Digital Health

Digital Twins in Healthcare: Simulating the Patient

Published 5 min read Acuitas Health Analytics

Digital Twins in Healthcare: Simulating the Patient

Engineers have used digital twins for years to model jet engines, factories and power grids. A digital twin is a virtual copy of a real system, linked to it by data, so the model changes as the real thing changes. In health care, the idea is to build a digital twin of a patient, an organ or a hospital to test choices before acting on them.

An official definition

The US National Academies of Sciences, Engineering, and Medicine studied digital twins in 2023. They define a digital twin as a set of virtual models that mimic a real system, are updated with data from it, can predict, and inform decisions that create value [1]. The key is a two-way flow: data flow from the patient to the model, and decisions flow back [1].

What makes it different from a simulation

A normal simulation is built once and run. A digital twin keeps learning from new data about the same person or system. The National Academies stress that tight links between models, data and decisions are what make a digital twin [1].

Possible uses in health care

The National Academies workshop on biomedical digital twins noted that they can represent molecules, cells, tissues, organs, whole patients or populations [2]. Possible uses include:

  • Heart: modelling a patient's heart to plan ablation or device placement.
  • Cancer: predicting how a tumour may respond to different treatments.
  • Diabetes: simulating how blood sugar responds to food, exercise and insulin.
  • Intensive care: modelling organ function to guide fluids and drugs.
  • Clinical trials: creating virtual control groups to reduce the number of people needed.
  • Hospitals: modelling beds, staff and patient flow to plan capacity.

Research gaps

The National Academies found gaps in the basic mathematics, statistics and computing behind digital twins, plus wider challenges in turning them into practice [1]. In medicine, these include:

  • Bodies are complex and change over time, often in ways we cannot measure.
  • Data are incomplete, noisy and spread across many systems.
  • Models must show how certain they are, so doctors know when to trust them.
  • Checking that a model is right for one patient is hard.
  • Privacy and security rules apply to the data that feed the twin.

Data sources

Digital twins need rich, ongoing data: health records, imaging, lab results, genetic data and readings from wearables. About 36% of US adults used a wearable health device in 2022 [3], which could feed some twins with daily data. Better interoperability between health systems is also needed.

Where things stand

Most medical digital twins today are research projects or narrow tools, such as heart models for planning procedures. Full "whole patient" twins are still a long way off. Progress will come step by step, organ by organ and use by use.

Key numbers at a glance

  • National Academies report on digital twins: 2023.
  • Biomedical twins can model molecules, cells, organs, patients or populations.
  • US adults using wearables: about 36% (2022).

Questions to ask about a digital twin tool

  • What decision does it support?
  • How often is it updated with new data?
  • How accurate has it been for real patients?
  • Does it show how certain its predictions are?

Go deeper

Want the full picture? Our report covers market size, growth, segments, countries and the key companies. Read the report or get a free sample.

Digital twins of hospitals

Not all health digital twins model the body. Some model hospitals: beds, operating rooms, staff and patient flow. These twins use live data from scheduling and records to predict bottlenecks, test new rotas or plan for surges. Because hospital processes are easier to measure than human biology, these twins may deliver value sooner.

Virtual patients in trials

Drug developers are exploring "virtual control arms", where models predict what would have happened to a patient without treatment. If accepted by regulators, this could reduce the number of people who need to receive a placebo. Strong validation and transparency would be needed for regulators to trust such models.

Human in the loop

The National Academies highlight that the two-way link between model and real system includes both automatic and human feedback [1]. In medicine, a doctor must decide whether to act on what a twin suggests. Twins should support, not replace, clinical judgement.

Building trust

  • Show how the model was built and tested.
  • Report how certain each prediction is.
  • Track whether predictions came true.
  • Update models as new data arrive.

A digital twin of a person holds a detailed picture of their health. People should know what data feed it, who can see it and how it is used. Clear consent and strong security are as important as accurate models.

In short

Digital twins link models, data and decisions in a loop.

What this means

Digital twins could make care more personal by letting doctors test options on a model first. The idea is powerful, but the science is young. Early value will come from focused tools, such as heart and diabetes models, and from hospital planning, before whole-patient twins arrive.

Where to read more

All figures in this article come from public sources listed below. For more detail, start with the National Academies report on foundational research gaps for digital twins. Each is free to read online. If anything here raises a question about your own health, take this article to your doctor or pharmacist and talk it through together.

Explore on Acuitas

Topics: Digital health · Artificial intelligence

Sources

  1. National Academies (2024): Foundational research gaps and future directions for digital twins
  2. National Academies: Opportunities and challenges for digital twins in biomedical research – workshop in brief
  3. Usage trends and data sharing practices of healthcare wearable devices among US adults (HINTS 2022)

Banner image: CreativeTools.se - PackshotCreator - 3D printed - ZPrinter - Heart-shaped cookie-cutter -… by Creative Tools from Halmstad, Sweden, CC BY 2.0, via Wikimedia Commons.

Topics: Digital health · Artificial intelligence

Back to Blog