The most appropriate means of capturing data from the Neuropsychiatric Inventory (NPI) must be understood to optimize use of this instrument in clinical trials. The utility of the composite score (frequency times severity) was recently demonstrated in mild and moderate dementia. Determination of frequency compared to composite scores in mild cognitive impairment (MCI) warrants investigation.

We used the NPI data from a randomized, placebo-controlled, multi-center, 24-week, clinical trial involving 160 patients who were diagnosed with amnestic MCI and had clinically significant neuropsychiatric symptoms (NPS). We calculated standardized changes for both frequency and composite scores.

There were improvements in NPI composite scores in both active drug- and placebo-treated patients, with significant superiority of active drug. Standardized changes in severity and composite scores tended to be larger than those in the frequency scores, whereas discrimination between treatment groups was similar for all three scores.

Our findings support the hypothesis that in MCI, as in dementia, the NPI frequency score is not more sensitive to treatment-related change than the composite score. As the severity score adds information, the use of the composite score has better performance characteristics.

We assessed eight-year mortality rates and predictors in a rural cohort of elderly individuals with cognitive impairment.

A total of 1,035 individuals, including 155 (15.0%) individuals with cognitive impairment, no dementia (CIND), and 69 (6.7%) individuals with clinically diagnosed dementia were followed for eight years from 1997. The initial assessment involved a two-step diagnostic procedure performed during a door-to-door survey, and mortality data were obtained from the Korean National Statistical Office (KNSO). The relationship between clinical diagnosis and risk of death was examined using the Cox proportional hazards model after adjusting for age, sex, and education.

During follow-up, 392 individuals died (37.9%). Compared to persons without cognitive impairment, mortality risk was nearly double among those with CIND (hazard ratio [95% confidence interval], 1.92 [1.46–2.54]), and this increased more than three-fold among those with dementia (3.20 [2.30–4.44]). Old age and high scores on the behavioral changes scale at diagnosis were two common predictors of mortality among those with CIND and dementia. Among the items on the behavioral changes scale, low sociability, less spontaneity, and poor hygiene were associated with increased mortality in individuals with CIND. Conversely, low sociability, excessive emotionality, and irritability were associated with increased mortality in patients with dementia.

Both dementia and CIND increased mortality risk compared with normal cognition in this community cohort. It is important to identify and manage early behavioral changes to reduce mortality in individuals with CIND and dementia.

This pilot investigation aimed to assess the relative and absolute test–retest reliability of commonly used functional performance measures in older adults with dementia residing in residential aged care facilities.

A total of 12 participants were tested on the Short Physical Performance Battery (SPPB), the Balance Outcome Measure for Elder Rehab (BOOMER), hand grip strength, anthropometric measures and Bio-electric Impedance Analysis (BIA). This study utilized a seven-day test–retest evaluation. Intra-class Correlation Coefficients (ICC) were used to assess relative reliability, Typical Error of Measurement (TEM) was used to assess the absolute reliability, and Bland–Altman plots were used to assess group and individual levels of agreement.

With the exception of Standing Balance (ICC = 0.49), 2.4-m walk (ICC = 0.68), functional reach (ICC = 0.38), and static timed standing (ICC = 0.47), all measures demonstrated acceptable (>0.71) ICCs. However, only the anthropometric measures demonstrated acceptable levels of absolute reliability (>10% TEM). Bland–Altman analysis showed non-significant (p > 0.05) mean differences, and eight out of the 17 measures showing wide Limits of Agreement (LoA).

Current measures of functional performance are demonstrably inappropriate for use with a population of older adults with dementia. Authors suggest aligning current measurement strategies with Item Response Theory as a way forward.

General Practitioners (GPs) from underdeveloped countries apply cognitive impairment (CI) assessment tools translated and adapted to cultural setting from other idioms, mainly English. As schooling in elderly from underdeveloped countries tends to be relatively heterogeneous, it is necessary to establish normative and cut-off scores for these CI instruments that are based on studies conducted locally. Some CI screening instruments frequently used by Brazilian specialists in dementia were analyzed to determine which could be most useful to GPs in their working sets.

Two hundred forty-eight patients aged 65 years or older that had been assisted by GPs in a tertiary hospital in Brazil were evaluated. Based on the MMSE and/or Short-IQCODE scores, 52 probable cases were identified on the basis of clinical data, performances on the neuropsychological tests and questionnaires (Functional Assessment Questionnaire/FAQ, Category Verbal Fluency/CVF, Clock Drawing Test/CDT) and blood tests and brain CT.

