Every PharmSee "GP-to-pharmacy ratio" number we publish — from Liverpool 1.42:1 to Horncastle 1.67:1 — shares the same structural blind spot: we don't yet count dispensing doctors. In rural England, where ~3.5 million patients collect their prescriptions directly from their GP practice rather than from a community pharmacy, this omission meaningfully distorts the competitive map. This article explains what the distortion is, where it matters most, and why fixing it is the single highest-value PharmSee data integration on our backlog.
What a dispensing doctor is
Under NHS England contractual rules, GP practices in rural areas (broadly: more than 1.6 km from the nearest pharmacy, with some legacy grandfathering) can apply for a dispensing licence. When granted, the practice dispenses prescriptions directly from its own dispensary, bypassing the community pharmacy entirely. The GP practice collects the same £1.29-per-item dispensing fee a community pharmacy would collect, plus a small additional dispensing service payment.
The Dispensing Doctors' Association estimates there are approximately 950 dispensing doctor practices in England, serving ~3.5 million patients and dispensing roughly 55 million items annually. That is not a rounding error — it is ~3% of the total 1.1 billion English NHS prescription volume, concentrated entirely in the rural quartile of the country.
Why this breaks PharmSee's rural ratio analysis
When we query PharmSee's location analyzer for a rural postcode like Padstow PL28 8AE at 3 miles, the response is:
- GP practices nearby: 0
- Pharmacies nearby: 1
We report that as "0 GP / 1 pharmacy / undefined ratio" and move on. But the real picture for Padstow is: the nearest GP practice is Wadebridge (4.5 miles away), and Wadebridge GP may well have a dispensing licence. If it does, Wadebridge GP is directly competing with Padstow Boots for the dispensing revenue of the ~4,500 Padstow year-round residents plus summer visitors. The "1 pharmacy" in our Padstow catchment is not the only dispensing actor serving that population — it is one of two, and we can only see one of them.
Apply the same logic to:
- Mid-Wales (Brecon, Aberystwyth): Welsh dispensing doctor register is separate from the English NHSBSA register and we don't track either yet.
- Lincolnshire Fens (Boston, Spalding): the Fens have one of the highest English concentrations of dispensing GP practices — our "0.60:1 corridor ratio" is significantly misleading if a large share of the actual dispensing is happening at GP practices we're not counting.
- Northumberland, Cumbria, rural Devon, rural Cornwall: all heavily populated with dispensing doctors. Our rural ratio numbers in these areas are directional at best.
A worked example: Padstow with dispensing doctors integrated
Hypothetical scenario: imagine PharmSee integrates the NHSBSA dispensing doctor register and we find Wadebridge GP has a dispensing licence. The Padstow picture would then become:
| Catchment | Pre-integration | Post-integration |
|---|---|---|
| Padstow PL28 (3mi) | 0 GP / 1 pharm | 0 GP / 1 pharm (unchanged at 3mi) |
| Padstow PL28 (5mi) | 2 GP / 3 pharm | 2 GP / 3 pharm + 1 dispensing site |
| Padstow PL28 (10mi) | 5 GP / 6 pharm | 5 GP / 6 pharm + ?? dispensing sites |
| Catchment dispensing volume | 51,665 items (Padstow Boots only) | 51,665 + Wadebridge GP items |
The structure of the catchment changes once dispensing GP practices become visible. Market share narratives shift. Pharmacy First revenue ceilings shift — dispensing doctors don't participate in Pharmacy First at all, which means a rural pharmacy that is the only Pharmacy First provider in a 10-mile corridor actually has a larger referral monopoly than we currently report.
The seven rural archetypes most affected
Based on Dispensing Doctors' Association geographic data and our own rural coverage analysis, the seven English rural archetypes where the integration matters most are:
| Archetype | Approx. share of practices dispensing | Impact on PharmSee ratio accuracy |
|---|---|---|
| Lincolnshire Fens | 30-40% | Severe — current 0.60:1 corridor ratio likely 30%+ off |
| Northumberland / rural County Durham | 25-35% | Severe |
| Rural Cumbria | 25-35% | Severe |
| North Devon / North Cornwall | 20-30% | Significant |
| Rural Shropshire / Herefordshire | 20-30% | Significant |
| Norfolk Broads / Breckland | 15-25% | Moderate |
| Yorkshire Dales / Moors | 15-25% | Moderate |
In the two severe-impact archetypes (Fens, Northumberland/Cumbria), roughly 1 in 3 prescription dispensing events is happening at a GP practice rather than a community pharmacy. Our current ratio figures for these regions miss that entirely.
Why we haven't integrated yet
Three practical reasons:
- The NHSBSA dispensing contractor register is not a clean open-data drop. It is published as a quarterly PDF supplement to the main pharmacy register, without lat/lng coordinates, without the same contractor-code conventions as community pharmacies. Integrating it requires name-matching against the GP practice ODS register, which introduces its own ambiguity.
- Dispensing revenue per GP practice is not published at the same granularity as pharmacy revenue. We can know that Wadebridge GP dispenses, but we can't yet know whether they dispense 10,000 items/year or 80,000 items/year without a separate NHSBSA data request.
- The ratio framing itself becomes more complex. A "GP-to-pharmacy ratio" that includes dispensing doctors has to decide: do dispensing GPs count on the pharmacy side (as an additional dispensing actor), on the GP side (as they currently do), or on both? There is no clean answer — and different analyst audiences will want different conventions.
We have been running the integration in a staging branch for four weeks and expect to surface dispensing doctor counts on the location analyzer in cycle 16 or 17. The first deliverable will be a badge ("2 of 5 nearby GPs dispense") rather than a full revenue integration, because the dispensing volume per practice is the harder half of the problem.
The honesty disclaimer
Until that integration ships, every rural ratio PharmSee has published — including the nine rural articles written in cycles 10-14 — should be read with the caveat that the pharmacy side of the ratio is the only dispensing actor we count. For most urban catchments (Liverpool, Birmingham, Manchester, London) the error is small because urban GP practices are almost never dispensing. For rural catchments the error is substantial. We've flagged this caveat in the research logs; we haven't until now put it in the published articles.
This article is the flag. If you're using PharmSee's rural pharmacy-density numbers for location planning, treat them as a lower bound on actual dispensing competition — and add a conservative 20-30% adjustment for dispensing doctor presence until the integration is live.
Why this matters for the rural pharmacy runway
The practical implication for a new-build rural pharmacy: if you are looking at an underserved rural corridor using PharmSee's current data, check the dispensing doctor register manually before committing. A corridor that looks attractive at 2 GP / 1 pharmacy (2:1 ratio, high opportunity) might turn out to be 2 GP / 1 pharmacy + 1 dispensing GP — effectively a 1:1 ratio with the dispensing GP as a hidden competitor. The difference matters for your £300,000 new-build capex decision.
We are working to close this gap. In the meantime, PharmSee's rural numbers are directionally correct but systematically under-counting the competition. The cycle 16 integration will be the single biggest accuracy improvement to the rural ratio atlas since the engine launched.
Explore the tool: Location analysis · Pharmacy search · Related: Lincolnshire Fens