Health‐economic innovation headroom in mild cognitive impairment due to Alzheimer’s diseas

Abstract Background Estimating health‐economic impact of new therapies for prodromal Alzheimer’s disease (AD) requires extrapolating short‐term effects on surrogate markers of efficacy to long‐term clinically relevant outcomes. This requires assumptions on the nature of the treatment (symptomatic versus disease‐modifying) and depends on the target population (which varies in disease progression associated to biological subtypes). We aim to estimate the innovation headroom (potential value) of AD treatment in mild cognitive impairment (MCI), describe the impact of assumptions on treatment nature (symptomatic/disease‐modifying) and disease progression (fast/slow progressors), and discuss challenges for health‐economic evaluation of future AD treatments; to ensure patients timely access treatment innovations. Method The IPECAD open‐source decision‐analytic model was used to estimating time spent with MCI, mild, moderate and severe dementia, discounted quality‐adjusted life years (QALYs) and costs, under different treatment strategies. Innovation headroom was estimated by simulating a usual care strategy, and compare it with a strategy reflecting a fictive disease‐modifying treatment with 100% efficacy, and a strategy reflecting a fictive 100% efficacious symptomatic therapy; in preventing all individuals to progress to dementia. Result The innovation headroom analysis resulted in 4.1 additional QALYs and k€13 lower costs of care with disease‐modifying, and 0.8 additional QALYs and k€71 lower costs of care with symptomatic therapy. The higher QALY gain for disease‐modification is mainly explained by an indirect effect on mortality, resulting in a 5.5 year increase in life expectancy. The innovation headroom was about twice as large in the persons with fast progression compared to persons with slow progression in MCI. The disease‐modifying therapy mainly generated improved health (QALY gains), while the symptomatic therapy mainly generated savings in costs of care; which is contingent on their mortality assumptions. Conclusion The potential health and economic benefit of (non‐)pharmacological interventions in MCI is substantial. However, its dependency on assumptions related to treatment nature argues for research efforts to better understand the relation between surrogate and long‐term outcomes. Its dependency on target population indicates a potential for individualized treatment based on profiles of disease progression. Finally, we recommend incorporating age‐related comorbidities as a result of the relatively large life years gained.