Skip to main content

Why health care needs to up its dose of digital twins

(Image credit: Shutterstock.com)

As the cost of health care continues its relentless ascent, global health spending is expected to top $10 trillion by 2022, according to the Economist Intelligence Unit. Not surprisingly, payors, providers, life sciences companies and other members of the health care ecosystem are eagerly exploring ways to leverage emerging technologies to stem the rising costs and improve patient care. One such technology is digital twins, which incorporates Big Data, artificial intelligence (AI) and sensors to create virtual replicas of a physical object or process.

As a concept, digital twins is not new. Various industrial sectors such as aerospace and automotive have been using them for years to optimise the operation and maintenance of assets, systems and manufacturing processes. Across industries, the digital twins market is projected to zoom to almost $36 billion by 2025, up from about $3.8 billion currently, says research firm MarketsandMarkets. And among those jumping on the bandwagon are a growing array of health care organisations.

As a technology, digital twins show great promise in health care. By synthesising and tracking patient data through medical records, medication history, imaging studies and patient-provided health data, health care professionals have the potential to identify information gaps in a patient’s medical records. They can synchronise all sources of data to create a complete picture of a patient’s needs. Such a view would allow medical professionals to run treatment simulations without harming patients. It would also prevent incorrect diagnoses, avoid unnecessary or ineffective treatment, and improve outcomes.

Challenges to overcome

To realise the potential benefits, however, health care organisations must overcome a host of challenges. These can include regulatory challenges, such as how and where data is processed; this can be addressed by hosting and analysing the data locally rather than overseas. Data siloes are another challenge that must be overcome. Over time, data can amass across multiple servers that don’t communicate with each other – or IT staff might not even be aware of such data. An end-to-end data management system can break down, unify and consolidate siloes to gather the relevant information for digital twins.

Already, health care professionals are using digital twins in a number of innovative ways. In a ground-breaking initiative, the Living Heart Project is using the technology to create 3D twins of patients’ organs such as their heart or liver. Thanks to connected sensors, doctors can monitor organs remotely – without the need for physical check-ups. GE, Alphabet and IBM are all using digital twins across medical imaging, preventative care and drug discovery with the aim of delivering fully personalised medical care.

Cathay Industrial Biotech is using digital twins within a digitised factory setting to reduce waste. Usually, when a batch of test drugs is discarded, most of the lab and manufacturing equipment needs to change, consuming time and resources. Digital twins simulates the drug production making the entire drug development process efficient.

Customising digital twins for health care

The successful creation of a digital twin depends on several factors. First, there needs to be a clear vision of the outcomes expected and a realistic assessment of the costs. The people involved with creating and running the digital twin need deep knowledge of the application domain, as well as a thorough understanding of the human aspects. Consideration must be given to the human demands and expectations of patients, health care providers, insurance carriers and others in the global health care ecosystem.

Additionally, for a digital twin to work well in a health care environment, the technologies and architectures must be scaled to meet a range of requirements. A number of key building blocks are needed. These include:

  • Connectivity and networks: The architects of solutions need to understand and identify the sources of information available for creating a digital twin, while also defining a strategy for obtaining the information that isn’t available. This includes techniques used to glean data from sensors and other devices and feeding that information into the digital twin to simulate the real-world scenario in a virtual environment
  • Data intelligence: A data analytics strategy is needed to transform the data into manageable information based on the needs of each application
  • Simulation: Large-scale problems need to be simulated on the digital twin using physics, math and machine learning models so that problems can be debugged without affecting the equipment already on the premises – be it in a hospital, lab or factory.
  • Simulation: Large-scale problems need to be simulated on the digital twin using physics, math and machine learning models so that problems can be debugged without affecting the equipment already on the premises – be it in a hospital, lab or factory.

Mission-critical digital twins

Now is the time for health care CTOs and CIOs to make sure they understand the requirements for implementing digital twins in order to build an underlying data foundation that can run all applications. Like any new technology, there will be challenges to address, as mentioned. Digital twins requires a shift in mindset. It is a move from transactional systems to a far more cooperative process. The technology heavily depends on the intelligence that is developed incrementally – which, in turn, means that the value it creates is incremental too.

Consequently, change management and culture issues have to be considered when developing and implementing a digital twin. Other factors play a role too, including the ownership of and accessibility to reams of data; compliance with GDPR and other legislations; and the ability to ingest a wide variety of data at different times and rates.

Ultimately, it’s imperative to recognise that a digital twin is about much more than managing and analysing data. It creates a complex computing architecture to connect different parts of a mission-critical ecosystem. That system might be an individual patient and the systems related to the body, or it might involve gathering information about different hospital equipment and lab apparatus within a health care setting. Regardless, the goal is the same: to glean actionable insights that promote faster and more accurate decision-making while delivering improved patient outcomes.

Subhankar Pal, Assistant Vice President of Research & Innovation, Altran