Improving Access to Psychological Therapies (IAPT) chatbot was an East Midlands AHSN demonstrator programme – a pilot to test and evidence the use of a person-centred chatbot solution, capable of continuous improvement, to support IAPT service user engagement.

EMAHSN worked with Trent PTS who are commissioned to provide IAPT services in Derbyshire and Nottinghamshire.

EMAHSN supported the pilot in these areas – from January 2021 to November 2021.

Solutions / action taken

Our IAPT chatbot demonstrator used the software provider Virtual Health Labs’ ‘Augmentbot’ chatbot solution. Augmentbot is an automated conversational agent, underpinned by a motivational interviewing algorithm and behaviour change principles.

Initially, two chatbots were proposed DNA and Step-Up with Solution and Prepare Client introduced later in the project:

• DNA chatbot - Offered to clients who do not attend (DNA) their first assessment. Aimed to establish if there is a problem with engagement and encourages clients to re-book another appointment.​
• Step-Up chatbot - Offered to clients who are on the waiting list. The chatbot gives suggestions of things the client could do whilst they wait for a therapy session
• Solution chatbot - Offered to clients before their first appointment. Intended to get clients to think about a current problem they have and takes them through brief solution focused coaching enabling them to identify small steps they can take to improve their situation
• Prepared Client chatbot - Offered to clients before their first appointment. It aims to prepare the client for their therapy sessions asking them about values, concerns and hopes for therapy. There is the option to forward the conversation to the therapist.

Nottingham Trent University (NTU) was commissioned to deliver an evaluation of the demonstrator.  

Key lessons learnt / outcomes

• Chatbots work best when offered to a patient when they initially make contact with the service compared to after an ‘event’. The DNA bot saw low engagement from patients and the evaluation confirmed use of a bot here did not deliver impact
• Clinical staff who champion/explain chatbots saw increased patient engagement with the bots
• Clients who had used the Step up chatbot had on average fewer appointments than those who did not use the chatbot.
• Trent PTS and NTU aim to complete further research to understand why Prepare and Solution chatbots attracted more engagement than the initial two chatbots

Resources