Aims LT clinics are joint educational clinics run by Paediatric and GP trainees, held within GP Training Practices (see also http://www.learningtogether.org.uk). Their aim is to improve outcomes for children and young people (CYP) via experiential joint learning. The aim of the evaluation was to test the case for national implementation by interrogating who learnt what and how; and to consider how effective the intervention would need to be (in terms of improved child health) to be considered cost effective.

Methods The evaluation looked at quantitative and qualitive data. Data included: 48 semi structured interviews; 65 ‘Learning Logs’ returned in 2014–15 from 75 clinics (81%); programme workshops that featured case discussions; a Quality Improvement Project (QiP) by a GP ST3.

Two researchers completed thematic analysis to map learning. Realistic methodology established how learning may occur.

Cost analysis was used to compare the data from 2014–15 joint clinics and what they displaced in usual training (for both types of trainee) for 353 patients seen in the joint clinics. Using the evidence in NICE guidelines, the results were used to formulate a rubric for the health gain required in a child for LT to be considered cost effective locally. The rubric was applied to the QiP before and after implementation of the joint clinics.

Results Several learning themes were identified: learning clinical practice; learning how to collaborate effectively; learning about the service or system; teaching/project skills.

Both GP and Paediatric registrars learnt in all major themes and changed practice as a result over 4–6 clinics. However, they learnt in different ways and over different time frames.

Each LT clinic costs the system £37, less than the cost of one GP appointment. The cost of 6 clinics would be recouped if, as a result, one child became symptom free for 2 additional weeks.

Conclusion The evaluation results reveal a complex, interdependent learning platform that is sufficient to change practice and improve health outcomes for CYP.

We conclude that this model is a ‘no brainer’ in terms of resource use: this model provides a way of learning to deliver better care.