Every program tracks non-attendance, and most treat it as a fact about patients. Some of it is a fact about geography, and that part is addressable. This is how to find out how much of yours is which, without a research project.
For a patient travelling in, an appointment is not an hour, it is a journey, a night away, a room, meals, time off work, and often a second person doing the same. The appointment is the smallest part of it.
Which means the decision to attend is being made against a much larger cost than the clinic sees, and the part of that cost a program can most easily reduce is the part it currently has nothing to do with.
This is a straightforward cut and it takes an afternoon. You need three things, and two of them are already in your systems.
Group patients into distance bands rather than plotting a continuous line. Something like: local, up to an hour, one to three hours, and anything requiring an overnight stay. That last band is the one that matters, because it is where lodging enters the picture.
Then look at attendance and completion by band. If the overnight band is materially worse than the rest, you have found something specific: not that some patients are less engaged, but that a particular group faces a cost the others do not.
Most drivers of non-attendance are hard to influence: transport in general, work patterns, caring responsibilities, health literacy. Distance is not something a clinic can change either.
But the cost of an overnight stay is, and it is the piece of the geography problem that is actually in reach. That is what makes this analysis worth doing rather than interesting: it separates a barrier you can act on from a set of barriers you cannot.
If the overnight band looks worse, the useful next question is what those patients currently do about lodging, and the honest answer at most programs is that nobody knows. That is the gap.
It is also the point at which lodging stops being a nice thing to offer families and starts being connected to a number the program is already measured on. Non-attendance has a budget line and an owner. A lodging page usually has neither.
Association is not causation, and distance travels with other things: patients who come furthest are often those with fewer local options, which can mean a more complex case or fewer resources. Treat the result as a place to look rather than a proof.
And be careful about a denominator that hides the effect. Patients who never started because the travel was impossible are not in your data at all, so any measure built from patients you already treat understates the problem by exactly the group most affected by it.
The cut is different. Most reporting shows a rate across all patients, which averages away geography entirely. Splitting by distance band separates a barrier you can act on from a set of them you cannot, and it is usually the first time the overnight group is looked at on its own.
A rough straight-line distance from the postcode centroid to your site is enough for banding, and any geocoding tool will do it. Do not spend effort on drive-time precision until you know whether the bands differ at all.
That is a genuinely useful result and worth knowing. It suggests your travelling patients are already well served or well resourced, and that attention belongs elsewhere. Check the denominator caution above before concluding it, though.
For a single visit, often not. For a protocol with repeat visits over months, it accumulates into one of the larger out-of-pocket costs a family faces, and it is concentrated in exactly the patients who travel furthest. That is why the overnight band is the one to look at.