Umeå University's logo

In the present study, we examine long-term population-level effects on episodic memory of an intervention over 15 years that reduces systolic blood pressure in individuals with hypertension. A limitation with previous research on the potential risk reduction of such interventions is that they do not properly account for the reduction of mortality rates. Hence, one can only speculate whether the effect is due to changes in memory or changes in mortality. Therefore, we extend previous research by providing both an etiological and a prognostic effect estimate. To do this, we propose a Bayesian semi-parametric estimation approach for an incremental threshold intervention, using the extended G-formula. Additionally, we introduce a novel sparsity-inducing Dirichlet prior for longitudinal data, that exploits the longitudinal structure of the data. We demonstrate the usefulness of our approach in simulations, and compare its performance to other Bayesian decision tree ensemble approaches. In our analysis of the data from the Betula cohort, we found no significant prognostic or etiological effects across all ages. This suggests that systolic blood pressure interventions likely do not strongly affect memory, either at the overall population level or among individuals who would remain alive under both the natural course and the intervention (the always survivor stratum).

Causal mediation analysis of observational data is an important tool for investigating the potential causal effects of medications on disease-related risk factors, and on time-to-death (or disease progression) through these risk factors. However, when analyzing data from a cohort study, such analyses are complicated by the longitudinal structure of the risk factors and the presence of time-varying confounders. Leveraging data from the Atherosclerosis Risk in Communities (ARIC) cohort study, we develop a causal mediation approach, using (semi-parametric) Bayesian Additive Regression Tree (BART) models for the longitudinal and survival data. Our framework is developed using static longitudinal exposure regimes and allows for time-varying confounders and mediators, both of which can be either continuous or binary. We also identify and estimate direct and indirect causal effects in the presence of a competing event. We apply our methods to assess how medication, prescribed to target cardiovascular disease (CVD) risk factors, affects the time-to-CVD death.

Motivated by the Swedish Betula study, we consider the joint modeling of longitudinal memory assessments and the hazard of dementia. In the Betula data, the time-to-dementia onset or its absence is available for all participants, while some memory measurements are missing. In longitudinal studies of aging, one cannot rule out the possibility of dropout due to health issues resulting in missing not at random longitudinal measurements. We, therefore, propose a pattern-mixture sensitivity analysis for missing not-at-random data in the joint modeling framework. The sensitivity analysis is implemented via multiple imputation as follows: (i) multiply impute missing not at random longitudinal measurements under a set of plausible pattern-mixture imputation models that allow for acceleration of memory decline after dropout, (ii) fit the joint model to each imputed longitudinal memory and time-to-dementia dataset, and (iii) combine the results of step (ii). Our work illustrates that sensitivity analyses via multiple imputations are an accessible, pragmatic method to evaluate the consequences of missing not at-random data on inference and prediction. This flexible approach can accommodate a range of models for the longitudinal and event-time processes. In particular, the pattern-mixture modeling approach provides an accessible way to frame plausible missing not at random assumptions for different missing data patterns. Applying our approach to the Betula study shows that worse memory levels and steeper memory decline were associated with a higher risk of dementia for all considered scenarios.

Aim: This study investigates the effects of degree of multilingualism on cognitive functions in adulthood, with focus on episodic memory recall and including measures of verbal fluency as well as global cognition.

Design: We studied a large population-based cohort cross-sectionally, and we also assessed changes over time through longitudinal measurements on four time-points over a 15 year period. Participants were drawn from the Betula prospective cohort study in Umeå, Sweden. The participants included in this study at baseline (n = 894, mean age = 51.44, 59.4% females) were divided according to number of languages into bilinguals (n = 395), trilinguals (n = 284), quadrilinguals (n = 169), and pentalinguals (n = 46). 

