The Promise of Artificial Intelligence: Evolving the Patient Journey
By Mélanie Bourlioux, Product Manager of Imagine, & Rie Christensen, Designer for Imagine
This is part 2 of a series on how health technologies impact our lives and seek to change the way patients interact with the healthcare system. In this post, we explore the inefficiencies of the traditional patient journey, and how we can use artificial intelligence to optimise healthcare.
You can read the first part of the series here, which raises the question: Should good healthtech make itself redundant?
A couple of years ago, IBM’s supercomputer Watson cracked a medical mystery that had baffled doctors for months. A female leukaemia patient in Japan didn’t respond as expected to treatments and eventually, the medical team turned to Watson for help. Barely 10 minutes passed before the results were in: The patient did suffer from leukaemia but it was a different and rarer form of leukaemia than the doctors initially had diagnosed her with. The supercomputer came to the conclusion after cross-referencing the patient’s genetic data with information from 20 million oncological studies.
Technology, like Watson, is changing the fabric of healthcare and holds the potential to transform the patient journey by empowering patients and augmenting doctors in their work — and the need is there for us to rethink the patient journey and create new paths to optimise healthcare through AI-powered technologies.
Patients get stuck in their journey towards recovery
If we look at the traditional patient journey, we’re faced with an often long, strenuous and ineffective path towards recovery. When discovering a skin issue, it can take weeks before the patient goes to see a healthcare professional where the risk of being misdiagnosed is high. In fact, you have a 50–50 chance of getting the right diagnosis for your skin condition when you see a general practitioner. You may go through several treatments before finding one that suits you — or discover that none of them work. And the waiting time to see a specialist is particularly lengthy due to the shortage of dermatologists — especially if you live in a rural area.
The World Health Organisation has estimated that there’ll be a shortage of almost 13 million healthcare workers in 2035. Within dermatology, this challenge is already painfully evident. In Germany, for example, there’s only 5944 dermatologists to serve a nation of more than 82 million. In the UK, about 650 consultant dermatologists are available to support a population of 66 million with more than 2000 different skin disorders.
And when 1 in 3 people will develop a skin disease at some point in their lives, it raises the question: Who will give the patients the support, they need?
Improving access and empowering patients through AI
Artificial intelligence holds the potential to drive meaningful change at several points along the patient journey while making different actions available to the patient.
More accessible healthcare is one way health technologies can improve the patient journey. Often, physical interactions can be substituted by online or telederm consultations thereby creating shortcuts in the patient journey and freeing up time for healthcare workers to focus on the patients who truly need a physical consultation. Health apps can also provide more personalised insights into your health and what actions to take.
Take the app Ada as an example. The AI-powered health guide aims at identifying the appropriate next steps when you discover some unwanted symptom. Ada thereby becomes a form of prescreening consultation. By plotting a number of symptoms into the app, Ada asks you a series of questions before delivering a report with suggested diagnoses and how you should act — for example “seek medical attention” or “can usually be managed at home”. The app was six years in the making to build the sophisticated digital medical knowledge database and train the underlying AI models — and with more than 5 million users today, the AI continues to learn and improve.
More accurate diagnoses is another driver for a smoother patient journey. With the increasing availability of healthcare data, AI-powered tools can detect complex patterns that may not be visible to the human eye. Our own skin tracking app Imagine demonstrates this as it, based on anonymised imaging data provided by users, now has developed an artificial intelligence that can diagnose psoriasis with an accuracy of 91%.
As an added bonus, the patients are far more likely to receive the right treatment faster when the starting point is a correct diagnosis. But even when no cure is available, health technologies can help patients manage their chronic condition and promote treatment adherence such as apps that allow you to monitor your health, track changes, and gain insights to promote well-being. When tracking your skin condition through the Imagine app, the goal is to better understand your disease and identify what lifestyle triggers you should be aware of.
For diabetes patients, BlueStar is another example of an app that combines coaching, education and insights into your blood glucose levels to help patients manage their condition.
These different paths along the patient journey give the patients a new form of agency. They become co-managers of their own health instead of passive recipients. Where healthcare professionals previously were the gatekeepers you had to go through to move along the patient journey, digital tools create new possibilities. By adding tech solutions to the mixer, only an estimated 10% of patients with a skin issue have to go through the long, traditional patient journey.
Think of AI as your copilot, guiding you along the way and providing you with more options for personalised health insights while augmenting doctors in their decisions through AI-powered diagnostic tools.
Humanising artificial intelligence for a better patient journey
Developing amazing tech solutions to influence the patient journey is one thing — another challenge is to create a user experience that generates the right value for patients.
So why do some AI experiences feel awesome and others… not so much?
First of all, designers have to understand the technology and identify the opportunity spaces for artificial intelligence. We also need a user-experience driven approach to healthcare, meaning designers have to make the patient journey human-centred. How? By humanising artificial intelligence. And by humanising, we mean how to take AI and connect it back to humans so it makes sense in the user journey.
We can divide some of the central opportunity areas when designing for artificial intelligence into three steps:
- Delegate complexity
- Implement empathy
- Score with assist
The strength of AI lies in its ability to do the heavy lifting on tasks that are either particularly complex, or repetitive and mundane. A good place to look when designing for AI is where users feel overwhelmed — or when people don’t know where to begin and where to end. Airbnb, for example, does this through its smart pricing system, which gathers information from more than 72 data points from different sources. Instead of leaving first-time hosts in the dark with regards to pricing, this system adapts your price to demand.
But when it comes to people’s health, it’s particularly important to build trust between humans and technology, and close the empathy gap. Here, designers can look to their parents for inspiration. Instead of simply giving smart answers, we need to consider the users’ intent — that is, taking their hand and leading the way like a caring parent would do, and create interfaces that allow for a more intuitive user experience. So ask yourself: “What information or recommendation can you give” or “what actions can you take on behalf of users” with the AI that will take them the whole way?
And last but not least, score with assist. If we’ve learned something from football, it’s that when you have good hand-offs, you’re way more likely to score. You need to keep the strengths and weaknesses of what you’re designing for in focus. Rather than trying to force everything into one device, pause for a moment and consider where you can provide value and where you’re better off leaving it to others.
Human AND machine will improve healthcare
The beauty of artificial intelligence lies in the ability to optimise a variety of existing practices — from your commercial flight using an AI autopilot to your e-mail spam filter that knows you’re probably not interested in inquiries from long-lost Nigerian “princes”.
But where AI goes so does the hype that your job is in jeopardy of being taken over by machines.
IBM Chief Executive Officer Ginni Rometty mentioned in an interview that she prefers the term ‘augmented intelligence’ over ‘artificial intelligence’ because the idea is that rather than replace, AI is going to help us when making important decisions — as doctors do when diagnosing patients or prescribing a certain treatment.
“I do believe that when it comes to complete job replacement, it will be a very small percentage. When it comes to changing a job and what you do, it will be 100% […] We really think this is about man and machine, not man vs. machine,” she said.
In a recent clinical study in Hamburg conducted by LEO Innovation Lab and Arbeitsgemeinschaft Dermatologische Prävention, 37 general practitioners were exposed to 31 patients with different skin diseases. With the aid of a digital clinical decision support system, the overall diagnostic accuracy was improved by 34% while patients also reported that they better understood the information provided during the consultation.
So doctors aren’t going to be replaced anytime soon but they will have new tools available to assist them in their work.
Mélanie and Rie are presenting at European Women in Technology Conference in Amsterdam on 29 November.
