Summary, in English
Given a growing population of older adults and the common occurrence of natural as well as pathological cognitive decline, i.e. dementia, there is a need to better understand how to accurately measure and classify level of cognitive functioning. Both using multiple test administrators to collect cognitive data and what year the participant is born (birth cohort effects) can influence participants’ performance on cognitive tests. The concept Mild cognitive impairment (MCI) aims to capture a prodromal stage of dementia. There is heterogeneity in the reported MCI prevalence and incidence estimates, partly due to varying MCI definitions applied in the research literature. Also, estimations of reversion rates (recovering from MCI) are reported to be fairly high (≈29%). Sampled participants used in these studies, aged 60-100, were randomly invited to partake in the Swedish population-based ageing study: Good Aging in Skåne.Paper I (n of participants =6686, Mage=71.3, n of test administrators = 21) examined test administrator influence on participants’ test scores. A series of mixed linear models revealed significant random effects corresponding to the test administrators for cognitive test measuring episodic memory, speed of processing and spatial ability (p<.01). The variation seen in test scores ascribed to the test administrator was between 1.4%-3.5%.Paper II (n of 60-year olds = 736 born between 1942-55, n of 81-year olds = 431 born between 1920-33) examined differences in participants’ tests scores in 3 sets of cohorts, at age 60 or 81, respectively. The ANOVAs found significant (p<0.05) differences in test scores measuring speed of processing, episodic memory, attention, executive functioning and vocabulary between the birth cohorts, where the later born cohorts outperformed the earlier born cohorts.Paper III (n for prevalence =3752, n for incidence = 1451, age-groups= 60-69, 70-79, 80+) examined prevalence and incidence of MCI across age, sex, subtypes and criteria of cognitive impairment. Prevalence of MCI were 21.4% and 6.6% for a lenient and strict inclusion criterion of cognitive impairment, respectively. The MCI incidence rates were 22.6 and 8.67 per 1000 person-years for a lenient and strict criterion, respectively. No sex-differences in MCI estimates were established and age differences in estimates were inconclusive. Paper IV (n=331) examined a 6-year MCI reversion rate and investigated factors that predicted reversion. There was a high reversion rate of 58%. The logistic regression found that lower age (p<0.05), better global cognitive functioning (p<0.02), good concentration (p<0.05) and single domain subtype (p<0.001) could predict reversion. In summary, test administrator effects and birth cohort effects were present in the cognitive data of this population based ageing study. Moreover, prevalent MCI was fairly common in a sample aged 60+, and that occurrence of MCI varies with the applied definition. Over half of the participants with MCI reverted back to normal cognitive functioning at 6-year follow-up. Correctly assessing cognition is important and there are many methodological aspects to take into consideration, from the stage of data collection to the stage of defining cognitive impairment.