Data and analysis: We analysed performance on tasks of episodic memory recall, verbal fluency (letter and category) and global cognition (Minimental State Examination, MMSE) both cross-sectionally and longitudinally. The control background variables were baseline age, gender, years of education, general fluid ability Gf (Wechsler Block Design Test), and socioeconomic status. We employed a linear mixed modelling approach with entropy balancing weights to assess effects of degree of multilingualism on cognitive functions.

Findings and conclusions: Using bilinguals as the reference group, our results indicated that all the other multilingual groups exhibited superior performance on episodic memory recall than bilinguals at baseline. The rate of change over time did not differ for trilinguals and pentalinguals compared to bilinguals. While quadrilinguals declined more over time than bilinguals, they still scored significantly higher than bilinguals at the last test wave. For letter fluency, similarly, all language groups scored higher than bilinguals at baseline, and none of the groups differed from bilinguals in rate of change over time. With regard to category fluency, quadrilinguals scored higher than bilinguals at baseline, but trilinguals and pentalinguals did not differ from bilinguals and none of the groups differed in change over time compared to bilinguals. Finally, for global cognition (MMSE), trilinguals and quadrilinguals scored significantly higher than bilinguals at baseline with no differences in change over time for any of the groups relative to bilinguals. Our study contributes to the understanding of multilingual cognition and sheds light into an under-researched cognitive domain known to decline in normal ageing, namely episodic memory recall.

Significance: Our study emphasizes the importance of researching less explored aspects of multilingualism on cognition, in particular on episodic memory recall, to aid our understanding of factors that could potentially aid cognitive decline in later adulthood.

Objective: Previous research has shown that daily activities are crucial for mental health among older people, and that such activities declined during the COVID-19 pandemic. While previous studies have confirmed a link between stringent restrictions and an increase in mental ill-health, the role of daily activities as a mediator in this relationship remains underexplored. We analyzed whether reductions in daily activities mediated the impact of these COVID-19 restrictions on mental ill-health during the pandemic’s initial phase.

Methods: We used data from Wave 8 SHARE Corona Survey covering 41,409 respondents from 25 European countries and Israel as well as data on COVID-19 restrictions from the Oxford Government Response  Tracker  (OxCGRT).  Multilevel  regression  and  multilevel-mediation  analysis  were  used  to  examine the relationships between restrictions, daily activities and mental ill-health.

Results: Reductions in walking and shopping showed a notably stronger association with increases in mental ill-health compared to social activities. Furthermore, declines in walking could account for about  a  quarter  of  the  relationship  between  restrictions  and  increased  mental  ill-health,  but  the  mediating effects of the other activates were negligible.

Conclusions: The study highlights the essential role of maintaining daily activities, particularly walking, to  mitigate  the  negative  psychological  effects  of  pandemic-related  restrictions  among  older  populations in Europe.

The prevalence of interpersonal violence has been reported at higher levels among Indigenous than non-Indigenous populations worldwide, but has not been thoroughly investigated among the Sámi population in Sweden. The aims of this study were to investigate: (1) the prevalence of emotional, physical, and sexual violence and violence by intimate partners, family members, acquaintances, and strangers among participants identifying as Sámi or Swedish, (2) whether reporting experiences of historical losses and discrimination mediated the anticipated association between identifying as Sámi and reporting experiences of violence, and (3) whether background characteristics were associated with reporting experiences of violence. Cross-sectional questionnaire data collected in 2021 for the “Health and Living conditions in Sápmi” study were used. All adults in an arctic region in Sweden were invited to participate (response rate: 41%). Respondents self-identifying as Sámi (n = 375; 24.7%) or Swedish (n = 1,144; 75.3%) were included in this study. Sámi respondents of both sexes more often reported violence by an acquaintance or stranger. Likewise, more Sámi than Swedish women reported family violence (16.4% vs. 9.2%), but there was no difference concerning intimate partner violence (13.3% vs. 15.4%). Mediation analyses revealed strong positive indirect effects of historical losses and discrimination on the different types of violence. Being female was the strongest predictor of reporting intimate partner violence, and younger age was associated with violence by all perpetrators except family members. In conclusion, interpersonal violence was more often reported by Sámi respondents, but the association was explained in full by experiences of historical losses and discrimination. The results underline the importance of a life-course and even intergenerational and historical perspectives when investigating interpersonal violence.

