Germany may be about to make AMNOG more adaptive. But the important question is not whether there will be a second assessment. It is whether the system knows, from the outset, what it expects to learn between the first and the second. 

Reports from last week’s Pharma Dialog at the German Federal Ministry of Health suggest that medicines in particular therapeutic situations – including orphan medicines and products entering through accelerated or conditional regulatory pathways – could receive an initial benefit assessment based on the evidence available at launch, followed by reassessment after approximately three years using additional evidence, including potentially routine care data from Germany’s Forschungsdatenzentrum Gesundheit (FDZ Gesundheit – Health Research Data Centre). 

For most new medicines, the existing AMNOG pathway would reportedly remain largely unchanged. 

‘No Ministry proposal has yet been published. What follows is therefore our interpretation of the reported discussions, not an assessment of draft legislation.’

The direction makes sense 

Modern drug development increasingly creates a problem that conventional HTA struggles to solve. 

Particularly in oncology, rare diseases and advanced therapies, important uncertainty may still exist at launch: long-term outcomes are immature, populations are small, treatment pathways evolve quickly, or generating conventional comparative evidence may simply take longer than regulatory decision-making allows. 

German HTA has traditionally required many of these uncertainties to be resolved at one point in time: market entry. If the evidence available then does not adequately answer the questions of the benefit assessment, the consequences for added benefit – and ultimately reimbursement – can be substantial. 

Acknowledging that some questions can only be answered after broader clinical use is therefore a sensible evolution. 

But moving the decision three years downstream does not automatically improve the decision. 

Reassessment needs prospective rules 

Germany already has some experience with repeated benefit assessment, and it provides a useful warning. 

An IGES analysis of orphan medicines reassessed after exceeding the statutory revenue threshold found a deterioration in the extent of added benefit in 32 of 52 patient populations, while only five saw an improvement or an added benefit quantified for the first time. 

This is not a direct analogue to the model now being discussed: importantly, the methodology changes once the orphan privilege is lost. Nevertheless, the experience demonstrates how strongly asymmetric the consequences of reassessment can become. 

A future two-stage AMNOG should therefore not wait until year three to decide what the second assessment is supposed to answer. 

At the first assessment, there should already be clarity about which uncertainties remain decision-relevant, which evidence could resolve them, which data sources are appropriate, and under what conditions new evidence can change the previous conclusion – in either direction. 

That does not mean freezing today’s comparator or treatment standard for years. Clinical practice will change. But it does mean agreeing prospectively on the questions, evidentiary principles and rules by which future learning will be judged. 

Symmetry matters too. If new evidence can reduce a benefit assessment and affect reimbursement, genuinely stronger evidence should also be capable of improving it under equivalent standards. 

The FDZ is promising – but the expectations placed on it are running ahead of experience 

The proposed role of the FDZ deserves particular scrutiny. 

Politically, expectations are high. The German Federal Ministry of Health describes the FDZ as the central infrastructure for the secure provision and use of health data for research and as an important step towards a more digital and knowledge-generating healthcare system. 

That ambition is understandable. The resource is potentially remarkable: statutory claims data covering around 74 million insured people, longitudinally available back to 2009, offer a scale and completeness that Germany has never had in one research environment. 

But operationally, the FDZ is still a very young infrastructure. 

It only went live in October 2025. By February 2026, more than 50 applications had been submitted, while the first application had only just completed the full process from submission to decision. The FDZ itself describes its procedures, quality assurance and technical infrastructure as subject to continuing development. 

More importantly, what is available today is primarily claims data. Data volunteered from electronic patient records are expected to become available from October 2026, while linkage with additional sources such as cancer registries remains a future development. 

Claims data can be extremely powerful for mortality, hospitalisation, treatment patterns, healthcare utilisation and many other questions of real-world care. But they do not systematically and reliably capture many of the variables at the centre of HTA disputes – such as disease progression, detailed tumour characteristics, laboratory parameters or patient-reported outcomes. 

This creates a potential mismatch. 

The policy debate appears ready to assign the FDZ an important role in resolving uncertainty for future reimbursement decisions. Yet the infrastructure has had less than a year of operational experience and is still developing precisely the richer clinical data linkages that may be required for some of those decisions. 

That is not an argument against the FDZ. Quite the opposite: its potential is substantial. 

But potential data availability and proven fitness for an HTA decision are not the same thing. 

The decisive question is therefore not simply whether real-world data should play a larger role in AMNOG. It is whether the data available for a particular reassessment can actually resolve the uncertainty that triggered that reassessment in the first place. 

The assessment architecture must not run ahead of the data infrastructure required to support it. 

For manufacturers, launch may become less of a finish line 

A genuine two-stage system would also change strategic evidence planning. 

Regulatory approval, EU Joint Clinical Assessment, initial German benefit assessment, post-launch evidence generation and subsequent reassessment could no longer be treated as largely separate exercises. 

Companies would increasingly need to understand, years before launch, which uncertainties can realistically be resolved before market entry, which may remain afterwards, and how evidence generated across the product lifecycle fits together. In selected cases, that will require an evidence architecture spanning several decision points rather than a succession of stand-alone evidence packages. 

There is also an economic consequence. 

If reimbursement remains materially open to revision several years after launch, the uncertainty surrounding future German cash flows increases. Depending on the final design, that may affect risk-adjusted launch economics, evidence-investment decisions and potentially launch sequencing. 

The real reform is not the second decision 

AMNOG 2.0 could address a genuine weakness of current HTA: the expectation that questions arising from immature evidence can always be answered at launch. 

A second assessment can help. 

But the second assessment is not the innovation. 

The real innovation would be a framework that establishes at the first decision point what remains uncertain, what evidence should resolve it, which data can credibly answer the question and how subsequent learning will translate into value. 

If Germany gets those rules right, AMNOG could move from judging uncertainty at a single moment to managing it over time. 

If it does not, the system may simply replace one uncertain decision at launch with another uncertain decision three years later. 

That would be reassessment. 

It would not yet be learning. 

Sources / background 

• Gunnar Göpel, “AMNOG-Reform: BMG-Plan sieht starken Umbau und Neubewertung vor”, Tagesspiegel Background, 4 September 2026.
• IGES Institut, analysis of orphan-drug reassessments after exceeding the statutory revenue threshold.
• Federal Ministry of Health, FDZ Gesundheit – questions and answers.
• Federal Ministry of Health, opening of the FDZ Gesundheit.
• Deutsches Ärzteblatt, “Forschungsdatenzentrum Gesundheit wird kontinuierlich weiterentwickelt”.