Out-of-Home Behavior and Cognitive Impairment in Older Adults

Abstract

This study explores differences in the out-of-home behavior of community-dwelling older adults with different cognitive impairment. Three levels of complexity of out-of-home behavior were distinguished: (a) mostly automatized walking behavior (low complexity), (b) global out-of-home mobility (medium complexity), and (c) defined units of concrete out-of-home activities, particularly cognitively demanding activities (high complexity). A sample of 257 older adults aged 59 to 91 years (M = 72.9 years, SD = 6.4 years) included 35 persons with early-stage Alzheimer’s disease (AD), 76 persons with mild cognitive impairment (MCI), and 146 cognitively healthy persons (CH). Mobility data were gathered by using a GPS tracking device as well as by questionnaire. Predicting cognitive impairment status by out-of-home behavior and a range of confounders by means of multinomial logistic regression revealed that only cognitively demanding activities showed at least a marginally significant difference between MCI and CH and were highly significant between AD and CH.

Keywords

out-of-home behavior Alzheimer’s disease mild cognitive impairment old age GPS tracking technology

Research shows that out-of-home behavior is critical for quality of life, independence, and well-being in old age (Mollenkopf, Hieber, & Wahl, 2011; Mollenkopf et al., 2004; Schaie, Wahl, Mollenkopf, & Oswald, 2003). We use in this article the term out-of-home behavior to address the full range of activities of moving from one location (particularly the home) to another (i.e., all forms of being mobile out of home) as well as activities that are conducted outside the home (Webber, Porter, & Menec, 2010). It is a clinically well-established fact that cognitive impairment is accompanied by changes of out-of-home mobility. In particular, cognitive impairments, such as some of the early manifestations of dementia (particularly, Alzheimer’s disease), but also types of MCI (Petersen et al., 2001; Petersen & Morris, 2003) include problems with out-of-home orientation, way finding, and spatial navigation (Hort et al., 2007; Tippett et al., 2009), which may result in a significant reduction of the use of out-of-home space. Problems with out-of-home mobility are also the single most frequently mentioned challenges of caring for people with dementia in the community; they are also among the difficult-to-manage behaviors (Silverstein, Flaherty, & Tobin, 2002). Change in out-of-home mobility may indeed be an early sign of mild cognitive impairment. The primary reason for this assumption is that out-of-home mobility is complex and resource intensive. For example, it demands continued cognitive attention, sensory motor coordination, and adjustment, due to issues such as changing weather conditions or unexpected occurrences in the physical environment (James, Boyle, Buchman, Barnes, & Bennett, 2011). Seen from a person–environment perspective, particularly the competence-press model of Lawton and Nahemow (1973; see also Wahl & Gitlin, 2007), the likelihood of a lack of fit between person competencies and environmental demands can be expected to substantially increase with the occurrence of cognitive impairment and this may cause unfavorable outcomes such as falls, getting lost, or even a transition to long-term care.

The relationship between mobility as a major indicator of out-of-home behavior and cognitive impairment has drawn considerable research attention in the past. First, experimental research has demonstrated a robust (negative) relationship between cognitive impairment—particularly, impaired executive function—and indicators of mobility (e.g., gait disturbances), which are more disturbed in those with impaired executive control (Ble et al., 2005; Holtzer, Verghese, Xue, & Lipton, 2006). Second, research related to the out-of-home behavior of older adults with dementia has frequently been conducted with residents of institutional settings (Miskelly, 2005; Silverstein et al., 2002), while much less is known about the out-of-home behavior of community-dwelling older adults. There is, however, some evidence pointing to reduced out-of-home life space and activity range in community-dwelling older adults with more severe cognitive impairment (Crowe et al., 2008; Shoval et al., 2011). In terms of methodology, most research has assessed out-of-home behavior, or rather single aspects of out-of-home behavior, such as out-of-home mobility and activities that are conducted out of home, via a questionnaire approach.

