The FDA can now see trial endpoints in real time. NICE recommended Imfinzi 17 days after MHRA approval. The evidence chain is compressing from both ends.

But underneath that acceleration, something else is happening. AI is no longer just supporting drug development. It is generating the evidence that regulators evaluate — patient selection, endpoint analysis, manufacturing quality, submission preparation.

The closer AI moves to a regulatory decision, the higher the credibility bar. The FDA’s 2025 draft guidance made this explicit: what matters is not whether AI is being used, but what decision the AI output supports. The EMA’s reflection paper says the same thing from the European side.

This 25-minute class covers where AI sits in the medicines lifecycle, why context of use determines regulatory significance, and what every regulatory affairs professional needs to understand before their next dossier lands on an assessor’s desk.

The question is no longer whether AI is being used. It is whether the evidence it produces can survive regulatory scrutiny.

Every global pharmaceutical company has a launch sequence. Almost none of them are designed for the world that exists in May 2026.

The traditional model — the United States first at premium pricing, then the EU Big Five, then Japan, then the rest — assumed three things that were true for two decades and are now simultaneously false. First, that European prices were a downstream consequence of HTA negotiation, not an upstream input into US pricing. Second, that each national market could be sequenced independently. Third, that the primary optimisation variable was speed to revenue.

In 2026, the launch sequencing problem has become a simultaneous optimisation across five variables: US price protection, international price visibility, speed to first patient access, clinical evidence reusability, and corridor erosion risk. Three structural shocks explain the shift — and they arrived within twelve months of each other.

This is not an incremental change. It is a regime change. And the sequencing logic that worked in 2024 will destroy value in 2026.

The three shocks that broke the old sequence

The first shock is the Most Favoured Nation framework. The executive order of May 2025, followed by the CMS models published in December 2025, created a direct mathematical link between what a manufacturer charges in selected international markets and what Medicare pays in the United States. Every European launch now carries a calculable US price consequence. The rational response — confirmed by a 35% decline in new EU pharmaceutical launches in the ten months following the executive order — is to delay, withdraw, or restructure European market entry.

The second shock is the EU HTA Regulation. Joint Clinical Assessments began in January 2025 for oncology medicines and ATMPs. Thirteen JCAs were initiated in the first year — ten oncology, three ATMP — and none had reached endorsement by December 31, 2025. The 2026 work programme projects approximately 35 oncology JCAs, 15 ATMP, 3 variations, and 5 device assessments. The JCA does not replace national pricing negotiation, but it restructures the evidence architecture — and it imposes a dossier submission timeline that most companies have not yet internalised.

The third shock is the UK–US pharmaceutical arrangement. Published on April 2, 2026, the deal raised the NICE cost-effectiveness threshold from £20,000–£30,000 to £25,000–£35,000 per QALY, committed the UK government to a 25% net price increase for prospective new medicines, capped VPAS repayment at 15%, launched the aligned MHRA–NICE pathway from April 1, 2026, and enrolled 27 companies as early adopters — with the first guidance expected in June 2026. The United Kingdom has moved from a market of reasonably fast access and relatively contained pricing to a market that is, by political design, less likely to be the low-price anchor in any international reference basket.

Together, these three shocks invalidate the sequencing assumptions that underpin every global launch plan currently in execution.

The data that should be on every sequencing slide

Before any framework can be applied, the baseline must be understood — and the baseline is worse than most headquarters presentations acknowledge.

The EFPIA W.A.I.T. 2024 Indicator — published in April 2025, covering 173 innovative medicines with centralised marketing authorisation between 2020 and 2023 — provides the most comprehensive public dataset on European access timelines. Total average availability across the EU stood at just 46% as of January 5, 2025. Full availability — unrestricted access — was 29%. Another 17% was available only with restrictions. The mean time from marketing authorisation to availability rose to 578 days, up from 531 days in the prior study. The disparity between the EU country with the highest and lowest availability exceeded 80 percentage points.

