatments or therapies that are specific and tailored to our genetics or biochemistry.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


There are already some successful examples of digital twins in medicine, such as an FDA-approved artificial pancreas that largely automates the administration of regular insulin injections for people with type 1 diabetes. In this application of digital twin technology, a monitor under a patient’s skin measures glucose levels every few minutes and transmits the data wirelessly to a digital twin on a handheld smart device. The mathematical model at the heart of the digital twin is calibrated to the specific patient to allow the simulation of their unique glucose metabolism. Using all this information, a control algorithm determines how much insulin the patient needs and instructs an attached insulin pump to administer the correct amount directly into the patient’s body. In this way, the digital twin carries out the glucose-regulating function that an unimpaired pancreas would normally have done. Other examples, described in a March Nature Computational Science report I co-authored, include a digital twin for the treatment of cardiac arrythmias and treatment optimization for breast cancer patients.

The time to develop this technology is now, thanks to three developments. First, we now have a greater understanding of how the body functions in health and disease. This has allowed us to create accurate computer simulations of human biology, ranging from the cellular level to the whole-body scale. Second, artificial intelligence and machine learning algorithms can continuously collect and analyze data from individuals as well as entire populations, allowing for constantly updated personalization of patient care. Third, the public has embraced personalized medicine as something that is beneficial and within reach. Likewise, the health care industry is increasingly relying on data-driven approaches that apply artificial intelligence to analysis of patient data, calculations of disease risk, and medical decision-making.

We must do more to scale up digital twin technology in medicine from scattered examples to a stable, mature, thriving industry, comparable to that of industrial digital twins today. First, just as NASA carried out early R&D for industrial digital twins in the U.S., public agencies, joined by private sources, need to launch a major funding initiative to support medical digital twin research to shoulder the substantial R&D start-up costs. It is important that these public/private partnerships are established in a way that assures that both costs and benefits are shared equitably. The European Commission is funding the comprehensive roadmap initiative Ecosystem Digital Twins in Healthcare, a large-scale interdisciplinary collaboration between industry and academia to bring this technology to patient care for cancer and other diseases. The recent report Foundational Research Gaps and Future Directions for Digital Twins by the National Academies of Science, Engineering, and Medicine has brought a lot of attention to such a move in the U.S. and has highlighted its far-reaching potential for biomedicine.

Second, we need a comprehensive framework for a supporting data ecosystem, comparable to the European Health Data Space. This European effort will provide a health-specific ecosystem comprising rules, common standards and practices, infrastructures and a governance framework, which empowers individuals to control their personal health data and, at the same time, provides a consistent, trustworthy and efficient set-up for that information’s deployment for research, innovation, and policymaking. Medical digital twins raise additional privacy concerns that will need to be addressed through legislation and regulation, since they are designed to functionally integrate a wide range of data that can capture information about a person’s biology as well as their lifestyle, potentially allowing forecasting of health needs, life span, and other features that have great potential to be misused. We need to answer questions about who owns and controls the data collected from people that is used to dynamically update their digital twin. Who will own our digital twin? We need to bring together all stakeholders in this debate to answer these questions now.

A final, and more challenging, requirement is funding the wide-scale implementation of this technology. We, as a nation, need to be able to afford the significant costs. Medical digital twins could be a powerful enabling technology for this purpose, by using it as a tool for predictive medicine. After all, one of the most important uses of digital twins in industry is for predictive maintenance, where it leads to considerable cost and risk mitigation. A similar effect could be expected in healthcare.

Medical digital twins are on the horizon, even though we are just at the dawn of this revolutionary new era. We should recall the history of aviation, where in the span of less than a century, we went from the first flight at Kitty Hawk to crossing the Atlantic in a Dreamliner jet, comfortably sipping champagne. But only a mere decade after the Wright Brothers’ maiden flight, aircraft design had advanced so rapidly that every major power in World War I deployed an air force. We are now at just such a moment of opportunity and competition. Medical digital twins may not yet be at the stage modern aviation is; but in their current state of development, they already offer important payoffs for curative and preventive medicine.

This is an opinion and analysis article, and the views expressed by the author or authors are not necessarily those of Scientific American.

Reinhard C. Laubenbacher is Dean’s Professor of systems medicine at the University of Florida. A mathematician and mathematics historian by training, his research focuses on the applications of mathematics to medicine. He is a fellow of the American Association for the Advancement of Science, the American Mathematical Society and the Society for Mathematical Biology, which he also serves as president-elect.

More by Reinhard C. Laubenbacher