The combination of a functional questionnaire with a cognitive instrument had higher sensitivity and specificity than using the instruments alone. A FAQ cut-off of 3 in conjunction with a CDT cut-off of 6 proved optimal (93% sensitivity and 92.5% specificity). A higher specificity (93.5%) was attained using a combination of the FAQ (cut-off of 3) with the CVF (cut-off of 10).

For low schooling elderly, the combination of the FAQ and CVF represented a very simple method of increasing the chances of correct screening. For those with higher schooling, the combination of the FAQ and CDT was more suitable.

Children of patients with young onset dementia (YOD) who are confronted with a parent who has a progressive disease, often assist in caregiving tasks, which may have a great impact on their lives. The objective of the present study is to explore the experiences of children living with a young parent with dementia with a specific focus on the children's needs.

Semi-structured interviews with 14 adolescent children between the ages of 15 and 27 years of patients with YOD were analyzed using inductive content analysis. Themes were identified based on the established codes.

The emerging categories were divided into three themes that demonstrated the impact of dementia on daily life, different ways of coping with the disease, and children's need for care and support. The children had difficulties managing all of the responsibilities and showed concerns about their future. To deal with these problems, they demonstrated various coping styles, such as avoidant or adaptive coping. Although most children were initially reluctant to seek professional care, several of them expressed the need for practical guidance to address the changing behavior of their parent. The children felt more comfortable talking to someone who was familiar with their situation and who had specific knowledge of YOD and the available services.

In addition to practical information, more accessible and specific information about the diagnosis and the course of YOD is needed to provide a better understanding of the disease for the children. These findings underline the need for a personal, family-centered approach.

Caregivers report early disturbances in social behavior among patients with behavioral variant frontotemporal dementia (bvFTD); however, there are few direct observational studies of these social behavioral disturbances. This study aimed to identify social behavioral themes in bvFTD by direct observation in naturalistic interactions. The identification of these themes can help caregivers and clinicians manage the social behavioral disturbances of this disease.

Researchers observed 13 bvFTD patients in their homes and community-based settings and recorded field notes on their interpersonal interactions. A qualitative analysis of their social behavior was then conducted using ATLAS.ti application and a constant comparison method.

Qualitative analysis revealed the following themes: (1) diminished relational interest and initiation, indicating failure to seek social interactions; (2) lack of social synchrony/intersubjectivity, indicating an inability to establish and maintain interpersonal relationships; and (3) poor awareness and adherence to social boundaries and norms. These themes corresponded with changes from caregiver reports and behavioral scales.

This analysis indicates that real-world observation validates the diagnostic criteria for bvFTD and increases understanding of social behavioral disturbances in this disorder. The results of this and future observational studies can highlight key areas for clinical assessment, caregiver education, and targeted interventions that enhance the management of social behavioral disturbances in bvFTD.

Occupation remains an unmet need in long-term dementia care. To increase residents’ occupation, knowledge of types of occupation related to wellbeing, and organizational and environmental characteristics encouraging involvement in these types of occupation, is indispensable.

In this explorative study, Dementia Care Mapping was used to study involvement in different types of occupation and wellbeing among 57 residents of 10 dementia care facilities. For each type of occupation, mean experienced wellbeing was studied. Occupation types with high mean wellbeing scores were classified as “wellbeing-enhancing occupation.” Care facilities were ranked according to the mean time residents spent in types of wellbeing-enhancing occupation. Using information on staff-to-resident ratio, individual space, and items of the Physical Environment Evaluation Component of Dementia Care Mapping, organizational and environmental characteristics of the facilities were compared to study their relationship with wellbeing-enhancing occupation.

Reminiscence, leisure, expression, and vocational occupation had greatest potential to enhance wellbeing, but these types were seldom offered. Much variation existed in the extent to which wellbeing-enhancing occupation was provided. Long-term care facilities that did so more frequently generally had a more homelike atmosphere, supported social interaction through the environment, and had no central activity program.

This study suggests that it is possible to engage residents in wellbeing-enhancing occupation, within current means of budget and staff. The physical environment and care organization might play a role, but the key factor seems to equip staff with skills to integrate wellbeing-enhancing occupation into care practice.

There is a paucity of research into PRN medication use in older psychiatric inpatients. This is an important topic given the risks of polypharmacy, adverse drug reactions, and high dose medication.

