Features Hub Opinion

Why go to the GP when you could send your health avatar instead?

Fri 26 Apr 2019 | Iain Buchan

The health avatar will harmonise machine learning, data analytics, and cloud technologies to make healthcare more interactive, personalised, and fulfilling

Professor Iain Buchan, public health and clinical informatics researcher at the University of Liverpool, coined the term health avatar. It describes a “digital mirror” of the future, that will cultivate an accurate picture of a user’s health and lifestyle over time using interoperable online services, new data streams and machine learning.

Among the many potential benefits of this digital reflection is the ability to match an individual’s health fingerprint to the care optimal to it; enabling individuals to “shop around” for treatment. A community of digitised selves will also enable individuals to connect with their “avatar kin” and specialist researchers.

Techerati asked Iain to outline health avatar fundamentals and its implications for the future of healthcare, wellbeing and fulfilment.

Why health avatars are the future of healthcare provision

A person’s health avatar is the interactive combination of their health and care records, programs, and preferences.

Sometimes it is a digital mirror, enabling a person to discover different potential health futures depending on lifestyle or self-care choices such as diet and physical activity alongside medication for type 2 diabetes. This digital mirror also allows a person to correct or fill in missing information through interaction, creating a more accurate picture of themselves, which in turn improves the record that guides their health professionals’ decisions.

Sometimes the avatar is a personal assistant, transacting on a person’s behalf, maybe ordering a repeat prescription or booking a visit for a blood test or consultation (which may be hastened or delayed, or moved between online or face-to-face, based on remote monitoring, and care pathway criteria).

Sometimes the avatar is the tutor of personal assistants, personalising interactions in ways that different assistants can learn from one another, leaving different signatures of your basic underlying “habit engine”, needs and life-contexts.

There may be no initial business incentive for competing companies/clouds with different AIs to cross-fertilise in this way, so the avatar is necessarily disruptive. Indeed, in Europe, the GDPR right to personal data portability might be interpreted as covering such personal ‘habit engines’, meaning a person’s health avatar must not be held hostage by being unreasonably binded to a single company’s technologies.

“I haven’t used the term “digital twin” as the data feasibly captured about a person’s health & care journey are very sparse”

Collaboration doesn’t naturally scale as competition does, but in a world facing increasing pressures on healthcare budgets, avatars that can ‘shop around’ for the best value care could be extremely valuable to society.

Why health avatar is a more appropriate term than digital twin

I haven’t used the term “digital twin” as the data feasibly captured about a person’s health and care journey are very sparse – future data will be more “high definition” but not conceivably close to a “twin”. Measurement error and missing data will remain a challenge for machine learning and humans trying to distil the models that constitute a person’s health ‘programs’.

Most current models that predict disease risks and treatment outcomes have been developed for looking at one disease at a time and making decisions about improving services for the average patient with that disease.

Real patients may have multiple interacting diseases and life-contexts. Science and technology are a long way from understanding how to build the union of models that accurately reflect a person’s health journey, learning more predictive accuracy with the data cumulating over a lifetime.

This kind of digital twinning is an important grand challenge, but long before it is achieved, relatively simple avatars can achieve much progress.

It may, however, be reasonable to characterise a person’s avatar as a digital twin if it attends a virtual clinic appointment on their behalf – the ease with which a multi-disciplinary team could review a patients “health journey” in a timely way may have great value.

A representation of how personal health records can integrate with healthcare, commercial governance, and research entities:

healthcareavatar — Source: Big data hurdles in precision medicine and precision public health, BMC Medical Informatics and Decision Making, 18(1) · December 2018

Interactions also extend beyond care providers. A person may wish their avatar to interact with researchers or their ‘avatar kin’ (people with similar health journeys) to advance knowledge and personal fulfilment. As a person’s avatar becomes a more “high definition” view of their health it will overlap with wider wellbeing and happiness. So, the avatar starts to broker a value chain cutting across many markets.

The ethics and governance challenges ahead

If health is top of the ethical tree then the health avatar is the acid test of ethics and governance for the personal assistant of personal assistants. I suspect pursuit of the health avatar will need something akin to W3C very soon – the economic implications of avatar communities may become profound. A long journey, but the first few practical steps are ready to take.

The UK’s formation of a grid of integrated care systems is a very promising nursery for the future health avatar. We need to focus that combination of personal records, programs, and preferences on bridging the care divides that commonly cause harm (e.g. discontinuities between social, primary and secondary care).