Screening and treatment for Alzheimer's disease: predicting population-level outcomes.

ABSTRACT

BACKGROUND: Advances in screening and treatment are needed to mitigate increasing prevalence of dementia due to Alzheimer's disease (DAT). Current proposals to revise Alzheimer's disease (AD) diagnostic criteria incorporate diagnostic biomarkers. Such revisions would allow identification of persons with AD pathology before the onset of dementia. The population-level impact of screening for preclinical AD and treating with a disease-modifying agent is important when evaluating new biomarkers and medications. METHODS: A published computer simulation model assigned AD-related event times, such that delays in disease progression due to therapy effectiveness can be estimated for a preclinical AD cohort. Attributes such as screening sensitivity/specificity, treatment efficacy, age at first screening, and rescreening intervals were varied. Outcomes included incident mild cognitive impairment (MCI-AD), incident DAT, and number of patients recommended for treatment. RESULTS: One-time screening at age 65 years, 50% efficacy, and literature-based proxy persistence rates yielded 12.4% incidence of MCI-AD and 0.9% decrease in DAT incidence from base case of no screening/treatment. Modest reductions in incident MCI-AD and DAT were observed with more sensitive testing. Reducing specificity yielded greater reductions in MCI-AD and DAT cases, albeit by treating more patients. Probabilistic sensitivity analysis predicted that for a cohort of patients aged 65 years, the number that needed to be treated to avoid one AD case was 11.6 (range 5.7-104). CONCLUSION: The reduction in MCI-AD and DAT depends on initial screening age, screening frequency, and specificity. When considering population-level impact of screening-treatment, the effect of these parameters on incidence would need to be weighed against the number of individuals screened and treated.