The country-level medians define the sequencing landscape. Germany: 52 days. England: 310 days. Italy: 391 days. France: 523 days. Spain: 567 days. In total availability for the 2020–2023 cohort, Germany reached 90%, Italy 83%, Spain 71%, and France 59%.

The deterioration is not uniform. In the 2022–2024 comparison, Germany worsened by just 2 days. Estonia worsened by 125 days. Portugal by 130. Romania by 219. France lost 4 percentage points of availability. Denmark lost 6. The European access problem is not static. It is diverging.

And it is not only a negotiation problem. EFPIA’s Access Hurdles Portal — covering 100% of EMA-authorised products from January 2021 to June 2024 marketed by EFPIA members — confirms that delay begins at filing, not at reimbursement decision. Mean time from marketing authorisation to filing: 90 days in the EU4 plus UK, versus 167 days in Central and Eastern Europe. Mean time from filing to reimbursement decision: 292 days in EU4 plus UK, versus 356 days in CEE.

The European access problem is simultaneously commercial and systemic — and the MFN framework has turned it from a revenue timing issue into a US price formation issue.

The MFN basket — and the confusion that is already circulating

The single most important clarification for any launch sequencing discussion in 2026 is this: there is no single MFN basket.

The executive order established the political mandate. CMS then designed three distinct operational models — and they use different country sets, different price references, and different calculation methods.

GENEROUS — the Medicaid model — uses a fixed basket of eight countries: the United Kingdom, France, Germany, Italy, Canada, Japan, Denmark, and Switzerland. The reference price is the second-lowest net price reported by the manufacturer, adjusted by GDP per capita at purchasing power parity.

GLOBE — the Medicare Part B model — and GUARD — the Medicare Part D model — use a broader basket of 19 economically comparable countries: Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, France, Germany, Ireland, Israel, Italy, Japan, the Netherlands, Norway, South Korea, Spain, Sweden, Switzerland, and the United Kingdom. In GLOBE, the default benchmark is the lowest country-level price, adjusted by GDP PPP. If the manufacturer submits proprietary net price data, CMS compares two methods and uses the higher of the two as the benchmark. In GUARD, the default benchmark is the lowest mean country-level price within the reference set.

The strategic inference is critical: not launching in a reference country does not eliminate MFN risk. CMS stated explicitly in the GLOBE proposed rule that the benchmark can be constructed from whatever subset of reference countries for which data is available. Under GUARD, an incomplete manufacturer submission may be accepted if justified and validated. The models are designed to operate with partial information.

The industry assumption that withdrawing from a reference country removes it from the MFN calculation is wrong. The models were built to function without full participation. That is the design, not the bug.

Country by country: the new strategic calculus

Germany — the signal market under pressure

Germany remains the most important European launch market, but for reasons that have partially shifted. The AMNOG system delivers near-immediate SHI coverage from marketing authorisation. The G-BA benefit assessment completes within six months. The manufacturer sets the price for the first six months — reduced from twelve — after which the negotiated reimbursement amount applies retrospectively from month seven.

That combination of immediate access, early price signal, and historical capacity to anchor negotiations across Europe explains why Germany has been the default first European market. The W.A.I.T. data confirms it: 52 days median to availability, 90% total availability.

But Germany is also adapting to the MFN era. The Medical Research Act, in force since October 30, 2024, introduced confidential reimbursement amounts. The GKV-Spitzenverband and manufacturers can now agree on reimbursement amounts that are not published — linked to local R&D investment requirements and a 9% additional discount. The strategic intent is transparent: enable deeper real discounts without automatically exporting the net price to every country that references Germany.

Germany is no longer just selling access speed. It is selling price confidentiality — and in the MFN era, that may be worth more than the access itself.

United Kingdom — the market that just changed position

The UK–US arrangement fundamentally repositions the United Kingdom in every launch sequence. The NICE threshold increase from £20,000–£30,000 to £25,000–£35,000 per QALY, the adoption of EQ-5D-5L, the 25% net price increase commitment, and the VPAS repayment cap at 15% collectively transform the UK from a market of moderate pricing into a market that is, by treaty design, protected against becoming a low-price anchor.