In 2013, we carried out a cross-sectional survey of the prescription and administration of sedative PRN medication to older adult inpatients on seven wards at a UK tertiary referral centre. We compared them with 242 patients of working age.

Of the 92 patients studied, 56 (60.9%) were prescribed PRN sedation and 25 (27.2%) had received one or more doses in the previous fortnight. In total, 70 doses had been administered; all by mouth and all but one as single doses. Lorazepam was by far the most commonly prescribed and administered PRN drug. Agitation was the main indication, although violence was the most commonly cited reason for administration but documentation of antecedents, non-pharmacological strategies and outcome including side effects was uniformly poor with only 37 (52.9%) doses recorded in the case notes. Those with organic disorders were just as likely to receive PRN as those with functional illnesses. Patients very rarely actually received high dose antipsychotics or antipsychotic polypharmacy as a result of PRN prescriptions. Older patients were less likely than adults of working age to be prescribed PRN and dosages were smaller.

Prospective studies of PRN prescription and administration are needed to better understand the reasons underpinning its use and to gain objective data upon its effectiveness or otherwise in this vulnerable patient group.

In dementia patients, dietary intake problems may occur despite the absence of swallowing problems. We investigated cognitive functions on food and taste in Alzheimer's disease (AD) and vascular dementia (VaD) patients.

Participants included 15 healthy controls (HC), 30 AD and 20 VaD patients. Food Cognition Test: Replicas of three popular foods in Japan with no odors were presented visually to each participant, with the instruction to respond with the name of each food. Replicas of food materials were subsequently presented to ask whether they were included in these foods. Taste Cognition Test: Replicas of 12 kinds of foods were presented to describe their expected tastes.

The AD/VaD groups exhibited significantly lower scores on Food/Taste Cognition Tests compared with the HC group. These scores correlated inversely with Mini-Mental State Examination (MMSE) scores in the AD group. Decreased dietary intake was observed in 12 of the 50 patients; 8 of the 12 exhibited decreased Taste Cognition Test scores, higher than that of the normal-intake patients. There was no difference in the filter paper taste disc test between HC/AD/VaD groups. To test the hypothesis that the insula is associated with taste cognition, two MMSE-matched AD subgroups (n = 10 vs. 10) underwent positron emission tomography. Glucose metabolism in the right insula was lower in the low taste cognition subgroup. The VaD patients with insular lesions exhibited impaired Taste Cognition Test findings.

It is important to consider the cognitive aspect of dietary intake when we care for dementia patients.

Neuropsychiatric symptoms (NPS) are prevalent in nursing-home (NH) patients with dementia, but little is known about the long-term course of these symptoms.

In this study, 931 NH patients with dementia took part in a prospective cohort study with four assessments over a 53-month follow-up period. NPS and level of dementia were assessed with the Neuropsychiatric Inventory scale and the Clinical Dementia Rating scale, respectively.

Mild, moderate, and severe dementia was present in 25%, 33%, and 42%, respectively. There was an increase in the severity of the dementia from the first to the fourth assessment. Agitation, irritability, disinhibition, and apathy were the most prevalent and persistent symptoms during the study period. The affective subsyndrome (depression and anxiety) became less severe, whereas the agitation subsyndrome (agitation/aggression, disinhibition, and irritability) and apathy increased in severity during the follow-up period. More severe dementia was associated with more severe agitation, psychosis, and apathy, but not more severe affective symptoms. Mild dementia was associated with an increase in the severity of psychosis, whereas moderate or severe dementia was associated with decreasing severity of psychosis over the follow-up period.

Nearly all the patients experienced clinically significant NPS, but individual symptoms fluctuated. Affective symptoms became less severe, while agitation and apathy increased in severity. An increase in dementia severity was associated with an increase in the severity of agitation, psychosis, and apathy, but not affective symptoms. The results may have implications when planning evaluation, treatment, and the prevention of NPS in NH patients.

Dementia affects 15% of Canadians 65 and older, and the prevalence is expected to double over the next two decades. Low socioeconomic status (SES) can increase the risk of Alzheimer's disease (AD) and the precursor mild cognitive impairment (MCI), but it is unknown what the relationship of SES is on initial clinical presentation to a memory disorders clinic.