Several shortcomings emerge from these streams and preferred methods of previous research. In most of the studies, only single aspects or indicators of out-of-home behavior were assessed, failing to emphasize its multidimensionality. However, an integrative and multidimensional understanding of out-of-home behavior needs multiple indicators. It may indeed be the case that different components of out-of-home behavior reveal different strengths of relationship with different cognitive status, and this can only be detected by using a multidimensional approach. Going further, assessing out-of-home behavior via self-reports or in highly controlled experimental designs—as has been done in many cases in previous work—has several disadvantages: Self-reports may be biased particularly regarding highly detailed aspects of out-of-home behavior (e.g., length of walking tracks), and mobility and cognition dynamics assessed in a laboratory context might only partially reflect the dynamics of out-of-home behavior in natural settings. For example, out-of-home behavior in natural settings is embedded in a full range of compensatory means, such as social partners’ impact on behavior, the use of highly familiar environments, and reliance on decade-long routines of walking.

In light of these limitations of the previous research, our work builds on prior empirical findings as well as conceptual reasoning supporting the view of out-of-home behavior as a multidimensional and multifaceted phenomenon that should be analyzed at various levels of complexity (Metz, 2000; Patla & Shumway-Cook, 1999; Webber et al., 2010). In particular, the distinction of different complexity levels allows us, unlike most approaches in previous research, to test for differential relationships between levels of out-of-home behavioral complexity and various degrees of cognitive status (Figure 1). First, at a rather low level of complexity, out-of-home walking indicators, such as walking duration and distance, number of walking trips per day, and walking speed reflect highly automatized out-of-home behavior developed over the life course; this behavior is typically executed in familiar settings in a rather narrow action range around one’s home. Second, global out-of-home mobility indicators, such as total time spent out of home and number of places visited, include the full range of out-of-home behaviors of all kinds of cognitive complexity, but in a rather general and unspecified manner. Therefore, we locate global out-of-home mobility indicators at the medium level of complexity of out-of-home behavior. A third level of complexity uses more content-circumscribed and higher-order units of behavior operating at a higher level of complexity. Here, we will use the term out-of-home activity in contrast to out-of-home mobility to indicate such out-of-home behavior. Going further, the distinction between physically demanding and cognitively demanding out-of-home activities, as used in the previous research on activities in the time budget and other research arenas (Horgas, Wilms, & Baltes, 1998; Karp et al., 2006; Wilson & Bennett, 2003), may be helpful regarding activities. In particular, cognitively demanding out-of-home activities, such as going to a bank for financial purposes or educational activities, such as visiting a library, were expected to be high in complexity; physically demanding out-of-home activities, such as doing gardening, were expected to be less complex. Concerning interdependencies between the derived out-of-home behavior dimensions (out-of-home mobility and out-of-home activity), a certain overlap is plausible as the exertion of a specific out-of-home activity such as shopping requires mobility performances. Therefore, a hierarchy of out-of-home behavior dimensions can be assumed, with out-of-home mobility on a lower hierarchical level than out-of-home activity.

Figure 1.

Conceptual framework of out-of-home-behavior.

Research Aims and Hypotheses

The goal of this study is to analyze the relationship between out-of-home behavior and cognitive impairment status, referring to different levels of complexity of out-of-home behavior and a multimethods approach (GPS tracking technology and questionnaire data). In regard to the lowest level of complexity of out-of-home behavior, reflected in out-of-home walking indicators, we expect no difference between cognitively healthy older adults (CH) and those with mild cognitive impairment (MCI) and Alzheimer’s disease (AD). With respect to the medium level of cognitive complexity of out-of-home behavior, expressed in indicators of global out-of-home mobility indicators, we expect differences between AD and both MCI and CH, whereas no differences between MCI and CH are expected. With regard to the highest level of complexity of out-of-home behavior, reflected in cognitively demanding out-of-home activities, we expect differences not only between AD and MCI and CH but also between MCI and CH. We will also examine whether such predictions will hold after controlling for covariates such as age, gender, education, household constellation, and physical functioning, all of which have been found to be related with out-of-home behavior (Murata, Kondo, Tamakoshi, Yatsuya, & Toyoshima, 2006; Shumway-Cook et al., 2007; Webber et al., 2010).

