The Framework wants proactive care for thousands of at-risk patients but barely mentions the tools to deliver it.

The NHS has spent the better part of a decade talking about integrated neighbourhood teams. But it seems that we may finally be moving beyond what is affectionately referred to in the NHS as the ‘Kumbaya’ moment, where meetings and good intentions give way to a concrete articulation of the job they should accomplish.

Though still loaded with reform agendas, goals and objectives, the Neighbourhood Health Framework does take a step forward by focusing minds on attenuating non-elective demand (admissions and bed-days), driven by at-risk patient cohorts including people with frailty and multiple chronic diseases. Sir Jim Mackey’s recent letter to NHS CEOs is more straightforward, asking “neighbourhoods [to play] a central role in implementing proactive care models for highrisk groups” as one of eight areas where he wants leaders to focus. This is smart: as research by the Health Foundation has shown, about 5% of patients drive over 50% of acute costs alone, never mind the wider system. Of these acute costs, about 60% are the result of non-elective admissions and bed-days. This means hundreds of thousands of bed-days and hundreds of millions of pounds for a typical hospital.

But if you read the Framework, there are but two mentions of technology, both about productivity tools in GP practices. This matters because, without technology, the ambition of neighbourhood health cannot be realised at a scale that would make enough of a difference. A Framework that implies proactively managing many thousands of patients, while barely mentioning the tools that would make this possible, risks setting neighbourhood teams up to fail. You can’t deliver at-scale proactive care with a spreadsheet.

Predicting who needs enhanced care

The Neighbourhood Health Framework rightly centres its attention on people with multiple long-term conditions and frailty. These patients often experience many unscheduled hospital care episodes in a year, bouncing between services in a way that is both costly and fragmented. We should organise care models around multimorbidity and accept that these patients require enhanced care designed to manage exacerbations early.

But predicting who is at risk of deterioration and unscheduled care requires relatively sophisticated modelling and data science. Machine learning is well suited to address this problem. Dynamic risk stratification — drawing on primary care records, secondary care data, and data collected remotely from patients themselves — can surface those most likely to benefit from intervention. The technology exists. But to drive action, it must be embedded in automated workflows that ingest and structure data from multiple sources.

Engaging patients proactively is a specialist skill

Neighbourhood models imply engaging with patients differently, proactively identifying people who have not yet sought help, reaching out across multiple channels, and building trust with populations who are often wary of the NHS. Waiting for patients to self-refer, or picking them up opportunistically in the clinic, will not reach the people who need care most.

The Framework implies neighbourhood teams must do this kind of outreach systematically, for large numbers of at-risk patients. Technology is indispensable for reaching this number of people. For example, at Doccla, our enrollment specialists use advanced voice AI technologies to maximise the number of patients they can contact and engage. We A/B test everything — call scripts, SMS message content and timing, outreach sequencing — to understand what actually works for different patient groups. The result is that each of our enrollment specialists onboards around 350 new patients per month, with those in the most deprived IMD joining at twice the rate of the rest. Although the human touch is certainly needed, it too must be augmented to achieve the scale necessary to meaningfully impact demand.

Scale is the central challenge

If neighbourhood health is to make a meaningful dent in acute demand, it needs to manage patients in the thousands, not the hundreds. A team managing three hundred patients can produce compelling case studies. A team managing ten thousand can move population-level outcomes and create headroom in hospitals.

Managing caseloads at that scale requires technology. AI-powered tools for prioritisation, care planning and clinical decision support are not a luxury add-on to neighbourhood health — they are a prerequisite for it working. Clinicians managing large caseloads need systems that surface patients before they deteriorate, flag gaps to guideline-directed care and enable frictionless teamworking.

We have invested in these capabilities for our clinical teams. Each of our doctors and nurses manage at least 120 patients; our pharmacists, focused primarily on medicines reviews and optimisation, manage around 600 patients each. This is what’s possible when technology is designed to extend clinical capacity rather than simply document it.

Creating the conditions for investment

Rob Webster and Sir John Oldham recently argued in the HSJ that “acute hospital measures of success … cannot dominate the implementation of neighbourhood health.” That’s right, but we should also understand that neighbourhoods need to create the conditions for their own investment at a time when the NHS budget is barely growing and will likely not do so for many years to come. Historically, growth money has gravitated towards hospitals. To reverse this pattern, neighbourhoods should demonstrate they can prevent unnecessary hospital activity at a scale that’s worth paying for.

Sir David Sloman recently said to me “hospitals should be hard to get into and easy to leave”. Making this a reality means treating technology as central, not an afterthought.

‍