NICE itself estimated that the new threshold could enable recommendation of 3 to 5 additional medicines or indications per year. The aligned MHRA–NICE pathway — operational since April 1, 2026, with 27 early adopter companies — promises to advance access by three to six months relative to the current sequential process. The first guidance under the pathway is expected in June 2026.

The implication for launch sequencing is direct. The United Kingdom is a reference country in every MFN basket — GENEROUS, GLOBE, and GUARD. If UK prices rise — and the trade deal was explicitly designed to ensure they do — the floor of the MFN calculation rises with them. Every manufacturer’s incentive to launch early in the UK has increased, because the UK is structurally less likely to be the price that drags the US benchmark downward.

The United Kingdom moved from a secondary sequencing priority to a primary one in a single quarter. Any launch sequence designed before April 2026 that treats the UK as a mid-tier European market is already obsolete.

France — the structural bifurcation

France remains the European market that is hardest to position strategically. The architecture rests on the HAS Transparency Committee’s dual assessment — SMR for clinical benefit, ASMR for added value — followed by price negotiation with CEPS. The ASMR rating directly determines the pricing corridor, and 60% of new medicines have historically received ASMR V — no improvement.

The access timeline data is methodologically contested but directionally unambiguous. The W.A.I.T. 2024 median is 523 days — or 499 days if products in Accès Précoce are counted as available from day zero. The LEEM 2025 barometer reported 613 days for all of France in 2024. Other sectoral analyses cite figures around 704 days. The conclusion is not that one number is correct. It is that France is structurally slow under every methodology.

The strategic read for launch sequencing is not that France will stop being launched. It is that France will bifurcate. Therapies with high unmet medical need — rare disease, oncology, conditions with no existing treatment — can enter through Accès Précoce or Accès Direct, achieving early access outside the standard CEPS negotiation timeline. Everything else enters the standard queue. And waits.

France is in the GENEROUS basket and the GLOBE/GUARD basket. Its list price is referenced by at least twenty countries. Its net price — confidential, rebate-laden, diverging further from list every year — is referenced by none. The €7.98 billion in gross contractual rebates invoiced in 2024 creates a structural gap between the price the world sees and the price France pays.

France is not a market you sequence early or late. It is a market you sequence differently depending on whether you can access the fast lane — and whether you can afford the price signal the slow lane generates.

Italy and Spain — the filing delay nobody discusses

Italy and Spain are frequently treated as interchangeable Southern European markets in launch planning. The data says otherwise.

Italy’s median time to availability is 391 days — 176 days faster than Spain’s 567. Italy’s total availability is 83% — twelve percentage points higher than Spain’s 71%. Italy is in both the GENEROUS basket and the GLOBE/GUARD expanded basket. Spain is in the GLOBE/GUARD basket only.

But the more revealing data point is where the delay originates. In the EU4 plus UK, the mean time from marketing authorisation to filing is 90 days. The mean time from filing to reimbursement decision is 292 days. The filing delay — the time companies themselves take to submit — accounts for nearly a quarter of total access time in Western Europe. In Central and Eastern Europe, where filing delay averages 167 days, it accounts for nearly a third.

The European access problem is not only governments being slow. It is companies filing late — often deliberately, to manage corridor exposure. The MFN framework has turned that deliberate delay from a pricing tactic into a strategic liability.

The EU HTA clock — and the timeline the industry got wrong

The Joint Clinical Assessment changes the evidence architecture, not the pricing decision. The JCA evaluates clinical domains only — no cost-effectiveness, no budget impact, no reimbursement recommendation. Those remain national competence. But the JCA report must receive “due consideration” in national processes, and national authorities must annex both the dossier and the JCA report in their own HTA documentation.

The timeline is more aggressive than widely understood. The process begins when the manufacturer submits the marketing authorisation application to the EMA. Scope and PICO finalisation completes 10 days after the CHMP adopts the list of questions. The manufacturer must submit the dossier within 100 days — reduced to 60 days for accelerated procedures and certain variations — and in all cases no later than 45 days before the CHMP opinion. The final JCA report must be completed by the date of marketing authorisation, and the HTACG must endorse it within 30 days of the Commission decision.

