Early intervention in Alzheimer's disease: a health economic study of the effects of diagnostic timing

Abstract

Background: Intervention and treatment in Alzheimer's disease dementia (AD-dementia) can be cost effective but the majority of patients are not diagnosed in a timely manner. Technology is now available that can enable the earlier detection of cognitive loss associated with incipient dementia, offering the potential for earlier intervention in the UK health care system. This study aimed to determine to what extent the timing of an intervention affects its cost-effectiveness.

Methods: Using published data describing cognitive decline in the years prior to an AD diagnosis, we modelled the effects on healthcare costs and quality-adjusted life years of hypothetical symptomatic and disease-modifying interventions. Early and standard interventions were assumed to have equal clinical effects, but the early intervention could be applied up to eight years prior to standard diagnosis.

Results: A symptomatic treatment which immediately improved cognition by one MMSE point and reduced in efficacy over three years, would produce a maximum net benefit when applied at the earliest timepoint considered, i.e. eight years prior to standard diagnosis. In this scenario, the net benefit was reduced by around 17% for every year that intervention was delayed. In contrast, for a disease-modifying intervention which halted cognitive decline for one year, economic benefits would peak when treatment effects were applied two years prior to standard diagnosis. In these models, the maximum net benefit of the disease modifying intervention was fifteen times larger than that of the symptomatic treatment.

Conclusion: Timeliness of intervention is likely to have an important impact on the cost-effectiveness of both current and future treatments. Healthcare policy should aim to optimise the timing of AD-dementia diagnosis, which is likely to necessitate detecting and treating patients several years prior to current clinical practice.

Figures

Figure 1

Effects on cognitive course of symptomatic and disease-modifying interventions applied 5, 8 or 1 year prior to a standard diagnosis of Alzheimer’s disease. Panel A symptomatic intervention at 5 (top), 8 (middle), and 1 (lower) years; Panel B disease-modifying intervention at 5 (top), 8 (middle), and 1 (lower) years.

Figure 2

Cost-effectiveness of early intervention with a modest symptomatic treatment in the nine years prior to standard detection and treatment. Panel A net benefit; B Gain in quality-adjusted life years (QALYs); C incremental costs. Symptomatic effects = 1 MMSE point, efficacy reduces linearly over 3 years. Costs calculated over a 10-year horizon. Mean cohort age at entry 75 years.

Figure 3

Sensitivity of net benefits obtained under a symptomatic treatment model to costing horizon and treatment effect size. Panel A: results when costing over 5, 10 and 20 years; Panel B: result of doubling or halving of treatment effect on MMSE score; Panel C: effect of mean cohort age at model start.

Figure 4

Cost-effectiveness of early intervention with a hypothetical disease-modifying treatment in the nine years prior to standard detection and treatment. Panel A net benefits; B Gain in quality-adjusted life years (QALYs); C incremental costs. Treatment effect is a 12-month halt in cognitive deterioration. Costs calculated over a 10-year horizon. Mean cohort age at entry 75 years. Panel A, net benefits; Panel B, gain in quality-adjusted life years (QALYs); Panel C cost savings. Treatment effect is a 12-month halt in cognitive deterioration. Costs calculated over a 10-year horizon. Mean cohort age at entry 75 years.

Figure 5

Sensitivity of net benefits obtained through a hypothetical disease-modifying treatment to costing horizon, treatment effect and age of cohort. Panel A: results when costing over 5, 10 and 20 years; Panel B: result of increased and decreased treatment effect, as expressed in months delay in deterioration in MMSE score; Panel C: effect of mean cohort age at model start.