Paging Doctor Bot
The rise of the bots — they’re everywhere! In this article we examine the benefits of using chatbots in healthcare.
Healthcare is a controversial topic. Rising costs, often attributed to factors such as increased life-expectancy and regulation, as well as a booming population add continuous pressure to reform and innovate in this space. What’s more, the quality of healthcare delivered is often questioned when compared to these rising costs.
There is a rich history of medical expert systems developed to reduce the increasing healthcare costs, meet new regulatory requirements, and satisfy wider requirements to improve the quality and effectiveness of healthcare while maintaining the operational efficiency of healthcare organisations. First there was the Telehealth movement, delivering healthcare at a distance. Then came the mHealth, or Mobile Health movement, making many similar promises but typically using the smartphone app as it’s platform for delivery.
In this article we touch on some of the biggest challenges faced by the sector and how chatbots can help shape the delivery of healthcare services.
So what’s gone wrong?
Trust and quality
When you visit your doctor, you trust that the advice given, or test and procedures recommended by them, are backed up by evidence-based guidelines developed as a result of numerous trials and tests, i.e. it has un unshakable foundation of truth. It’s your health after all! It’s natural to assume that the same assumption will be made for mobile healthcare applications.
There have been numerous cases of apps either not working or getting it wrong, whether it’s an app that falsely claimed to aid refugees at sea or misreads hypertension range. Reports like this make it harder to build trust around mobile healthcare apps. What’s more — the vast majority of healthcare apps are not regulated.
Security and privacy
Security and privacy are some of the biggest concerns to app users, even more so when dealing with healthcare data. As end users we expect applications to be free of malware or viruses, and that any data we provide won’t be used for anything we haven’t explicitly consented to.
Personal and healthcare data is strictly regulated. For examples see HIPAA in the USA, the Data Protection and e-Privacy Directives in the EU, and the DPA in the UK. From May 2018 the GDPR will affect every organisation that processes EU residents’ personally identifiable information. App developers need to be able to assure users that their apps comply with these regulations — something that is not visible
Connected electronic health records
Mobile health apps generate a lot of data. For this data to be most useful it needs to be merged from a variety of sources into a single central electronic health record and accessible by your GP or doctor.
Healthcare applications will have to be universally covered by insurers before widespread adoption is likely to occur — much the same as a new drug that hits the market. The process of getting a drug to market is a long and expensive one, and involves clinical trials and classifying it’s effectiveness as a treatment.
Poor user experience
The biggest blocker for mobile health app adoption is perhaps the poor user experience. Think about it — there are so many hoops to jump through to get the most out of a a healthcare app:
- You need to download and install the app.
- You may need to configure the app.
- You need to remember add data to the app. Possibly every day.
Entering data into an app not only reminds you of your illness, it also feels cold and impersonal. No wonder it seems like work! We don’t want to spend time ‘working’ on our mobile devices. We use them for communication, information and entertainment.
The Quantified Self (or life logging) movement incorporates technology into personal data capture to improve one’s health and well-being. Technologies (such as apps, GPS enabled devices and physical activity trackers) allow individuals to track all aspects of their daily lives, including their total activity, number of steps, food they eat, amount of sleep, heart rate, arousal, blood oxygen levels and mood.
Users set goals for themselves, view their stats and even share progress on social media. It’s highly personal, highly engaging and has element of gamification to it which increases engagement by users.
Quantified self-advancement have enabled individuals to quantify data about themselves that they never knew existed. In many cases data is collected automatically through wearable technology, but this isn’t always the case. Some data still has to be logged manually.
So what does this mean for delivering healthcare services via mobile devices?
It’s not enough for healthcare services delivered via mobile to simply connect and provide information. To maximise effectiveness it also needs to be delivered in a personalised and engaging manner.
Smartphone users spend more time than ever in messaging apps. This is where the benefits of a conversational UI become clear.
- Conversation is how we normally communicate — it feels natural. It’s familiar territory so there’s no learning curve.
- We spend more time than ever in messaging apps on mobile devices. If the trends are anything to go by this is likely to increase.
- Conversations can be personalised based on the users profile and responses. The tone of the conversation can be changed to match the users wants and needs, making for a much more empathetic exchange.
- Conversations are linear so it’s easy to remember the conversational context by simply checking the last message and picking up where you left off.
- Receiving alerts and prompting users to engage with the service feels natural and friendly. It’s a relationship rather than an interface. A conversation with a confidant.
- Last but not least, the simplicity in terms of what the end user sees. There is no designed interface, simply a familiar layout thanks to all the existing messaging apps and platforms out there.
Where are we today
Although there have been advances in AI, natural language processing is still not capable of generating convincing dialog. The longer the conversation, the harder it becomes to automate it.
So what are chatbots good for right now, you might ask? The simple answer is short exchanges. Chatbots are great for checking in with users on a regular basis and collecting data.
And the practical benefit to mobile healthcare services?
The service becomes an ongoing conversation. You always have an open channel of communication, and you can push messages right to the user’s phone without needing to rely on emails or websites.
Reduce human time required for routine activities
Your bot can take care of routine activities, such as checking in with users and collecting data by asking standard questions. Certain keyword responses or actions can trigger an alert to a human (case manager) to intervene in the conversation.
Improved quality and timeliness of data collection in the field
Because the notifications and conversations are happening in the customer’s favourite messaging app, friction is reduced and so we expect a higher response and engagement rate. There is no need for users to install an app, or log into a website and type into forms.
Rapid follow-up and intervention
Based to responses to questions you can set up rules to trigger alerts to case workers or practitioners. For example if a user mentions words related to suicide, or if they have not responded to us for 2 days, or if they have consistently answered negatively to the question “are you feeling positive about today”.
Anonymity is possible
Why do people feel more comfortable talking candidly with strangers? You can enable users to treat your bot as a confidant. It is not necessary to require users to create an account, or provide any details in order to converse with the bot. We can then request personal details as and when they become necessary for us to provide further services.
So who is already doing interesting things in this space?
The HealthTap chatbot allows users to type a question into Facebook Messenger. Questions are reviewed by any of the more than 100,000 physicians doctors that span 141 specialties, with most users getting responses within a day. If the answer isn’t sufficient, users can opt to do a live consultation with a doctor using video or text chat from within HealthTap’s platform.
mindBot is a chatbot that uses introspective questioning and AI to help you organise your mind and concentrate on doing what matters most to you. It’s aim is to help users understand their thought patterns and how they can make sure they are not led astray by them, but use them rather to lead to positive behaviours to live an optimised life, helping them manage their intentions and be accountable for the goals they have set.
Your.MD is a personal AI healthcare assistant. You can input your symptoms and the Your.MD chatbot will suggest conditions based on your input. It will also serve up medical information that have been written and reviewed by doctors (and verified by the UK’s National Health Service). If you need further help, it has partners integrated within the app to provide specialist services and information.
In April 2016 Boston Children’s Hospital announced a new “skill” for Alexa powered devices which provides health advice on children’s symptoms. Parents can list the noticeable symptoms and the KidsMD app will quantify if the child should see a doctor. It can also provide information on dosing guidelines depending on the child’s weight or age.
Can you help me build a bot?
Yes! We’re one of only a few companies that specialise in building chat bots. We built mindBot mentioned above.
Get in touch — we’re super excited about this space, we have built our own technology platform and developed a process to bring new bot products to market quicker than ever. We’d love to share our technology and everything we’ve learnt with you.