.png)
SummaryMeasured purely on data quality, Poland is the best free medicine price source we examined in Europe. Manufacturer price in the file, pack size as its own field filled to 100 percent, a valid GTIN on every row of the main sheet, four price levels in a single line, and a daily register alongside the quarterly price list. The obstacle is not technical. It is four characters in a licence name.
The health ministry publishes the reimbursement list as an announcement with an attached spreadsheet. Alongside it runs the register of medicinal products, which is updated daily.
| Property | Reimbursement list | Product register |
|---|---|---|
| Price level | Manufacturer price, net selling price, plus wholesale and retail levels | Not a price source |
| Format | XLSX, roughly 1 MB | XML and CSV, daily |
| Rhythm | Quarterly | Daily |
| Code | EAN code plus the national code | GTIN, pack count, ATC |
| Pack size | Own field, filled to 100 percent | Structured |
| Licence | CC BY-SA 4.0 | CC BY 4.0 |
Two corrections to widely circulated descriptions are worth stating. The rhythm is quarterly, not two monthly, evidenced by the publication dates of 18 December 2025, 17 March 2026 and 17 June 2026, each effective at the start of the following quarter. And a spreadsheet version of about 1 MB exists alongside the 97 MB PDF, which removes any need to parse a PDF.
Three properties, rarely found together:
The manufacturer price is in the file. No tax step, no margin inversion, no band basis to prove. Poland publishes the net selling price directly, and three further price levels sit in the same row.
The pack size is a dedicated field, filled everywhere. Not parsed from a designation, not inferred, not partially populated.
Every row of the main sheet carries a GTIN. All 4,242 values in the main sheet are numeric and exactly fourteen digits.
For anyone who has fought with free text pack descriptions in France or Spain, the practical difference is measured in weeks of implementation work that simply do not arise.
The hospital segment is covered through additional sheets in the same file, which means the outpatient and hospital views come from one source with one schema.
The register is published under CC BY 4.0 and is free to use commercially. The price list itself is published under CC BY-SA 4.0.
Share alike is the problem. A licence that requires derivative works to be distributed under the same terms sits awkwardly with a paid product built on that data, and the question has to be settled by a lawyer rather than by an engineer. Two sub questions decide the outcome: whether the obligation extends to a database that merely incorporates the data alongside other sources, and whether official materials are excluded from the copyright regime under national law in the first place.
Until that is answered, Poland does not belong in a commercial build, however good the file is. That is why it sits in the second wave of our own plan rather than the first: the quality argues for building immediately, the licence argues for waiting, and the licence wins because it cannot be fixed afterwards.
Poland publishes prices for the reimbursement market. Non-reimbursed medicines have free prices and are not published officially anywhere.
That boundary is sharper here than in most countries, and it is worth stating in any report: a Polish average price describes the reimbursed segment, and the unregulated segment is not merely missing from the file, it is unpublished by design.
The reimbursement list carries the manufacturer level and the levels above it in the same record. For a cross-country dataset that is unusually convenient, because it removes the derivation entirely and lets you compare Poland at whichever level a given comparison needs.
It also makes Poland useful as a control country. Where another market's manufacturer price has to be derived from a consumer price through a margin scale, Poland shows what the full chain looks like when every step is published rather than reconstructed. A derivation that produces implausible margin structures elsewhere can be sanity checked against a country where the whole ladder is visible.
The register and the price list answer different halves of the same question, and a pipeline needs both.
The register is the identity layer: daily, machine readable, with GTIN, pack count, ATC code and the procedure type. It tells you what exists and how to identify it.
The price list is the price layer: quarterly, with four levels per row. It tells you what the reimbursed products cost.
Joining them is straightforward because both carry the GTIN. That is the whole reason Poland scores as well as it does: in most European countries the identity source and the price source use different keys, and the join is where the project loses its schedule.
A quarterly price list is slower than the French monthly file and far slower than the Belgian daily export. For a price monitoring product that is a real limitation, and it should be stated rather than smoothed over: between two announcements, the Polish price data does not move.
What compensates is predictability. The publication dates are known in advance and the effective date is the start of the quarter, so a pipeline can schedule the fetch rather than poll for changes. Predictable and slow is easier to operate than fast and irregular, as long as the promise made to users matches the rhythm of the source.
The price list link has to be parsed from the page rather than constructed. The register, by contrast, is retrievable directly, and we downloaded it on 5 September 2026 at HTTP 200 and 16,996,036 bytes.
That asymmetry shapes the pipeline. The daily register can be fetched unattended. The quarterly price list needs a small amount of page parsing, four times a year, which is a modest cost given the rhythm. What it does mean is that a link change will break the price branch silently, so the fetch should verify that the file it found is newer than the one it holds.
Even with four published levels, the Polish file is a list price file. Rebates, tender outcomes and hospital contract prices sit below the published manufacturer price and are not published, exactly as in every other European market. A published manufacturer price is the best starting point available, and it is still a ceiling rather than a transaction.
For sourcing work that distinction is operationally important: the published figure tells you what the regulated maximum is, and the negotiation happens underneath it.
Cross-border sourcing. Poland is a regular origin market in parallel distribution, and a published manufacturer price per pack with a GTIN is precisely what a sourcing decision needs.
Reference pricing analysis. Countries whose formulae read Poland can be modelled directly, without a derivation layer in between.
Benchmarking a data model. If your schema cannot represent the Polish file cleanly, with four price levels, a GTIN, a pack count and two codes per row, the schema is the problem.
Poland is the clearest case in Europe of a country where the engineering is finished before it starts and the legal work decides the schedule. Manufacturer price included, pack size complete, GTIN everywhere, four price levels in one row, a daily register beside a quarterly list. And a share alike clause that has to be resolved in writing before any of it can be sold.
Related reading: European Drug Pricing Database: how cross-border prices work and Pharma pricing Germany: how the system works, plus ex-factory price and list price vs net price.
pharmazie.com is the consolidated pharmaceutical data platform by DACON Datenbank Consulting GmbH that bundles 25+ specialist databases into a single search, exclusively for healthcare professionals. Price coverage focuses on the DACH region and a number of further EU countries, with more countries following in the coming months.
