The Digitalisation of Healthcare

Introduction

In recent years, technology has slowly embedded itself in almost all human activity. And not just in a supporting role. In fact, it is paving the way to new practices and pushing innovation forward.

Healthcare is one of those sectors that can benefit from incorporating innovative technology; like Artificial Intelligence, for instance. Let’s take Machine Learning, or ML for short. ML is the concept at the foundation of web search engines; it helps you find what you might be interested in based on previous searches. It is commonly used to find patterns in large amounts of data. Interestingly, the more data is fed to it, the better the ML algorithm does its job. The question is: can we translate this into healthcare?

Picture it: a national, secure, digital healthcare database, shared by all healthcare providers, with countless health records from across the country. Your doctor feeds your data to the intelligent analytics engine, which compares not only your symptoms but also your medical history, in a privacy-preserving manner, to those of other people. The machine learning algorithms work quickly and effectively, enabling your doctor to narrow down your diagnosis more easily to a few options, saving time and resources.

That may seem a futuristic utopia, but it’s not as far away from reality as we might think.

The FAITH Concept

Let’s have a look, for example, at what the Irish-led, EU-funded FAITH research project is trying to achieve. Launched in January 2020, FAITH is aiming to provide an AI-based mobile application that supports patients who have undergone cancer treatment. The goal is to monitor and safeguard their psychological well-being.

Almost one-fifth of cancer patients and survivors experience episodes of anxiety and depression. Consequently, they can experience a diminished quality of life and a longer recovery path from their illness. Therefore, improving their mental health status is crucial for their overall well-being.

Usually, healthcare personnel can identify and promptly treat symptoms of mental distress and depression while patients are under hospital care. However, this becomes more difficult when the patient returns home and regains some freedom after hospitalisation. In addition, medical follow-up might focus mainly on the patients’ oncological profile, thus neglecting other relevant factors, such as mental health aspects.

Enter, the FAITH solution. An App on the user’s phone will monitor their conditions by combining proactive engagement (through interactive questionnaires) and passive tracking (i.e., sleep, activity, and outlook tracking). Then, a Machine Learning algorithm will analyse the data collected, looking for targeted depression markers and identifying potential downward trajectories.

This will occur without the need for the patient to have face-to-face consultations with their doctor. Finally, only in case of detection of a possibly negative trend, the App will notify the patient’s health team, allowing well-timed intervention.

Artificial Intelligence and Machine Learning

The market for healthcare technologies is giving strong signs of growth, suggesting a shift towards a more patient-centric, digital-service-enabled healthcare system. Besides the FAITH solution, concepts such as AI and Machine Learning are finding increasing applications in other critical phases of healthcare, like diagnosis. But what is Artificial Intelligence, exactly?

At its core, AI is a wide-ranging branch of computer science that deals with building smart machines able to perform tasks that usually need human intelligence. At the basis of any AI there are algorithms; the harder the task, the more complex the algorithms ruling it. Ideally, a good AI is capable of rationalising and taking actions to achieve a specific goal.

Machine Learning is a specific kind of AI that focuses on using algorithms to find patterns in large amounts of data. This enables them, for example, to accurately predict what could occur next and provide recommendations to support decision-making.

What separates Machine Learning from regular computer software, then? Well, mainly the fact that in regular software the coder writes specific instructions on how a computer must perform; instead, machine-learning models are taught how to reliably make informed decisions, based on their training on large amounts of data.

Tackling technological challenges

Artificial Intelligence is primed to revolutionise healthcare. Such an ambitious objective doesn’t come without concerns to address, though. Whenever new technological solutions are found, new issues arise and these need to be resolved to facilitate adoption. The same is happening with the application of AI to healthcare.

Handling medical records is not easy, as health records contain some of the most sensitive personal data. When it comes to sharing this data to an AI platform, a possible solution to guarantee users’ privacy, which FAITH adopts, is relying on a specific type of Machine Learning, namely Federated Learning.

One of its features is that Federated ML does not require to move the data to a centralised cloud to analyse it. Instead, the computation moves to the data, thus setting and updating algorithms and models directly to the user’s device. As the data never leaves the phone, the user’s privacy is safer. Furthermore, Federated ML has the advantage of producing more personalised models, which adapt to the users’ specificities.

Another issue is that regarding trust: in order to accept what a machine suggests, physicians and healthcare providers need to understand what’s behind its logic and why it has produced a certain output. That is, for instance, why it predicted negative trends in the mental health of a certain patient. This transparency can be achieved by adopting an Explainable AI framework; something that FAITH factored in when designing its solution, to allow healthcare providers to check on the system’s accountability and decision-making process.

A human challenge: designing healthcare solutions

Digital tools like in-phone Apps could have a huge impact on healthcare. Nonetheless, as much as any other App, if it just sits there and doesn’t engage its users, it is no better than a GPS tracker, monitoring movements but unable to retrieve any other data.

This is utterly important for healthcare Apps because mere passive tracking can only collect a limited amount of quantitative data; for example, how much the user sleeps, but not sleep quality. In order to have a complete picture of people’s well-being, gathering also qualitative data is fundamental. Unfortunately, this data is impossible to collect passively: it requires active interaction with the users.

Sometimes, when designing Apps for healthcare purposes, keeping the data collection both smooth and efficient is a major challenge. How can the App gather enough valid data without vexing the end-user with too many questions or requests for input?

This challenge can be faced by designing the App around the user experience, a term that encompasses all aspects of the end user’s interaction with the tool. This is a crucial step for the App to be easy to adopt by its end-users; a step that was pivotal for the FAITH project. In fact, regular user engagement consequently improves the quality of data. Keeping the end-users at the centre of the FAITH concept is crucial to delivering a successful assistive digital tool able to support both cancer patients’ well-being and their healthcare providers’ efficiency.

FAITH is led by Waterford-based Walton Institute and is being delivered by a multi-disciplinary team of clinicians, SMEs, and research institutions, enabling the needs of all stakeholders to be met with the essential expertise of scientists, health care specialists, technologists, and business partners.

Conclusion

The market for healthcare technologies is growing, and AI-powered tools and virtual interactions between patients and care providers are multiplying. Data-driven interoperability and personalised diagnostics will accelerate the delivery of predictive, analysis-driven precision care. This will pave the way for better and more efficient decision-making for patients and healthcare teams.

With an increasingly ageing population, the demand for technologies and services of care supporting patients after discharge from the hospital will grow as well. Acceptance of digital technologies and new forms of healthcare delivery has grown as a result of the recent COVID 19 pandemic. Therefore, the rising demand for digital healthcare services and their adoption suggest one thing; computation, AI, and data-driven, personalised medicine is on the way to becoming a staple of healthcare.

The digitalisation of healthcare could potentially be a major turning point in the sector’s history; so much so that, in the future, we might regard it as the introduction of penicillin. The path to achieving truly helpful digital assistive tools is still in the making, but so long as research projects like FAITH will keep working on it, established, trustworthy healthcare Apps will come sooner than we expect.