Self-care and the future of digital health

So, patients will be able to manage extensive therapies at home. Sounds marvellous, right?

Let’s take a step back. Digital therapeutics have a lot of potential, however as with any new innovation, they need to be approached with careful consideration. According to Dr Eric Topol, author of “The Creative Destruction of Medicine”, there are four key areas we need to tackle to make sure digital therapeutics are deployed safely and effectively:

1. Use of AI for clinical paperwork
2. Use of machines to reduce errors
3. Harvesting of data to improve outcomes
4. Treating patients at home, not in hospital

You can read Dr Topol’s most recent review here.

The use of AI and machines to remove burdens and reduce errors seems like a no-brainer. However, in delegating control to machines, we risk creating unknown issues for both healthcare professionals and patients. Such issues can include inherent negative bias in the system, whereby the creator has not correctly allowed for their own gender, racial and age stereotyping when designing them. The short-lived Genderify system, as well as a study by Harrisburg University on facial recognition to identify criminality, both fell victim to this. As a result, the biases and beliefs of these creators, however unintentionally, were transferred into the systems they created.

Clearly, AI and machines are only as capable or as fallible as their creators, something that needs to be considered in the development of any digital health solution.

Harvesting data can also have its drawbacks when the systems used to analyse data are not created using an appropriate level of due diligence. The data gathered from users need to be collected and shared anonymously. Certain markets are also more comfortable with sharing data than others. In the UK and the rest of Europe, users are used to secure data sharing and tend to trust the people they share it with. By contrast, in the USA many users are wary of how companies may be using their data. The USA doesn’t have the same protections afforded to people in Europe, so it’s worth noting that the security of data sharing will need to be reinforced for any digital therapeutic in this market, if we are to gain trust from our users.

When we consider Topol’s fourth point, treating people at home, it’s important to consider how this will apply across the globe. Technology and innovation in the African continent, for example, sometimes happens in a very different way to traditional western economies. SMS services are used for cash approvals, for example, as data networks are not widespread, whereas 2G phone networks normally are. One of the interesting problems for digital therapeutics will be how a patient issue can be diagnosed by text message in places where primary care centres simply don’t exist.

Changing behaviours

In addition to Dr Topol’s suggestions, user behaviour is another key area that needs to be considered with the introduction of digital therapies.

Medical apps play a large role in digital therapeutics, however it’s a huge behaviour change to shift healthcare to people’s smartphones, one that needs careful thought and standardised measurement. This can be done through frameworks such as ABACUS, an assessment framework designed specifically to measure the efficacy of apps in how they influence behaviour change.

Typing or talking?

We use our smartphones for everything, typically with a default method of interaction: typing. We type messages and searches, record notes, write lists, add calendar invites – all using text input. When considering digital therapeutics, should we still choose text as the default input method?

As highlighted by Jennifer Heaper, Vixen Labs, at the recent MUXL Festival of UX & Design, there’s evidence from studies by Mindshare and Neuro-Insights that interaction with voice systems like Alexa is much less taxing on the brain compared to performing the same actions using text. As a result, we may be adding cognitive load and impacting adherence by defaulting to text-based digital solutions. AI systems are ideal for linguistic analysis, so you can begin to imagine voice interfaces becoming the norm for adding patient feedback rather than asking patients to fill in endless forms.

Conclusion

Ultimately, there is a lot of promise in digital therapeutics as a tool for self-care by patients outside of the hospital. To get there however will require careful thought, planning and consideration. As the COVID-19 pandemic has shown us, now is clearly the time to be thinking of ways to alleviate pressure on our healthcare systems and find new ways to help patients to help themselves.

 

Extra reading

Amisha et al. “Overview of artificial intelligence in medicine.” Journal of family medicine and primary care vol. 8,7 (2019)