Our Model

Our methodology is based on a multi-stage probabilistic model for the incidence and progression of Alzheimer's disease (see figure below). According to the model, healthy persons may transition to stage 1 disease and then to stage 2 disease; where stage 1 and 2 disease could be defined as mild or moderate and late stage disease, respectively. A healthy person has an annual probability of onset of Alzheimer's disease (r) which may vary by age, gender and calendar year. A person with stage 1 Alzheimer's disease has an annual probability of transitioning to stage 2 disease (&gamma) which may vary by calendar year. The annual probability of onset of Alzheimer's disease and the annual transition rate to stage 2 disease depend on calendar year to allow for the introduction of interventions that delay disease onset or progression.

Figure: Multi-stage probablistic model for the incidence and progression of Alzheimer's disease

Death acts as a competing risk at each stage of the model. Healthy persons have an annual probability of dying (d0) which depends on age and calendar year; we refer to this annual probability of death as the background mortality. Persons with stage 1 and 2 disease have annual probability of death d1 and d2, respectively.

Information on the annual probability of onset of Alzheimer's disease, the annual probability of transitioning to stage 2 disease and the annual probabilities of dying are used to generate age, gender and stage specific prevalence rates of Alzheimer's disease. To forecast Alzheimer's disease prevalence, the prevelance rates are applied to demographic population projections.

The program available on this website will produce Alzheimer's disease incidence and prevalence projections for the years 2001 to 2050 and age-specific prevalence rates by gender, stage of disease and age (60 to 100). In addition, Disability Adjusted Life-Years are estimated for the years 2001 to 2050 and the median survival for persons with Alzheimer's is estimated to measure the impact of Alzheimer's disease on mortality, where our survival function, S(T ≥ t), is defined to be the probability of experiencing death at age t or at any age greater than t. Projections of the cost of Alzheimer's disease are produced for forecasts for the United States or a specific country or region of interest to the user.

Key references:

The publication on our key findings can be found here

Further technical details of our model can be found here

The publication on our work estimating the incidence of Alzheimer's disease can be found here

The publication on our work estimating the impact of Alzheimer's disease on mortality can be found here

An exammple of the application of our model and web-based tool to project the prevalence of Alzheimer's in New Zealand can be found here