Method

Project Design, Study Samples, and Recruitment Strategy

Data were gathered within the project “The Use of Advanced Tracking Technologies for the Analysis of Mobility in Alzheimer’s Disease and Related Cognitive Diseases” (“Senior Tracking”; SenTra), an interdisciplinary study of German and Israeli psychologists, psychiatrists, geographers, and social workers (Oswald et al., 2010; Shoval et al., 2008). Potential participants with AD and MCI were identified in Germany, through the memory clinics of the Department of General Psychiatry, Heidelberg University, Heidelberg (Germany), and the Central Institute of Mental Health, Mannheim (Germany), and in Israel, through the psychogeriatric center of the Tel Aviv Sourasky Medical Center. The participants were part of the normal work-up of outpatients with cognitive disorders at either institution. A comprehensive medical, neuropsychological, and neuropsychiatric assessment, including the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) standardized procedure for the evaluation and diagnosis of patients with cognitive impairments (Morris, Heyman, Mohs, & Hughes, 1989; German version by Thalmann et al., 2000) in Germany and the Camdex-R (Roth, Huppert, Tym, & Mountjoy, 1998) in Israel, was carried out by a multidisciplinary team (including experienced [neuro-]psychologists, neurologists, and geriatric psychiatrists) in the local memory clinics. CH individuals were drawn in Germany at random from official local public registers. In Israel, CH individuals were identified via senior centers and other organizations because public registers are not readily available. CH individuals underwent the same assessment procedure as the other participants.

Participants with cognitive impairment fulfilled either the diagnostic criteria for AD or MCI (Levy, 1994; Winblad et al., 2004). Because of our intention to use questionnaires as well as the use of GPS technology across a period of 4 weeks (see below), we only included individuals with early-stage AD, based on a global deterioration scale (GDS; Reisberg, Ferris, de Leon, & Crook, 1982) score of 4 or a Clinical Dementia Rating (CDR) score of 0.5 or 1.0. Exclusion criteria were other types of dementia (e.g., vascular, frontotemporal), other severe psychiatric disorders (e.g., major depression, schizophrenia, or severe personality disorders), substance abuse, severe motor disturbances (e.g., caused by Parkinson’s disease), sensory deficits potentially affecting mobility, severe somatic illness (e.g., cancer), and use of prescription drugs that could potentially affect cognition and functioning (e.g., neuroleptics). Inclusion criteria for the CH included no subjective cognitive complaints, no activities of daily living (ADL) impairments, and performance within 1 standard deviation—according to norm scores—in all cognitive domains. We also strived to enlist participants with a caregiver or significant other available in the household because we were also interested in their experience regarding the out-of-home behavior of the target person (we are not using this information in the current paper).

All participants were informed about the project and the assessment procedure by means of individual invitation letters, followed by personal telephone calls. If respondents agreed to participate, they were enrolled in the study after informed consent, following the ethical guidelines and procedures for formal ethical consent. In particular, expected ethical considerations regarding the study’s use of GPS/GIS technology (e.g., fear of being observed, intrusion into privacy) were discussed during the informed consent. Major reasons to refuse participation in the study were distrust, fear, lack of interest and/or time, general health problems, and occurrences with significant others (e.g., recent death of partner). We generally recruited a to some extent positively selected group of CH, MCI, and AD participants regarding outcome variables such as health. Written informed consent was received from all participants, and the Ethics Board Review of the Universities of Heidelberg and the Helsinki Committee (IRB) at the Tel Aviv Sourasky Medical Center approved the study.

Because we expected difficulties in recruiting participants and their family members who would commit themselves for 4 weeks of GPS tracking as well as interviewing and cognitive testing, our sampling goal was to gather as large as possible groups in the 2-year period available for the recruitment process. Specifically, we strived to enroll at least 30 individuals in the AD group. On the other hand, we purposefully planned to collect at least 100 CH older adults for additional analysis with a sole focus on those without cognitive impairment (not reported in this article). In total, data were obtained from 35 AD participants (15 from Israel), 76 MCI participants (39 from Israel), and 146 CH participants (46 from Israel). In all three groups, the majority (66 to 75%) reported the availability of a car in their household, with no significant group difference in car availability. Nearly all study participants (84%) were retired (or currently not working). Table 1 provides a description of the sample. As can be seen in the Table 1, no differences were observed between the three groups regarding age, gender, number of persons in household composition (statistical tendency that individuals with MCI and AD more frequently lived with another person), perceived health, physical functioning, and income satisfaction. However, the MCI and AD groups revealed a significantly lower level of education. Education is considered as one of several influences on the cognitive reserve (Snowdon, 2003; Stern, 2003), which may contribute to prevention or delayed onset of cognitive impairment. This may be an explanation why most of the highly educated persons belonged to the CH group. Groups also varied by MMSE (Folstein, Folstein, & McHugh, 1975) and two common tests of executive functioning, Trail Making Tests A and B (Reitan, 1958; Spreen & Strauss, 1991).

