In order to solve the problems mentioned in the previous section, cognitive impairment could be defined as scoring lower than would be expected given the persons age / sex / education etc. This is normally done using the standard deviation. An example from a Swedish population?based study, the Kungsholmen Project, is shown in figure (Figure 7).

The 1435 nondemented persons in the study were split into 14 groups according to their age and educational level. Mean performance on MMSE was then computed for each group, as well as the standard deviation. All persons scoring one standard deviation below the mean for their age and education were then classified as cognitively impaired. Other cutoffs based on 1.5SD or 2SD etc are also commonly used.

The advantage of this approach is that it takes into account normal levels of functioning for specific cohorts, ages, etc. However, there is one major disadvantage. Using a statistical calculation such as the standard deviation results in a specific prevalence rate. Fifteen percent of the population will, by definition, score 1SD below normal on the task, so using this approach alone would mean that the prevalence of cognitive impairment will be 15% in all age groups (Figure 8).

Página en inglés. Es necesario registro en www.ttmed.com/dementia