Building a proactive INT from the ground up: Arbennek PCN

About this evaluation

An independent assessment

This case study draws on an independent evaluation conducted by Future Care Capital (FCC), a UK charity dedicated to driving sustainable impact in health and care. The evaluation was led by Dr Melanie Fraser, with support from Professor Andy Jones, and involved 13 in-depth interviews with a diverse range of stakeholders across the Arbennek PCN Integrated Neighbourhood Team — including GPs, care coordinators, social prescribers, VCSE representatives, ICA managers, and local authority staff.

The evaluation was not focused on assessing Brave AI as a technology. Instead, it examined the process of creating, operating, and leading the INT — using Tuckman's stages of group development as its theoretical framework. The resulting report provides one of the most detailed independent accounts of an NHS INT in operation.

Context

Arbennek PCN, Cornwall

Arbennek PCN covers a population of approximately 33,000 people across four GP practices in Central Cornwall — Brannel Surgery, The Clays Practice, Probus Surgery, and the Roseland Group Practice. The PCN encompasses significant socioeconomic diversity: from the wealthy retired communities along the Roseland Heritage Coast to the former clay mining villages around the Clays, where older patients are likely to have spent their working lives in a low-wage industry now largely gone.

Arbennek was selected for this evaluation because it was one of the earliest PCNs to receive access to Brave AI, as part of a pilot programme, and because early indications suggested an unusually positive and impactful implementation — making it an ideal case study for learning dissemination.

Results

49 out of 50 patients kept out of hospital

The headline result from Arbennek's first cohort was striking. Of 50 high-risk patients reviewed by the INT, 49 were kept out of hospital. The one patient who passed away died at home — which had been their wish. As one civil servant described it to the evaluation team:

Brave Scores declined across the cohort as a result of INT actions, providing measurable evidence of reduced risk alongside the admission avoidance data.

How it worked

Brave AI as a catalyst for collaboration

Before Brave AI was introduced, INT meetings at Arbennek felt unfocused. One participant described pre-Brave meetings as "nebulous" — lots of professionals in the room, discussing a concept that hadn't yet taken concrete form. The introduction of Brave Scores changed that:

The INT used monthly meetings to review patients with high Brave Scores, agree action points, and revisit them the following month. Actions ranged from a joint home visit by a GP and care navigator, to arranging a volunteer visitor for a patient experiencing loneliness. The Brave Score gave the team a shared, evidence-based starting point — while the human relationships and judgment of the team determined what happened next.

As one INT leader put it: "The reason I wanted to use Brave was I felt there was a great opportunity to kickstart — to become a catalyst for collaboration and working. Brave is just a tool. It's a way of identifying patients."

The team

A genuinely integrated neighbourhood team

What distinguished the Arbennek INT from a standard MDT was the breadth of its membership. Meetings grew from around a dozen participants to over 35, drawing in professionals from across the health and care system: GPs, social prescribers, district nurses, adult social care, Age UK, VCSE organisations, and community health staff. One attendee cascaded the views of 27 other VCSE organisations.

This breadth created some operational challenges — large meetings are time-consuming, and not every patient discussed is relevant to every attendee. But participants consistently rated the cross-sector relationships built through these meetings as one of the INT's most important outputs, enabling faster, more joined-up responses when individual patients needed multi-agency support.

Energy and hope

The human dimension

The evaluation's sentiment analysis — conducted using NVivo — found 555 references coded for positive sentiment against 211 for negative: a ratio of more than 2:1. Participants used words like "exciting," "hope," and "transformatory" repeatedly. One GP practice manager summed up the feeling:

Key lessons

What other INTs can learn

The Future Care Capital evaluation is explicit that the Arbennek model cannot be directly replicated — it emerged from a specific set of personal relationships, local conditions, and bottom-up energy that are not transferable wholesale. But the learning points are: