Requirements

Getting Started

Background

The NHS is the largest non-military public organization in the world. It is evident that recently it has come under immense pressure, thus exhausting its resources. Incidentally, the NHS is also receiving a surge in medical appointments for trivial problems, which are not an effective use of clinicians’ time.

It is more important than ever to develop systems that increase the efficiency of NHS clinicians by helping them spend time where it really matters. The NHS has decided to design automated self-service medical advice chatbots.

bot

Patients can be advised to first interact with these chatbots when they have a problem, and they will receive advice for either further medical examination or where they should look for more relevant info. This takes significant workload off doctors.

The NHS use Voiceflow to design these chatbots because their interface is the most intuitive on the market. This is important as NHS clinicians - not engineers - are designing these bots. However, Voiceflow designs currently can only be used for smart speaker devices, such as Google Home or Amazon Alexa. This is not very inclusive. Nonetheless, at least they can make some chatbots.

Problem

The NHS had to make a compromise on who would have access to these chatbots (only those with smart speakers) because there aren’t any viable alternatives to design bots for other platforms. This is a key limitation as smart speakers are far from widely adopted, and some patients do not even have internet access. So what about the other tens of millions of Britons?

Ideally, the NHS would like to design SMS text bots to reach more patients. However, as Prof. Joseph Connor from the NHS puts it:

Normally, developing a text bot requires a clinician to speak to a technologist, discuss the requirements and bot design, the technologist develops that in 6 months, and tests it in an environment to evaluate its impact for a further 12 months.

In addition, patient-sided problems include long waiting times when calling NHS 111 only to receive quick, generic advice. Accessing health information on the NHS website might be more useful but requires greater effort and is therefore difficult.

Goals

SMSITPAGE2

This is where we come in. The end goal is to allow Voiceflow bot designs to be converted and deployed as SMS chatbots and, as a result, allow anyone with a GSM cell phone to access these health services.

With this project, I’d expect the timescale for developing such a text bot to reduce by a year. - Prof. Joseph Connor

Client

joe

This project is developed for Prof. Joseph Connor and the NHS.

Prof. Joseph Connor is a CTO in HLP (Health & London Partnership). He is a member of the BSI committee for validating AI and healthcare. He is writing the framework for the implementation of AI in Wales. Joseph expects to use this project as an example of user-centered design and its easier deployment using the services of IXN students.

Contact

The team has done an exceptional job at empowering NHS clinicians to convert Voiceflow formats into SMS bots. This is the first implementation of its kind in the NHS. It is important that it will enable clinicians to engage people who have not got easy access to the internet in dialogs that are important to their health and wellbeing. We look forward to testing this with non-clinical interactions in 2021. - Prof. Joseph Connor