Table 1.

Group Comparison (CH, MCI, and AD) in Sociodemographic and Cognitive Measures.

Variable N	Total, n = 257	CH, n = 146	MCI, n = 76	AD, n = 35	Statistical test
Sociodemographic measures
Age (M, SD, range)	72.9, 6.4, 59-91	72.5, 6.1, 61-91	72.9, 6.5, 59-88	74.1, 7.1, 59-86	ns
Gender (%) male	51.0	50.0	48.7	60.0	ns
Education (years; M, SD, Range)	13.6, 4.2, 2-26	14.5, 4.2, 2-26	12.3, 4.2, 4-23	12.5, 3.2, 9.5-21	*** (CH > MCI, AD)
Number of persons in household	1.8, 0.7, 1-5	1.7, 0.6, 1-5	1.9, 0.7, 1-5	1.9, 0.6, 1-3	*
Perceived health^a	2.8, 0.8, 1-5	2.8, 0.7, 1-5	3.0, 0.9, 1-5	2.8, 1.0, 1-5	ns
Physical functioning^b	80.5, 19-7, 0-100	82.6, 16.9, 30-100	77.2, 21.8, 0-100	78.7, 24.8, 15-100	ns
Income satisfaction	7.1,2.2, 0-10	7.3,2.0, 2-10	6.8,2.3, 0-10	7.1,2.4, 0-10	ns
Cognitive measures
MMSE^c	27.5, 2.3, 19-30	28.6, 1.3, 26-30	27.0, 2.1, 21-30	24.1, 2.4, 19-30	*** (CH > MCI > AD)
Trail-Making Test A (sec)^c	60.8, 34.5, 19-300	51.2, 21.8, 19-129	66.5, 33.9, 25-171	91.1, 56.4, 36-300	*** (CH<MCI<AD)
Trail-Making Test B (sec)^c	141.6, 80.1, 26-416	121.7, 69.2, 26-370	160.5, 88.3, 38-416	200.2, 75.5, 68-300	** (CH<MCI, AD)

Note: CH = cognitively healthy control persons; MCI = persons with mild cognitive impairment; AD = persons with early-stage Alzheimer‘s disease. Statistical test for differences: Analysis of variance (ANOVA) F test (means) and chi square test (frequencies).

Higher scores indicate lower perceived health (1 = excellent, 2 = very good, 3 = good, 4 = sufficient, 5 = bad).

Physical functioning was assessed by the SF-36 (Bullinger & Kirchberger, 1998), with higher scores indicating better physical functioning (range = 0-100).

MMSE = Mini-Mental State Examination, with higher scores indicating better performance; in Trail-Making Test A and B, higher scores indicate more time needed for completion and thus worse performance (all based on the CERAD procedure; see further explanation in text).

ns = not significant. *

p < .05. **p < .01. ***p < .001

Measurement of Out-of-Home Behavior

Measurement of out-of-home walking and global out-of-home mobility indicators

This part of the assessment was done via a GPS data collection approach. Participants received a GPS tracking kit and instructions concerning its use. The kit consists of a GPS receiver with a Global System for Mobile communications (GSM) modem and a monitoring unit located in the home that let researchers know whenever the tracked person leaves home (Murakami & Wagner, 1999; Shoval et al., 2010, 2008; Shoval & Isaacson, 2006). The GPS device had a size of about 4 (length) X 2 (breadth) X 1 (height) inches and a weight of about 1 pound. Participants could choose how to carry the GPS unit, for example, in a belly pouch, in a shoulder bag, or in any other way that is convenient to the participant. The participant took the unit with him or her at all times for a period of up to 4 weeks. The GPS was programmed to obtain locations every 5 seconds when the tracked person is outside the home. The data collected in Israel and in Germany were sent by General Packet Radio Service (GPRS) protocol to the project server at the Hebrew University of Jerusalem (Shoval et al., 2011).

