AI based Conversational Healthcare Service for improved care and better patient outcomes (Guest blog by Tata Consultancy Services)
Background:
One of the key challenges NHS service continues to face is high wait times for patients, more so in the cases of specialist secondary referral areas like Neurology and Cancer. Through the pandemic this challenge has been exacerbated multi-fold with waiting time for specialised care an all-time high across UK.
TCS, through engagement with our global healthcare customers in areas like clinical trials, pharmacovigilance, research and omnichannel patient engagement to name a few, has seen that, with the advancement of AI and large language models, specialist clinical and healthcare professionals can be assisted by reimagining traditional ways of working to reduce various manual touchpoints and information gathering that take up valuable consultant’s time in the delivery of patient care.
Specific NHS Trust Context:
Late 2020, at the peak of the pandemic, Senior Clinicians from a neuroscience NHS Trust in the North of England, approached TCS to explore how to use technology to help serve patients better.
During our discussions with the Trust, we demonstrated how an intelligent conversational agent can accelerate the process of gathering patient information prior to their first or follow up consultation for a patient with epilepsy. This sparked a great interest from the clinicians.
Understanding the Challenge and Shaping up Approach:
We understood from the Trust that specialist clinicians were spending 40% or more of their time gathering and validating basic information such as a patient’s condition and medication details, leading to an inefficient use of time and delays in patient treatment. Referrals for headaches generate the highest volume of outpatient referrals, with patients waiting on average around four months for their first outpatient appointment with a consultant.
Further workshops followed during 2021 including the Trust’s specialist clinicians and TCS healthcare team consisting of senior consultants from the public sector, Cloud, AI and Innovation practices. They jointly discussed and defined the high-level approach, objectives and outcomes in addressing this challenge through an intelligent digital solution that leverages conversational AI.
This led to finalising of a Minimum Viable Product (MVP) phase of work leading to a pilot to be evaluated through specialist clinicians, students and selected patients of the Trust. Thus, the NHS Trust and TCS partnered to create a pioneering, AI-based chatbot for headache patients starting end of 2021.
This digital service is envisioned through chatbot technology which can ask patients questions and provide a detailed summary report to clinicians about the patient’s symptoms and prior medical history. It can also provide useful information to the patient to help them prepare and get the best out of their first appointment.
How We Did It:
While in today’s world conversational agents or chatbots are omnipresent, to apply the same in a complex neurological area presented several challenges in terms of capturing and summarising information on various types, nuances and pathways of headaches, and trade-offs between guided and unstructured conversations as means to capture useful information prior to first and subsequent consultation.
Hence, we followed a very collaborative, design-led approach applying enough structure while not restricting point of views and scope of innovation. Some of the key elements of this journey is illustrated in the points below:
- Discovery phase – Agree the conceptual outline, plan, milestones and approach, use of collaboration tools like miro (https://miro.com/), to conceptualise target outcomes, design thinking approach (https://www.tcs.com/who-we-are/newsroom/tcs-in-the-news/design-thinking-unlocks-enterprise-creativity)
- Collaboration methods: Online tools and design-led thinking techniques using miro, design of the conversation flow for all headache types.
- Graphic design: Logos, illustrations, report, Chat Icon and UX wireframe using Adobe XD (https://www.adobe.com/products/xd/learn/get-started/what-is-adobe-xd-used-for.html).
- Development – Agile sprints. The complexity necessitated a structured and guided conversation design rather than free flow and NLP/NLG based approach. This was guided by business need, and not technology choices.
- Embedding the service into the current Trust workflow: A joined-up approach with the operational teams, clinicians and IT to outline the Workflow for embedding the new service for the Trust and patients.
- AWS partnership: For architectural guidance and hosting (https://aws.amazon.com/partners/tataconsultancyservices/)
Specifically, the following three key elements can be highlighted for success in the MVP phase and beyond:
- Business Stakeholder Alignment – Doctors and Innovation champions from the Trust engaged with TCS Public Sector Healthcare domain experts leading to quick initial agreement and alignment within the first few meetings
- Collaboration was key both from TCS and Trust side – Internal teams / departments from both sides were aligned including IT, Business, Commercial and Legal.
- Contextual knowledge and TCS experience – TCS domain experts’ deep knowledge of UK healthcare and TCS breadth of experience in chatbot technology across the globe.
Concluding Remarks:
This new online Headache Chatbot/service, the first of its kind within the NHS, has been developed in partnership with TCS, as a prototype to improve the GP referrals process and reduce waiting times for patients who have been referred to see a neurologist. The chatbot is currently in the pilot stage, and is being trialled with clinicians, medical students, and some patients at the Trust before being rolled out to patients later this year.
By using the innovative, artificial intelligence-based chatbot, patients being referred to a neurology or headache specialist at the NHS Trust will be provided with a review of their medical history of headaches sooner than if they were to wait for a face-to-face appointment through the existing referrals pathway. This is expected to reduce the consultation time by 40% thereby reducing patient wait times and improving patient care and outcomes - through early digital connect and subsequent intervention.
The design and build of the infrastructure and workflows leverage Amazon Web Services based, cloud-native, serverless technologies and conversational AI services ensuring the chatbot is safe, effective, and compliant with the latest international digital technology healthcare standards.
Going forward:
As this program was carried out as part of Innovation initiative of the specific NHS Trust, post go-live of this service for headache, plans are there to:
- Further refine and extend the capabilities of this online service e.g. inclusive design for ease of adoption, diagnostic recommendations for consideration of clinicians, redline high risk patient and symptoms for urgent care, conversational AI to include empathy and follow up questions etc.
- Expand into other areas like epilepsy
- Co-develop and offer the services to other NHS Trusts through NHS app
- Similar services through NHS app can be extended for other areas like routine pain management
Get our tech and innovation insights straight to your inbox
Sign-up to get the latest updates and opportunities from our Technology and Innovation and AI programmes.
Authors
Anjan Ruj
