Six minutes into Apollo 13’s mission to the moon In 1970, its oxygen tank exploded. The event prompted NASA to develop a new approach to predicting possible failures in its spacecraft. The approach relied on continuous sensor data, which then fed deep digital simulations, enabling much more rigorous testing of complex spacefaring systems. It was the very first use of “digital twin” technology.
Today, digital twin systems are used across industries to improve operations and accurately simulate any change in a system. Tech companies like Apple and Tesla use digital twins to monitor product performance in the field and determine whether or not specific system components require maintenance.
Digital twins have also been used in healthcare, albeit largely in drug research and development. Its greatest potential, however, is in chronic disease management. By coupling machine learning and Internet of Things technology with digital twin AI, an approach that originated with something as vast as space exploration has the potential to make healthcare truly individualized.
Digitizing traditional care has failed
Modern medicine has made incremental moves toward personalized care over the past decade by giving patients a voice in decision-making, and toward precision medicine through advances in genomic research. Both have helped tailor care to the individual, but for the most part, our healthcare system takes a “large group” approach to care delivery.
It’s evident in the way we manage chronic disease. Every one of the 133 million Americans currently living with one or more chronic diseases is set upon a planned care pathway – a treatment regimen, a fad diet, often a number of medications – and their improvement is measured in batches of thousands of other individuals who share their condition.
This approach hasn’t worked. Notoriously, U.S. spending on diabetes, heart disease, and cancer continues to rise, and technology’s impact on outcomes and costs has been limited. In digital management of diabetes, weight loss, and other conditions, that impact has been a non-factor.
In March, a report published by Peterson Health Technology Institute underlined this lack of sustained results. The report found that all of the evaluated solutions perform poorly on engagement and outcomes over time. As a result, weight loss, A1C reduction, medication elimination, diabetes reversal, and the health, well-being, and economic benefits of these solutions are both limited and unsustainable.
That’s because most solutions just digitize an ineffective template for care. They don’t account for individual differences. Every person brings their own set of cultural, biological, dietary, behavioral, and environmental factors that influence their health at a deeply individual level.
Moving from ‘personalized’ care to individualized care
Digital twin AI promises a departure from the template. Core to the technology is the concept that every individual is an N of one. An individual’s digital twin is informed by a continuous measure of their unique clinical and behavioral variables, and uses that data to shape care guidance toward the best and healthiest version of that individual.
The power of digital twin technology is in its attention to the small things – the things we eat and do – and how they impact our current and future selves. In practice, digital twins can accurately predict the effect a steak dinner will have on a specific person’s metabolic or cardiovascular health. To the extent that impact may be negative, digital twins can offer ways to mitigate the repercussions. It might suggest a 10-minute walk or an alternative dessert. Instead of ice cream, maybe it’s banana nut bread with Greek yogurt and fresh berries or simply a different sequence.
In this way, digital twin AI can show an individual what’s in store for them if they stay on their current trajectory and the big changes that can occur by making small adjustments over time. Keep up your current routine, and you’ll be able to stop taking metformin in three weeks. Fall back into old habits, and you can expect to pick up a refill.
It’s potent technology, and while its impact on healthcare has largely been recognized only in academia, it is beginning to find its role in commercial use cases. In 2014, Dassault Systemes and the FDA launched SIMULIA Living Heart, a project that works with device manufacturers to develop and refine cardiac devices at a faster pace. At the onset of the pandemic, OnScale’s Project BreathEasy developed a digital twin of the lungs of COVID-19 patients to improve and optimize the use of ventilator resources.
Medical researchers are also using digital twin disease models to predict the effectiveness of pharmaceutical interventions based on complex, extremely individual biological processes. Takeda Pharmaceuticals has embraced the technology to shorten pharmaceutical processes and make realistic input-output predictions for biochemical reactions. More recently, researchers used digital twin technology to simulate therapy outcomes and determine the best treatment for oropharyngeal carcinoma based on the individual.
Chronic disease management is the next frontier
A recent paper published in Nature asserts that digital twins are “poised to make substantial contributions” to cancer care, especially in monitoring the progression of the disease and evaluating treatment responses, which infamously vary individual by individual. The same paper analyzes cardiac digital twins fed by imaging, EHR, genetic, and continuous wearable data, and their potential to predict acute cardiac events.
These advancements will give way to life-changing healthcare technologies. Their power lies in a concept core to their purpose: nothing complex is static.
This is especially true of our biological systems. A digital twin requires thousands of data points per day, per individual, to truly understand the interplay between an individual’s biology, culture, lifestyle, preferences, and health. Some of this data is already being captured by wearables and mobile apps, but without a model that puts that data into the context of the individual and their care journey, it is rudderless.
In the world of chronic disease management, the small things can very quickly become big, life-threatening things. And while digital health has raised the hopes of patients with language like “personalization,” the tools and approaches that have been offered to people have not addressed their unique needs and preferences.
Digital twin AI will turn this approach on its head by helping us better understand and improve our health on a deeply personalized level. It’s a technology poised to fulfill the promise of individualized care.
