The lag table
Median days from EMA authorisation to each agency's first published decision on the same molecule. Computed from the joined records, not surveyed. Four rows are open; subscribers see all 17 systems, every molecule behind each median, and the timelines for the US, Japan, the UK, Switzerland, Canada and Australia.
| Agency | Median lag | Molecules | Decisions from |
|---|---|---|---|
| HAS · France | 198 days | 373 | 2003 |
| G-BA · Germany | 250 days | 324 | 2011 |
| TLV · Sweden | subscriber data | 424 | 2003 |
| SMC · Scotland | subscriber data | 565 | 2002 |
| NICE · England | 398 days | 338 | 2004 |
| CIPM · Spain | subscriber data | 163 | 2019 |
| Medicinrådet · Denmark | subscriber data | 152 | 2017 |
| AIFA · Italy | subscriber data | 134 | 2017 |
| AOTMiT · Poland | 629 days | 187 | 2016 |
| Nye Metoder · Norway | subscriber data | 212 | 2014 |
Each body is restricted to molecules authorised on or after its own first recorded decision, shown in the last column, so a short decision archive does not read as a slow country. The figure is the interval to that body’s published decision or opinion — not to price agreement, national listing or first patient use. Intervals outside −400 to 4,000 days are excluded as join failures or decisions too late to attribute; that removes 77 observations in total. Since 2021 an EU authorisation no longer starts the clock in Great Britain, so the England and Scotland rows are anchored to a date that is increasingly not the operative one: 131 of the English and 125 of the Scottish observations are post-2021 EU authorisations.
Ten bodies, ten different questions
The table above is not a league table of national speed, and reading it as one will mislead you. Each row is the interval to a different kind of decision. France’s HTA opinion (Commission de la Transparence) is a clinical view that arrives before any price is agreed. Norway’s is a decision that already carries the price. Comparing 198 days against 670 days is comparing an opinion with a conclusion.
Germany is the row most often misread. A new medicine there is reimbursed by statutory health insurance from the day it launches, with the manufacturer free to set the price for the first six months. The G‑BA resolution that lands a median 250 days after EU authorisation gates the price, not the patient. Nobody in Germany waits 8.2 months for access.
| Agency | What its published decision is |
|---|---|
| Medicinrådet · Denmark | Medicinrådet recommendation |
| NICE · England | technology appraisal guidance |
| HAS · France | HTA opinion (Commission de la Transparence) |
| G-BA · Germany | benefit assessment resolution |
| AIFA · Italy | Gazzetta determination, innovativeness assessments |
| Nye Metoder · Norway | Beslutningsforum decision |
| AOTMiT · Poland | AOTMiT recommendation |
| SMC · Scotland | HTA advice |
| CIPM · Spain | interministerial price and financing decision |
| TLV · Sweden | reimbursement decision, outpatient scheme only |
Why the archive window is in the table
A decision archive that opens late makes a country look slow. If we hold records for Poland only from 2016, then a molecule authorised years earlier contributes an interval that measures our coverage rather than that country’s process. Measured without that restriction Poland reads 875 days; restricted to molecules authorised once the archive was open, it reads 629. That is a 246-day difference produced entirely by record-keeping, and it moved Poland from 10th to 9th in the table.
Every row here is restricted that way, and the last column shows the year each archive opens so you can judge it yourself. Mature archives barely move; the correction matters most exactly where the headline numbers were worst.
What this does not measure
It does not measure when a patient can get the medicine. Between a published decision and a dispensed pack sit price negotiation, national or regional listing, tendering and local formulary adoption, and those add months in most of these systems. If time‑to‑availability is the question, EFPIA’s W.A.I.T. Indicator measures the interval to appearance on the national reimbursement list and is the right source; it is not the same quantity as this one and the two should not be compared row for row.
Two further limits worth stating. Wales is not in the table. AWMSG would be the fastest and largest series here at 154 days over 596 molecules, but since 2010 its remit has been medicines not on NICE’s work programme — a different population, not a faster country. And since 2021 an EU authorisation no longer starts the clock in Great Britain, so the England and Scotland rows are anchored to a date that increasingly is not the operative one — 131 of the English observations are post‑2021 authorisations. We do not publish an MHRA‑anchored alternative: the GB authorisation dates we hold include later licence events, so they do not reliably mark first authorisation, and a median built on them moves by more than a hundred days depending on how the outliers are treated.
Questions this data answers
How long after EU approval does NICE publish its decision?
A median of 398 days across 338 molecules authorised from 2004. Treat the recent end of that range with care: since 2021 an EU authorisation no longer binds Great Britain, so for the 131 post-2021 observations the EU date is not the operative start of the domestic process.
Which European HTA body decides fastest?
On this measure France, at 198 days over 373 molecules. But HAS’s decision is HTA opinion (Commission de la Transparence), which is an earlier step than several of the others, so the ordering reflects what each body publishes as much as how quickly it works.
Why is your figure for a country different from published estimates?
Almost always because of the endpoint. Published comparisons usually measure time to reimbursement or to patient availability; this measures time to the HTA body’s own published decision, which comes earlier. Where a source states its endpoint we can reconcile the two; where it does not, the figures should not be placed side by side.
How many molecules sit behind each figure?
Between 134 and 565, shown per row. Coverage differs by body because each archive opens at a different date and each body assesses a different share of EU authorisations, so no two rows share a denominator.
How current are the approval and decision dates?
Data to 2026-09-21. The HTA archive refreshes twice a month; regulatory approval dates refresh as each agency publishes.
Citing this
Altima Partners, “Approval to HTA decision: median lag by body”, altimapartners.co.uk/regulatory-timelines, data to 2026-09-21. Figures are medians of days from EU marketing authorisation to each body’s first published decision, restricted to molecules authorised on or after that body’s earliest recorded decision. If you are quoting a row, quote its n and its archive year with it.
Two molecules, end to end
The same two worked examples as our HTA tool. Every dot is a dated primary record.
The full layer is one subscription away
Every molecule, every authority, every lag, in the portal and over the API.
See pricing