The present study aimed to characterize profiles of cognitive aging and how these can be predicted frominterindividual differences in demographic, lifestyle, health, and genetic factors. The participants were1,966 older adults (mean baseline age= 71.6 years; 62.9% female), free from dementia at baseline and with atleast two cognitive assessments over the 15-year follow-up, from the population-based Swedish NationalStudy on Aging and Care in Kungsholmen. The cognitive assessment comprised tests of semantic andepisodic memory, letter and category fluency, perceptual speed, and executive function. First, we estimatedthe level and change within each of the cognitive domains with linear mixed effect models, based on whichwe grouped our sample into participants with “maintained high cognition,” “moderate cognitive decline,” or“accelerated cognitive decline.” Second, we analyzed determinants of group membership within eachcognitive domain with multinomial logistic regression. Third, group memberships within each cognitivedomain were used to derive general cognitive aging profiles with latent class analysis. Fourth, thedeterminants of these profile memberships were analyzed with multinomial logistic regression. Follow-upanalyses targeted profiles and predictors specifically related to the rate of cognitive change. We identifiedthree latent profiles of overall cognitive performance during the follow-up period with 31.6% of the samplehaving maintained high cognition, 50.6% having moderate cognitive decline, and 17.8% having acceleratedcognitive decline. In multiadjusted analyses, maintained high cognition was predicted by female sex, highereducation, and faster walking speed. Smoking, loneliness, and being an ε4 carrier were associated with alower likelihood of maintained high cognition. Higher age, diagnosis of diabetes, depression, and carryingthe apolipoprotein E ε4 allele increased the likelihood of accelerated cognitive decline. Factors at baselinethat could significantly predict profile membership within the specific cognitive domains included age, sex,years of education, walking speed, diabetes, and the ε4 allele. Of note, these factors differed across cognitivedomains. In sum, we identified demographic, lifestyle, health, and genetic factors of interindividualdifferences in domain-specific and general cognitive aging profiles, some of which are modifiable.

Studies of memory trajectories using longitudinal data often result in highly non-representative samples due to selective study enrollment and attrition. An additional bias comes from practice effects that result in improved or maintained performance due to familiarity with test content or context. These challenges may bias study findings and severely distort the ability to generalize to the target population. In this study we propose an approach for estimating the finite population mean of a longitudinal outcome conditioning on being alive at a specific time point. We develop a flexible Bayesian semi-parametric predictive estimator for population inference when longitudinal auxiliary information is known for the target population. We evaluate sensitivity of the results to untestable assumptions and further compare our approach to other methods used for population inference in a simulation study. The proposed approach is motivated by 15-year longitudinal data from the Betula longitudinal cohort study. We apply our approach to estimate lifespan trajectories in episodic memory, with the aim to generalize findings to a target population.

This study aims to evaluate the performance of Item Response Theory (IRT) kernel equating in the context of mixed-format tests by comparing it to IRT observed score equating and kernel equating with log-linear presmoothing. Comparisons were made through both simulations and real data applications, under both equivalent groups (EG) and non-equivalent groups with anchor test (NEAT) sampling designs. To prevent bias towards IRT methods, data were simulated with and without the use of IRT models. The results suggest that the difference between IRT kernel equating and IRT observed score equating is minimal, both in terms of the equated scores and their standard errors. The application of IRT models for presmoothing yielded smaller standard error of equating than the log-linear presmoothing approach. When test data were generated using IRT models, IRT-based methods proved less biased than log-linear kernel equating. However, when data were simulated without IRT models, log-linear kernel equating showed less bias. Overall, IRT kernel equating shows great promise when equating mixed-format tests.