Presentation (CTI)

OBJECTIVES: Germany's AMNOG process evaluates the added benefit of new medicines against an appropriate comparator therapy (ACT). As evidence requirements and assessment practices evolve, outcomes may shift over time. We examined 15-year trends in added-benefit ratings and the influence of G-BA-defined subpopulations, comparator selection, and evidence generation on outcomes.
METHODS: All AMNOG assessments (2011-2025) were analyzed using the AMNOG Monitor. Outcomes were the proportion of dossiers receiving no-added-benefit and the magnitude of added benefit, compared across single versus multiple G-BA-subpopulations, active versus BSC/watch-and-wait comparators, and patient-individualized therapy (PIT) versus "OR"-linked comparators.
RESULTS: No-added-benefit ratings rose from ~50% (2011) to 67% (2025), while mean G-BA-subpopulations per dossier fell from 2.3 to 1.6. Dossiers with multiple subpopulations showed lower no-added-benefit rates (52% vs. 67%) and more high added-benefit ratings, more often including RCTs versus the ACT (64% vs. 46%). No added benefit occurred in 30% of BSC/watch-and-wait assessments versus 64% with active comparators. PIT use rose from 2% to 16% and "OR-linked" dossiers from 30% to 50%. PIT yielded more no-added-benefit than "OR" (68% vs. 63%); however, only 14% of PIT cases involved accepted evidence failing to demonstrate sufficient effects, versus 36% for "OR"—suggesting most no-added-benefit outcomes reflected evidence that could not be generated or accepted rather than absent treatment effects.
CONCLUSIONS: The decline in subpopulations per dossier (2.3 to 1.6) likely reflects narrower indication labels and smaller target populations. Multiple subpopulations favor positive ratings—when the pivotal RCT can be matched to a subpopulation whose population and ACT align with the trial, benefit is demonstrable even if the broader indication exceeds the study population. Fewer, smaller subpopulations make RCTs versus the ACT harder to conduct, plausibly driving rising no-added-benefit rates. This underscores the need for early G-BA alignment on subpopulations, comparators, and feasible evidence, and for adapting AMNOG to small populations through proportionate, attainable evidence standards.