Chatbots Can Transform The Healthcare Industry — But Patients May Need More Convincing

I’m rarely unwell, but when I am I tend to want to see a GP as soon as possible. No hanging around, riding it out or seeing what further symptoms present for me. Think frantic Googling, then straight on the phone to the doctor’s surgery. If I can’t get seen quickly, I’ve used a couple of the (now many) doctor-on-demand services available and have been impressed with the speed and service most of the time.

So the other week when I was experiencing flu-like symptoms, I called my local surgery where an administration assistant answered and told me a doctor would call me later that day. So I waited. And waited. I went to sleep. When I woke up, I found a message from a doctor on my ansaphone telling me she’d tried calling, but would call me tomorrow. The next morning, the doctor called as promised and asked me the standard triaging questions and decided that I should come into the clinic to see her as there were a couple of symptoms she wanted to check.

Unfortunately there were no appointments for two days and by that time I was over the worst, so I cancelled the appointment. I’d wasted time and energy — not only by calling, but also worrying about the symptoms the doctor had wanted to see me to discuss, which turned out to be just side effects of a heavy cold.

So the whole process was just an exercise in pointlessness.

To make it clear, I am not blaming the doctor. I understand the pressures on the NHS and it was definitely a good thing that she was being overcautious regarding my ‘unusual’ symptoms. But calling me back, the over-the-phone triaging when the doctor could be seeing patients with more pressing need, and the lack of appointments immediately available, wasted a lot of time.

Enter Fred

Imagine that instead of calling the surgery I had access to a chatbot — let’s call him Fred. Fred could have worked out after just a few questions whether I needed to collect a prescription, go to a hospital immediately, arrange an appointment with a GP or deal with the issue myself (e.g. sleep, have a lemon tea). If Fred had found my symptoms weren’t part of his range of scenarios, I could have been transferred as soon as possible to a GP for diagnosis. In such a scenario, the doctor could be left to do what he/she does best: patient care.

Fred could also have followed up to remind me to take my medication and monitored my progress until my condition was cured. As it was just a common cold this wasn’t necessary. But in another situation, Fred could have been like my own personal assistant, taking the stress out of the appointment-booking process and helping with patient follow-ups. Chatbots used in such a way ease the pressure off GPs and help to improve the patient experience.

The elephant in the room

Most people don’t need to see a doctor for the majority of medical conditions. The NHS is under huge pressure and a lot of interaction could be taken over by chatbots and instead spent on patients who need to see a doctor the most. Ironically these patients often don’t get seen quickly enough. Patients who can have their needs solved online, will no longer be delayed and will be less frustrated by the whole process.

Different types of chatbots

Chatbots are already here but haven’t yet been widely adopted in the NHS. They can take on many forms, not just as coordinators of the patient-doctor experience, but also as responsible for internal record-keeping, sending medical records on to A&E and so on.

Voice Assistants can be used in conjunction with chatbots to provide better elderly care

Here are some other ways different chatbots can be used in healthcare:

• Wellness chatbots — can be used to help people count calories, keep records of sleep patterns and exercise routines. For example, Fitbit has integrated Alexa with its Fitbit Versa 2 but customer service has been slow in the past due to the lack of real-time communication. Chatbots have the potential to streamline customer support in Fitbit and similar devices designed to track users’ health.
• Chatbots which help the elderly and disabled — caring for the elderly and immobile can be improved by using a voice personal health assistant. Chatbots with natural language processing (NLP) abilities offer a great option to older people. They can remind people to take medication and get patient queries processed quickly. Other ways these chatbots could be used include introducing memory games to help patients with long-term conditions like Alzheimer’s. And doctors could be quickly alerted in the case of an emergency. These types of chatbots can also help people suffering from different disabilities.
• AI applications alongside chatbots — AI can be used in conjunction with chatbots to improve elderly patient care. The University of Missouri has developed sensors that can identify problems such as determining how dangerous a fall is, detecting when oven doors have been left open and whether potential mental health issues are getting worse. All of these applications working together mean that vulnerable members of society can live more independently.

The problem with chatbots

There are a number of problems that have emerged when it comes to chatbots, especially those used in healthcare.

• Data protection — Google’s recent acquisition of Fitbit has led to users threatening to ditch their watches despite the fact that Google has said it won’t sell the data or use it for advertising, showing that user trust is at an all-time low when it comes to data privacy. Google’s also been put under the microscope after its secretive deal with Ascension and Facebook has also had questions raised after it rolled out Preventative Health, despite the fact that the social media site has pledged to introduce strict privacy safety guards. In order to put consumer fears at risk, businesses will need to communicate that they’ve implemented strict security to ensure their platform or app is resistant to cyber-attacks.
• Effectiveness — Chatbots must be effective, they have to work and they have to be straightforward to use. If a user is promised technology that can help streamline the diagnosis procedure, the chatbots need to do just that. Users don’t want to be disappointed with technology that doesn’t work or have promises made that can’t be kept . It’s important to remind patients that chatbots can help to a certain extent, but they aren’t miracle makers.
• Trustworthiness — Trust isn’t just about data protection or effectiveness, it’s also about developing relationships with consumers or businesses. If you’re a healthcare practitioner wanting to introduce an application to a patient, you need to be confident that information provided is clear, transparent and trustworthy. One US study to test whether consumers trust chatbots with their health found that they were concerned about relying on unfamiliar brands as well as HIPAA compliance. For businesses looking market their products, case studies, a review system and an accessible privacy policy as well as the right core messaging can help.

Chatbots can help global healthcare providers increase service efficiency, remove some of the burden from HCPs and improve the patient experience. However, there’s also room for things to go very wrong. Healthcare isn’t like other industries; serious harm can be caused if incorrect advice or guidance is given. This is why it’s so important to ensure that the user journey is smooth and that patients are updated with accurate information at every stage.