The shortcuts we weren’t supposed to take: Desire lines, junctions and health infrastructure
There’s a concept in urban design called desire lines.
You’ve seen them - even if you didn’t know their name. They’re the faint, unofficial paths worn across parks and green spaces. The shortcuts that tell you, more honestly than any formal signpost, where people actually want (or need) to go.
They’re not (usually) vandalism or rebellion. They’re instinct.
They are human behaviour made visible.
People don’t walk that way to break rules - they walk that way because it makes sense. These lines are a quiet critique of design. They are the mark that says, ‘you almost got it right, but not quite.’
Desire lines tell us something powerful - people don’t move through places according to the lines on the plan. They move in the way that makes sense to them - usually according to logic, instinct and need. They are something to be observed and learnt from.
And lately, I've been thinking about what desire lines look like in our health system.
Because we keep seeing the same pattern - empty health buildings, but full cafés down the road. For those of us planning NHS estates, neighbourhood hubs, or integrated care, this should feel uncomfortably familiar.
The desire lines of health
These aren't the patient journeys we map on paper. Desire lines show up in the real, lived routes connecting GP surgeries, hospitals, homes, and communities - tracing links between housing, transport, social care, and support networks.
Prevention begins in the daily interactions and physical movement through communities that keep people well long before anyone becomes a patient. Every NHS site sits within a wider web of movement, yet we plan our buildings like they exist in isolation. We prioritise health spaces for delivery - because of financial pressures or perceived suitability - or worse, assume our NHS buildings automatically sit at the heart of the community and try and cram them with services better placed elsewhere.
At the same time, the library is packed. The lunch club is buzzing. The shortcut to the bus stop is worn bare. Trust and connection already exist in communities - just not where we built it for them.
If we want a prevention-led system, we need to understand how people already move through their neighbourhoods, and whether the health system blocks that movement or supports it.
Integration is spatial
We talk endlessly about integration in health policy - seamless care, joined-up systems, wraparound support. But integration isn’t just a policy, organisational or operational concept.
Integration is spatial.
It exists in how people move through buildings and across neighbourhoods, how facilities relate to one another, how local places connect.
The NHS estate (all 25 million square metres of it) isn’t just the backdrop to care and treatment. It’s the stage. It is the connective tissue of the health system.
Designed in isolation, it creates islands. Designed as part of a network, it enables flow.
The truth is patients, staff, families, and communities make their own routes as they navigate health, wellbeing, treatment and care. They carve invisible lines through their communities and through the NHS system, crossing boundaries, ownership, and postcodes.
And if we want health infrastructure that truly works - particularly at neighbourhood level - we need to see those routes, and help make them easier.
The corridors we don’t see
In the world of health infrastructure, we’re pretty good at drawing buildings.
We broadly understand the hospital, the clinic, or the park. We can show you where a new emergency wing will sit, or where a new neighbourhood health centre will replace existing poor quality accommodation.
We’re not quite so good at drawing what happens between them - the informal connections, the corridors between buildings, the pauses, the shortcuts, the social spaces where lives intersect.
Unlike in the park, there’s no worn grass to follow. In health systems, these desire lines are all but invisible.
And they rarely appear in an estates strategy.
Our challenge - how do we see them, understand them, and design our health infrastructure around them?
Finding the junctions
Every desire line meets a junction - where personal choices and formal systems collide. In healthcare, junctions might be a pharmacy, a discharge desk, or a community space. We often like to label them as ‘hubs’. Some of these flow beautifully. Others create friction, confusion, inefficiency, or delay.
So if we want seamless, preventative, community-based care, we must design better junctions - and plan infrastructure based on how people actually live.
Not every building wants an NHS badge
Crucially, not all junctions can - or should - wear an NHS badge.
Many already exist where trust and belonging live - in spaces like cafés, libraries, gyms, places of worship, or markets.
People don't move the way we plan for them to move. The NHS’s role in neighbourhood health development isn't to claim every junction - sometimes an NHS label is actively detrimental. Instead, strategically recognise, connect, and strengthen what's already working. Our health buildings won't always be the junction, and that's okay.
You cannot build connection by declaration.
A neighbourhood health centre doesn’t become a community hub just because we put it on the sign, or said it would in the business case. We cannot naïvely copy-and-paste success stories places like Bromley-by-Bow into a different context, expect it to ‘just work’, and proclaim it is some sort of community centred, one stop shop for health and wellbeing.
A new NHS one-stop-shop that no one will ever use will never be better than two-stops that actually work.
Where desire lines meet junctions
This is where the magic happens.
Desire lines reveal where junctions want to be.
Staff shortcutting to the bus stop, communities navigating between GP practices, housing, and social care - these are the beginnings of junctions that deserve attention and investment.
Prevention flourishes when infrastructure supports these existing patterns of life- not when we force new patterns through NHS programmes or new buildings.
We can’t declare a building to be a junction and expect connection to appear. The NHS’s role is to map, connect, and strengthen the natural infrastructure intersections that people already trust and use - not assume we can create them by decree.
Mapping the desire network
We already hold fragments of the intelligence we need - in patient feedback, staff complaints, footfall data, community asset maps. We just don't treat it as spatial insight.
Imagine if we actually mapped:
Where people walk
Where they wait
Where they ask for help
Where support naturally clusters
Where they’d have a cup of tea and a chat
Where trust already exists
Where young people gather
Overlay this with service models and a desire network emerges - the real connections of prevention and care. A system shaped by human movement, not where our health buildings are.
This isn't abstract. It's strategic.
First, it would help us build better new neighbourhood health centres - ones that actually understand the ecosystem they're entering, not ones that assume they'll become the centre of it.
Second, it would let us use our existing NHS estate in better ways - actively strengthening community connections instead of destabilising them by forcing inappropriate services into the wrong places.
This mapping tells us where to invest, where to partner, and critically - where to step back.
We can't build a community-focused, prevention-led health system on guesswork about where people actually are.
What if we started with the network that already exists?
The shift we need
Infrastructure shouldn’t start with buildings. Strategic health infrastructure planning should pivot from a building-centric mindset to a lines-and-junctions mindset.
Follow the shortcuts. Find the junctions. Track the flows. Invest where people actually go.
If desire lines show us how people move, and junctions show us where they meet, then health happens where those two things touch.
Maybe the future of NHS infrastructure isn’t about drawing new lines - but noticing the ones already there.