Data from 127 AD and 135 MCI patients who presented to our Memory Disorders Clinic from 2004 to 2013 were analyzed retrospectively. We examined the relationship between SES (measured using Hollingshead two-factor index) and (1) diagnosis of either AD or MCI; (2) age when first presented to clinic; (3) objective cognitive tests to indicate clinical severity; and (4) the use of cognitive enhancers, medication for treating mild-to-moderate AD patients.

AD patients had lower SES than MCI patients (p < 0.001, r = 0.232). Lower SES was associated with a greater age at initial time of diagnosis (χ2 = 11.5, p = 0.001). In MCI patients, higher SES individuals outperformed lower SES individuals on the BNA after correcting for the effect of age (p = 0.004). Lower SES was also associated with decreased use of cognitive enhancers in AD patients (p < 0.001, r = 0.842).

Individuals with lower SES come into memory clinic later when the disease has progressed to dementia, while higher SES individuals present earlier when the disease is still in its MCI stage. There were more higher SES individuals who presented to our memory clinic. Higher SES is associated with better cognitive functioning and increased use of cognitive enhancers. The health policy implication is that we need to better engage economically disadvantaged individuals, perhaps at the primary care level.

This study was a part of a larger study exploring the impact of assistive technology on the lives of young people living with dementia (YPD). This paper focuses on one of the most useful devices, the simple remote control (SRC). The objective was to explore the reason why the SRC is significant and beneficial in the everyday lives of YPD and their caregivers.

This qualitative longitudinal study had a participatory design. Eight participants received an SRC. The range for using it was 0–15 months. In-depth interviews and observations were conducted at baseline and repeated every third month up to 18 months. A situated learning approach was used in the analysis to provide a deeper understanding of the significance and use of SRC.

Young people having dementia spend a substantial amount of time alone. Watching television was reported to be important, but handling remote controls was challenging and created a variety of problems. YPD learned to use SRC, which made important differences in the everyday lives of all family members. Comprehensive support from caregivers and professionals was important for YPD in the learning process.

The SRC was deemed a success because it solved challenges regarding the use of television in everyday lives of families. The design was recognizable and user-friendly, thus allowing YPD to learn its operation. Access to professional support and advice regarding assistive technology is vital for establishing a system for follow-up and continued collaboration to make future adaptations and adjustments.

The Computer-Administered Neuropsychological Screen for Mild Cognitive Impairment (CANS-MCI) is a computer-based cognitive screening instrument that involves automated administration and scoring and immediate analyses of test sessions. The objective of this study was to translate and culturally adapt the Brazilian Portuguese version of the CANS-MCI (CANS-MCI-BR) and to evaluate its reliability and validity for the diagnostic screening of MCI and dementia due to Alzheimer's disease.

The test was administered to 97 older adults (mean age 73.41 ± 5.27 years) with at least four years of formal education (mean education 12.23 ± 4.48 years). Participants were classified into three diagnostic groups according to global cognitive status (normal controls, n = 41; MCI, n = 35; AD, n = 21) based on clinical data and formal neuropsychological assessments.

The results indicated high internal consistency (Cronbach's α = 0.77) in the total sample. Three-month test-retest reliability correlations were significant and robust (0.875; p < 0.001). A moderate level of concurrent validity was attained relative to the screening test for MCI (MoCA test, r = 0.76, p < 0.001). Confirmatory factor analysis supported the three-factor model of the original test, i.e., memory, language/spatial fluency, and executive function/mental control. Goodness of fit indicators were strong (Bentler Comparative Fit Index = 0.96, Root Mean Square Error of Approximation = 0.09). Receiver operating characteristic curve analyses suggested high sensitivity and specificity (81% and 73% respectively) to screen for possible MCI cases.

The CANS-MCI-BR maintains adequate psychometric characteristics that render it suitable to identify elderly adults with probable cognitive impairment to whom a more extensive evaluation by formal neuropsychological tests may be required.

The Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE) were compared with and without the addition of a brief processing speed test, the symbol digit modalities test (SDMT), for vascular cognitive impairment (VCI) screening at three to six months after stroke.

Patients with ischemic stroke and transient ischemic attack were assessed with MoCA and MMSE, as well as a formal neuropsychological battery three to six months after stroke. VCI was defined by impairment in any cognitive domain on neuropsychological testing. The area under the receiver operating characteristic curve (AUC) was used to compare test discriminatory ability.

