Could virtual assistants reduce the strain on the NHS?

Posted:

It’s no secret that the NHS is over-stretched and under-funded. With a recent surge in Coronavirus cases sweeping the nation, the service is struggling now more than ever. Pandemic aside, the average waiting time for a GP appointment is anywhere between 14 and 20 days, while A&E waiting times regularly exceed 4 hours. What is even more troubling is that up to 17 million hospital visits a year may be entirely unnecessary, while more than 9 million A&E visits require simple advice that could be handled by 111. This in turn puts a huge demand on the advice service, with hold times amid current widespread concern reaching a worrying 24 hours. So, what can be done to address these issues?

Virtual Assistants

In recent years, there’s been a surge in demand for virtual assistants, or ‘Chatbots’. These are particularly popular with universities who have been looking to drive student engagement and reduce drop-out rates. In this scenario, powered by artificial intelligence and machine-learning, the chatbots act as personal assistants, helping students to manage their time effectively, improve their attendance, access resources and contact key members of staff when needed.

Tied into a data platform, these chatbots go beyond the scope of simply answering a few set questions. They can analyse patterns in student language and behaviour, picking up on early signs that a student may be struggling or depressed. The university can then intervene and offer the support required at an early stage, potentially preventing a drop-out.

How could chatbots be used in healthcare?

Use case #1 – responding to Covid-19 FAQs from the public

Last week, the government launched a chatbot via WhatsApp to address frequently asked questions from the public. This is an example of a very simple chatbot that can be used to offer straight-forward, trustworthy advice, reducing the strain on 111. With this chatbot, you have to select from a list of numbered topics such as ‘Reply with number 3 for symptoms’ or ‘Reply with number 5 for travel information’ so at this stage, the bot cannot engage in conversation. However, by utilising AI and machine-learning, the bot’s capabilities could be greatly expanded. The first step would be to incorporate more conversational qualities, such as responding to the initial ‘Hi’ sent by a user with a human-like response as oppose to a list of information. The ability to then analyse questions and draw from a reliable source of data to respond with complete and accurate answers would push this chatbot into the industry 4.0 space.

Use case #2 – Reducing A&E waiting times

Beyond this use case, the potential for chatbots in healthcare is huge. As well as responding to FAQs, chatbots could act as a tool to triaging patients entering A&E. This would take the strain off triage nurses, streamlining the process and allowing them to focus on more uncertain, complex or urgent cases. The chatbot could be quickly downloaded via a QR-code upon arrival to A&E. It could then assign the patient a number and begin to assess them, asking relevant follow-up questions and providing an initial assessment. The aim here would be to filter the number of people looking for basic advice regarding minor injuries, in turn reducing A&E waiting times.

A chatbot like this could also provide other useful information to improve the patient waiting experience by estimating when they are likely to be seen and where they can get food and drink while they wait. The chatbot would then provide real-time updates to notify the patient when the clinician was ready to see them. Upon approaching discharge, the chatbot could offer to pre-book a taxi for the patient, allowing them to head straight off, reducing the strain on staff and maximising efficiency.

Use case #3 – Providing extra health and wellbeing support to patients

Another viable use case could be to extend this application to inpatients. The chatbot could act as an assistant, providing additional support to inpatients such as directions around the hospital, accessing resources and answering any general queries. It could check up on a patient’s wellbeing, provide guided mindfulness and even help to tackle any underlying anxiety that a patient may be experiencing in a new and unfamiliar environment.

The past few years have seen significant improvements to the AI underpinning the bots and with the right architecture, we believe they could prove incredibly useful to both the advice service and the wider NHS as a source of reliable information in periods of unprecedented strain.

Download our free open-source chatbot

To help alleviate the pressures on the NHS’ 111 service, we have made our open-source virtual assistant, Basebot free to download, allowing for the creation of a simple FAQs chatbot in just half an hour. This can quickly answer questions from the general public. To download Basebot on GitHub, click here. Our chief engineer, Nathan Gaskill has also made a walk-through demo to help you deploy the chatbot in under 30 minutes. Check it out below:

Want to know more about BaseBot and it is deployed? Check out our chatbot hub.