Exploring the impact of prevalence and mortality on incidence of dementia in the oldest old: the sensitivity of a deterministic approach.

BACKGROUND

There has been considerable debate about whether dementia is an age-related or an aging-related phenomenon. When the evidence is restricted to prevalence data, the inferences are not sound because incidence, mortality and differential mortality are not taken into account. The aim of the paper is to demonstrate the use and sensitivity of a deterministic model based on the relationship between mortality, incidence and prevalence with the example of the chronic disease dementia.

DESIGN

The simple deterministic model used here allows the calculation of one component given the other three. The sensitivity of calculated incidence to individual and combined changes in mortality, differential mortality (as measured by the mortality odds ratio) and prevalence is examined using published data.

MAIN RESULTS

Calculated incidence continues to increase with age despite extreme but plausible changes in each of the other components. Calculated incidence declines amongst the oldest old only when the changes are combined.

CONCLUSION

The deterministic model is shown to provide robust interrogation of the relationship between prevalence, incidence and survival. The effect incidence and survival have on the age-related and aging-related debate is illustrated. Only with extreme assumptions can the incidence of dementia be lessened. Clearly, more information on the oldest old is needed, but incidence studies of sufficient size at such great ages are problematic, and therefore there is a need to use flexible models for maximising the value of empirical data.