That is not a 90-day window. The 90-day figure circulating in the industry is wrong. It is 100 days standard, 60 days accelerated, with a hard backstop of 45 days pre-CHMP opinion.

The first year’s output was smaller than projected. The 2025 programme anticipated 17 oncology JCAs and 8 ATMP JCAs. Thirteen were initiated — ten oncology, three ATMP. None reached endorsement by year-end. By December 31, 2025, there was one confirmed dossier, three in confirmation check, five in dossier preparation, and four still in scoping.

The 2026 programme scales dramatically: approximately 35 oncology JCAs, 15 ATMP, 3 variations, and — for the first time — approximately 5 joint device assessments beginning in June 2026.

The JCA does not tell a country what to pay. It tells a country what the evidence says — using a single assessment that every member state must consider. For companies that designed different evidence narratives for different markets, that flexibility just narrowed to a corridor.

The five questions — a decision tree for 2026

The synthesis of every data point, every regulatory shift, every basket design, and every access timeline in this analysis converges on five sequential questions. No single source publishes this framework. It is the inference that emerges when the MFN architecture, the JCA timeline, and the W.A.I.T. data are placed on the same table.

Question one: Is this country in an MFN basket?

If yes — and fifteen of the nineteen GLOBE/GUARD countries are in Europe or the OECD — every price generated in that market feeds directly into the algorithm that determines US Medicare reimbursement. The launch decision is no longer bilateral between the manufacturer and the national payer. It is trilateral — manufacturer, national payer, and CMS.

Question two: Will the observable price be list or net?

Germany’s confidential reimbursement amounts. France’s €7.98 billion rebate architecture. The UK’s VPAS framework. The distinction between the price the world references and the price the health system actually pays is the central variable in corridor management. A market that publishes a low list price causes more corridor damage than a market that negotiates a low net price confidentially.

Question three: How fast does the first real patient get access?

Germany: 52 days. England: 310 days. Italy: 391 days. France: 523 days. Spain: 567 days. The spread between the fastest and slowest major European market is 515 days — nearly eighteen months. That gap is not just a revenue timing issue. It is a window during which the manufacturer is generating — or not generating — price signals, evidence, and reference points.

Question four: Can the clinical assessment be reused from the JCA, or does the market require additional national friction?

The JCA report must receive due consideration in every EU member state. But member states retain the right to add national evidence requirements — cost-effectiveness analysis, budget impact modelling, local clinical practice data. The markets that accept JCA evidence with minimal additional friction become faster to access. The markets that layer national requirements on top become slower. The JCA does not eliminate national HTA. It creates a two-speed Europe — and launch sequencing must account for which speed each market operates at.

Question five: Does this market generate enough early NPV to compensate for the potential corridor damage?

The eight core markets — the United States, Japan, Germany, France, Italy, Spain, the United Kingdom, and China — concentrate 89% of cumulative sales in the first five years post-launch. But that 89% figure masks enormous variation in timing. Germany’s 52-day access delivers revenue in month two. France’s 523-day access delivers revenue in month eighteen. The NPV of early revenue must now be calculated net of the corridor damage that early pricing creates — and that calculation has never been harder.

The early window still matters enormously. But its value can no longer be measured only in revenue brought forward. It must be measured in US price preserved. That is the conceptual pivot of launch sequencing in 2026.

Four sequencing logics for the new architecture

Applied to the current landscape, the decision tree produces four observable sequencing strategies — not as fixed templates, but as gravitational logics around which different therapeutic areas and different corporate risk appetites will organise.

The first logic: the United States remains the centre. It remains the first launch market for the majority of innovative medicines, and it remains the market where the premium price is set. What has changed is that the US price is no longer independent of international pricing decisions. Every subsequent launch in a reference country carries a calculable impact on Medicare reimbursement. The US-first logic persists — but the second and third moves in the sequence now carry US price consequences that did not exist two years ago.