One hundred and eighty-nine patients out of 327 (58%) had VCI, of whom 180 (95%) had vascular mild cognitive impairment (VaMCI), and nine (5%) had dementia. The overall AUCs of the MoCA and MMSE scores and performance at their respective cut-off points were equivalent in detecting VCI (AUCs: 0.87 (95% CI 0.83–0.91) vs. 0.84 (95% CI 0.80–0.88), p = 0.13; cut-offs: MoCA (≤23) vs. MMSE (≤26), sensitivity: 0.78 vs. 0.71; specificity: 0.80 vs. 0.82; positive predictive value: 0.84 vs. 0.84; negative predictive value: 0.72 vs. 0.67; and correctly classified 78.6% vs. 75.5%; p = 0.42). The AUCs of MMSE and MoCA were improved significantly by the SDMT (AUCs: MMSE+SDMT 0.90 (95% CI 0.87–0.93), p <0.001; MoCA+SDMT 0.91 (95% CI 0.88–0.94), p < 0.02).

The MoCA and MMSE are equivalent and moderately sensitive, and can be supplemented with the SDMT to improve their accuracy in VCI screening.

An association between metabolic syndrome (MetS) and disturbances in neurocognitive function has been identified in Caucasians but the nature and extent of impaired cognition in Asian MetS patients, who may be at greater risk of degenerative cognitive decline, remains unspecified.

A cross-sectional study was conducted at the National University Hospital of Singapore. Participants were recruited from a diabetes clinic at the National University Hospital. Fifty-three patients who met MetS criteria and 44 clinical controls were recruited. All participants were 55 years and above and community ambulant. Neurocognitive function was assessed using the Cambridge Neuropsychological Test Automated Battery (CANTAB). CANTAB performances between MetS and control groups were examined with analysis of variance (ANOVA) and the relative contributions of vascular risk, and intrademographic factors on CANTAB scores were dilineated with stepwise regression analyses.

Participants with MetS consistently performed significantly worse than controls across all CANTAB subtests. Education and Chinese race were found to be potential protective factors.

Executive and memory impairment is present in Asian patients with midlife MetS who may be particularly vulnerable to the detrimental impact of MetS in midlife.

This study examines the relationship of unmet dementia-related care needs of community-dwelling persons, and their caregivers (CGs), to measures of caregiver burden.

Cross-sectional baseline data were analyzed from participants in a dementia care coordination trial of community-residing persons with dementia (PWD) (n = 254) and their caregivers (n = 246). Participants were recruited from Northwest Baltimore, Maryland. The Zarit Burden Inventory (ZBI) was used to measure subjective caregiver burden. Objective burden was measured by estimating the total hours per week spent doing things for the PWD and/or how many hours CGs missed paid work in the prior month due to caregiving responsibilities. The Johns Hopkins Dementia Care Needs Assessment was used to identify unmet dementia-related care needs. Bivariate and multivariate linear regressions examined the relationship of unmet needs, demographic, clinical, or functional characteristics with caregiver burden measures.

In adjusted multivariable models, patient neuropsychiatric symptoms and caregiver unmet emotional needs explained 22% of the variance in ZBI scores. In adjusted multivariable models, caregiver need for respite, patient functional dependency, and caregiver unmet specialty medical needs explained 26% of the variance in the hours per week spent caregiving. PWD's level of functional dependency was the sole correlate of missed time at work, explaining 11% of the variance.

Addressing potentially modifiable unmet caregiver needs may reduce subjective and objective caregiver burden.

Considering the discussion on implementing routine dementia screening in Germany, the objective of the current study was to validate the German version of the Perceptions Regarding Investigational Screening for Memory in Primary Care (PRISM-PC) questionnaire and to determine the acceptance of Alzheimer's disease screening in elderly German adults.

The German version of the PRISM-PC was administered to a subsample of participants who attended the Berlin Aging Study II (n = 506). The questionnaire was validated by exploratory as well as confirmatory factor analysis.

Regarding acceptance of Alzheimer's disease screening (Section B) a single factor structure fitted best. In terms of attitudes regarding Alzheimer's disease (Section D), a hierarchical factor structure was modeled with the higher-order factor “Harms” covering the domains “Family Burden,” “Dependence,” “Emotional Suffering,” “Stigma,” and “Medical Care” on the one hand and the domain “Future Planning” on the other hand. Internal consistency of the different scales reached from α = 0.67 to α = 0.94. Overall, 71.2% of the participants indicated that they wanted to be screened for Alzheimer's disease on a regular basis.

