How digital twins and small data are making medicine more personal, precise, and human, according to Dr Dean Ho

Digital medicine, Dr Dean Ho explains, uses real-time health metrics to improve how interventions are delivered — whether in clinical treatment or preventive care. Rather than relying on fixed protocols or population-level assumptions, his approach builds digital twins for each individual, adjusting recommendations based on how they respond over time. 

Acknowledging that digital medicine is “a relatively new frontier”, he defines it “as the use of data, whether it’s obtained from wearables or through biomarker testing… to create digital twins that we can use actually to drive better treatment or intervention.”

Dr Ho, who moved to Singapore with his family from the United States in 2018, is provost’s chair professor at the National University of Singapore, where he also serves as director of the N.1 Institute for Health, director of the Institute for Digital Medicine (WisDM), and head of the Department of Biomedical Engineering. Across these roles, his work emphasises small data and clinical applicability.

Individual data over population models

This approach is applied in CURATE.AI, a clinical decision support system that sidesteps common challenges in healthcare artificial intelligence (AI), including dependence on large datasets and algorithmic bias. “We do not use population data to train algorithms to treat the individual. We primarily look at a patient’s response to varying drug doses,” says the biomedical engineer.

“This information is used to create a digital twin that does not require personal demographic data to construct. It’s based on understanding each person’s dynamics to hyper-personalise their own regimens.”

Using a small number of dose-response points, CURATE.AI calibrates to each patient and updates its recommendations throughout the treatment course. The platform is being applied in prospective trials for oncology, infectious diseases and post-transplant care. Unlike retrospective analyses, it supports real-time decision-making while ensuring clinicians remain in control. 

“Whenever we make dose recommendations, the clinician ultimately has the final say,” he adds. “Our work is truly a partnership between human and technology.”

Play

The programme, exclusive as it may sound, is not reserved for a wealthy few. CURATE.AI was designed to integrate with routine clinical practice. Dr Ho’s team interviewed clinicians, caregivers, and patients to better understand their daily needs. Their input influenced how the workflow was structured: From how and when data is collected to how dose recommendations are delivered.

“We built trust and empathy into our digital medicine tools by understanding what quality of life means to the users,” he says. “That feedback shaped the workflow around their lives, not the other way around.”

The team has also trialled the platform in infectious disease and post-transplant scenarios, including studies on dosage management in liver transplant patients. These studies demonstrate CURATE.AI’s ability to support decision-making in complex, evolving conditions without requiring large datasets or standardised demographic profiles.

Testing the model firsthand

To showcase how the same principles of digital twinning could be applied beyond disease treatment, Dr Ho initiated the DELTA study, using himself as the test subject. Over several months, he adjusted fasting duration, training intensity and sleep, tracking how his biomarkers responded.

“It’s actually an interventional study where I’m optimising sleep performance,” he says. “I was able to look longitudinally at my biomarkers as a function of varying things like duration of fasting and intensity of fitness training.”

What emerged has helped shape a broader effort to guide development. “Some of the findings are generalisable to the masses, and what we’re doing is taking some of these learnings and then crafting a much larger study for the population.”

The study reflects a growing focus at WisDM and the N.1 Institute on healthspan. “Young people are getting sicker earlier. People are also living longer, but their years spent in good health don’t match up,” he says.

What began as a personal experiment has since influenced those around him. “A totally unexpected outcome of the DELTA study so far is that it motivated my kids to prioritise their sleep — because they see me, their father, doing it,” he shares. One measure of success, for Dr Ho, is whether these efforts help others adopt more sustainable health behaviours.

That same principle of measurable change driven by personalised data is reflected in CURATE.AI’s clinical trials. A published case report described how a patient with a rare cancer remained on treatment longer, avoided transfusions, and required a lower drug dose — all while maintaining clinical benefit.

“Our trials are prospective, which means we are guiding actual dose recommendations, and to see patients do well on these trials is always a strong motivator for the team.”

Measurable outcomes and long-term value

As Dr Ho’s work shows, real progress depends on more than innovation and technical solutions alone. “Technology alone will not improve healthcare,” he says. “We need to pair unique technologies with clinical experience and a science and technology-driven ecosystem.”

His team structure reflects that thinking. In addition to engineers and clinicians, it includes social scientists, behavioural economists, industrial designers, health economists, and, critically, patients and caregivers.

“It really takes a redefinition of what multidisciplinary means to take a technology from ideation to real-world validation,” he says. That diversity of expertise, he adds, is essential for moving an idea from the lab to the patient in a way that works.

Progress took time and persistence, says Dr Ho. “The work that we do is different, and it took time to validate our ideas and bring them into clinical trials,” he emphasises. “Ultimately, we can redefine what impact means: life-saving outcomes.”

To make medicine as adaptive as the individuals receiving it, “we want to expand our approach of viewing each person as a Story instead of a Snapshot. (After all) Healthcare is about people.”