The second logic: Germany as the European signal market. Germany’s combination of near-immediate access, early price signal, and — now — confidential reimbursement capability makes it the natural first European launch. The 52-day median access time means the price signal generates within two months of EMA approval. The confidentiality option means the negotiated net price need not cascade. But the free-pricing period has been halved to six months, and the dynamic manufacturer discount introduced by the GKV-Beitragssatzstabilisierungsgesetz will compress AMNOG-negotiated prices over time.

The third logic: the United Kingdom as a rising sequencing priority. The UK–US deal, the NICE threshold increase, the aligned MHRA–NICE pathway, and the 25% net price commitment collectively elevate the UK from mid-tier to early-sequence. The three-to-six-month access acceleration, the structural price protection, and the VPAS cap create a market that is — for the first time — designed to attract early launches. Companies that are among the 27 early adopters of the aligned pathway have a six-month head start on the new architecture.

The fourth logic: France and Spain as late or selective entries. France’s structural access delay, its bifurcated pathway system, and its confidential rebate architecture make it a candidate for Accès Précoce where eligible and standard late-sequence entry where not. Spain’s 567-day median access time and its absence from the GENEROUS basket — though presence in GLOBE/GUARD — position it similarly: a market where filing delay is often deliberate, and where the corridor calculus rarely favours early entry.

The old sequence was linear: US, Germany, UK, France, Italy, Spain, Japan. The new sequence is conditional: the order depends on which basket the country sits in, what price it publishes, how fast it grants access, and whether the JCA evidence carries through. There is no single correct sequence. There are five questions — and every product must answer them differently.

The consequence nobody is modelling

The structural incentive created by the MFN framework extends beyond Europe. Markets outside the US reference baskets — the Middle East, Latin America, Southeast Asia — gain relative attractiveness as early-revenue markets with lower corridor contamination risk. This does not yet manifest as a documented systematic shift in launch sequencing toward non-referenced markets. But the economic logic is clear: the GENEROUS, GLOBE, and GUARD baskets are heavily concentrated in Europe and high-income OECD countries. Every market outside those baskets is, by definition, a market where early pricing does not feed the US benchmark.

Reuters documented a 35% decline in new pharmaceutical launches across Europe in the ten months following the MFN executive order. That decline is not random. It is the market solving the optimisation problem that this decision tree frames — withdrawing from markets where the corridor damage exceeds the revenue gain.

The companies that navigate 2026 successfully will not be the ones with the fastest launches or the highest prices. They will be the ones that understood, before their competitors, that the launch sequence is no longer a revenue maximisation problem.

It is a corridor preservation problem. And the corridor is no longer European. It is transatlantic.

Every launch plan in the industry was built for a world where European prices were a consequence. They are now an input — into the algorithm that determines what Medicare pays, into the JCA that structures what every EU member state evaluates, and into the reference pricing cascade that connects Paris to Lisbon to Athens to Washington. The companies that redesign their sequences around these five questions will preserve their pricing architecture. The companies that do not will watch it erode — one launch at a time.

Subscribe now

Beyond Approval publishes weekly strategic intelligence on market access, HTA, pricing, and regulatory shifts that reshape how medicines reach patients — written from inside the industry, for the people whose decisions move drugs to patients.

It comes from seven years sitting on the other side of the table — inside reimbursement committees, reading dossiers, watching submissions succeed and fail, and understanding why the outcome rarely had everything to do with the science.

This framework is built on one core premise: reimbursement committees don’t just evaluate evidence. They evaluate credibility, risk, and trust. And most pharma teams walk into that process without understanding how the room actually works.

Inside this document you’ll find the seven criteria committees use to assess a submission, the five arguments that consistently move decisions, the five mistakes that consistently kill them, and the unwritten rules that nobody puts in the official guidance but that shape the outcome of every single negotiation.

This is the resource I wish had existed when I was on the industry side.

⬇️ Download the PDF below. Read it before your next submission.