This study suggests that acceptance can reliably be assessed with the section “Acceptance of Alzheimer's disease screenings” of the German PRISM-PC questionnaire. Furthermore, the majority of elderly German adults would like to be screened for Alzheimer's disease regularly, which might be an effective starting point in order to implement routine dementia screenings. As the sample is a convenience sample of (relatively) healthy older adults, generalizability of these results is limited.

In many countries, illiteracy rates among aged people are quite high. However, only few studies have specifically investigated the impact of illiteracy on depression.

Data for 1,890 elderly individuals (aged ≥65 years) were obtained from a nationwide dementia epidemiological study conducted in South Korea. Based on their reading ability, the participants were divided into three groups: totally illiterate, partially illiterate, and literate. The Korean version of the Geriatric Depression Scale – Short Form (SGDS-K) was used to detect depression (cut-off score = 8). Multivariate logistic regression analyses were used to assess the association between illiteracy and depression. To explore clinical features of depression in illiterate people, we performed logistic regression to calculate odds ratios of positive responses (or negative responses to reverse-coded items) for each SGDS-K item using literate individuals as the reference group.

Totally illiterate participants had 2.41 times the odds and partially illiterate individuals had 1.59 times the odds of being depressed compared with literate participants after controlling for other variables. Compared with literate individuals, illiterate elderly persons were at increased odds for responding negatively to the majority of SGDS-K items, including “having memory problems,” “others are better off than me,” and “feeling worthless” even after controlling for various demographic and clinical factors.

Illiteracy in elderly individuals was associated with a higher rate and increased severity of depression. Illiteracy negatively affected depression symptomatology, especially factors associated with self-esteem. Therefore, clinicians should carefully monitor for the presence of depression in illiterate elderly adults.

Late-life depression is a heterogeneous disorder, whereby cognitive impairments are often observed. This study examines which clinical characteristics and symptom dimensions of late-life depression are especially impacting on specific cognitive domains.

Cross-sectional data of 378 depressed and 132 non-depressed older adults between 60–93 years, from the Netherlands Study of Depression in Older adults (NESDO) were used. Depressed older adults were recruited from both inpatient and outpatient mental healthcare institutes and general practices, and diagnosed according to DSM-IV-TR criteria. Multivariable associations were examined with depression characteristics (severity, onset, comorbidity, psychotropic medication) and symptom dimensions as independent variables and cognitive domains (episodic memory, processing speed, interference control, working memory) as dependent variables.

Late-life depression was associated with poorer cognitive functioning. Within depressed participants, higher severity of psychopathology and having a first depressive episode was associated with poorer cognitive functioning. The use of tricyclic antidepressants, serotonergic and noradrenergic working antidepressants, and benzodiazepines was associated with worse cognitive functioning. Higher scores on the mood dimension were associated with poorer working memory and processing speed, whereas higher scores on a motivational and apathy dimension were associated with poorer episodic memory and processing speed.

Heterogeneity in late-life depression may lead to differences in cognitive functioning. Higher severity and having a first depressive episode was associated with worse cognitive performance. Additionally, different domains of cognitive functioning were associated with specific symptom dimensions. Our findings on the use of psychotropic medication suggest that close monitoring on cognitive side effects is needed.

Memory complaints in older adults may be a precursor of measurable cognitive decline. Causes for these complaints may vary across age groups. The goal of this study was to develop classification models for the early identification of persons at risk for memory complaints using a broad range of characteristics.

Two age groups were studied, 55–65 years old (N = 1,416.8) and 65–75 years old (N = 471) using data from the Longitudinal Aging Study Amsterdam. Participants reporting memory complaints at baseline were excluded. Data on predictors of memory complaints were collected at baseline and analyzed using logistic regression analyses. Multiple imputation was applied to handle the missing data; missing data due to mortality were not imputed.

In persons aged 55–65 years, 14.4% reported memory complaints after three years of follow-up. Persons using medication, who were former smokers and had insufficient/poor hearing, were at the highest risk of developing memory complaints, i.e. a predictive value of 33.3%. In persons 65–75 years old, the incidence of memory complaints was 22.5%. Persons with a low sense of mastery, who reported having pain, were at the highest risk of memory complaints resulting in a final predictive value of 56.9%. In the subsample of persons without a low sense of mastery who (almost) never visited organizations and had a low level of memory performance, 46.8% reported memory complaints at follow-up.

The classification models led to the identification of specific target groups at risk for memory complaints. Suggestions for person-tailored interventions may be based on these risk profiles.