In terms of validity of the tracking data, interviewers carried out weekly phone conversations with the participants during the 4 weeks to inquire about possible difficulties in using the GPS kit. Missing data may result from various sources, such as problems with the mobile phone connection due to underserved areas, connection problems occurring in the data transport from Germany to Israel, or simply participants forgetting the device itself or forgetting to charge it. Therefore, a validity classification was used for periods of 24 hr, and only days that do not have less than 1 hr of missing data were considered as “valid days” for full time-space analysis. In addition, we only used tracking data on days with out-of-home behaviors; days that were completely spent at home are excluded from the following analyses. Applying these validity criteria, the mean number of valid days in our sample was 20.5 (SD = 5.9); that is, on average, 70% of the days within the tracking period of our study participants were considered as valid. In eight cases only, the number of valid days was fewer than 10. There were no differences between the three groups regarding the mean number of valid days.

According to the components of out-of-home mobility that we conceptually distinguished, we included the following GPS tracking data. Regarding out-of-home walking indicators, we refer to three GPS-based indicators, that is, walking distance, walking duration, and walking speed. Walking tracks were identified based on a speed criterion, with all tracks with a speed less than 5 km/hr considered as walking tracks, a criterion that proved useful in previous GPS tracking research (Shoval et al., 2010). Because only a minority in our sample reported using bicycles regularly, we did not include an indicator of cycling tracks in our analyses. “Driving tracks” were also not included because it could not be determined based on GPS technology if the study participant was the driver and because tracks by car and tracks by public transportation are not easily distinguishable. In terms of global out-of home mobility indicators, we refer to the time spent out of home and the number of visited nodes (places) per day. Nodes were defined as places respondents stayed for at least 5 min; that is, nodes are an indication of (the number of) visited places, such as supermarkets, physicians, or the apartment of a relative.

Measurement of out-of-home activities

Participants filled out an activity list in which 23 out-of-home activities were included. They gave information about which of the activities listed they participated in (Yes/No) and how frequently they engaged in them (frequency was coded as follows: 1 = never, 2 = less than once/month, 3 = once/month, 4 = more than once/month, 5 = once/week, 6 = more than once/week, 7 = daily). To identify the (most) cognitively and physically demanding activities within the activity list, an expert rating was initiated. Ten experts from different academic disciplines (predominantly psychology, gerontology, and geropsychiatry), with much scientific and practical knowledge on older adults, evaluated the cognitive demands of every single activity, using a Likert-type scale ranging from 0 (not demanding) to 10 (very demanding). All activities with a mean cognitive demand rating above the total average (M = 6.10) and with a small interrater deviation (SD < 1.6) were categorized as cognitively demanding out-of-home activities. These activities were volunteering, conducting businesses (e.g., bank, post office), visiting a library, accompanying someone in need of support (e.g., grandchildren), and engaging in educational activities (e.g., course in continuing education). In a similar way, activities with a mean physical demand rating above the total average (M = 6.42) and with an interrater deviation SD < 1.22 were classified as physically demanding out-of-home activities (e.g., shopping, gardening, sports activities). Activities that were rated as both physically and cognitively demanding were excluded from both activity classes to avoid an overlap between the two activity categories. The resulting ratings tended to be highly consistent (for cognitively demanding activities, Cronbach’s alpha was .84; for physically demanding activities, Cronbach’s alpha .89).

Covariates

Because of the rather small sample, we only considered a minimum of covariates, that is, age, gender, household constellation, education, and physical functioning. Education was assessed as duration of school and professional education (in years). The Physical Functioning subscale of the SF-36 (Bullinger & Kirchberger, 1998) is a highly reliable and valid measure assessing with 10 items the extent to which health problems impair everyday (physical) activities (like walking more than 1 mile) and self-care. Possible answers on each item are “yes, strongly impaired”; “yes, slightly impaired”; and “no, not impaired.” The sum score ranges from 0 to 100, with higher values indicating better physical functioning. Finally, we also controlled for country (Germany, Israel) to counterbalance differences in recruitment as well as for differences in terms of geographical and cultural context.

Data Analysis Strategy

We used parametric analyses to test for group differences (analysis of variance [ANOVA] F test, followed by pair-wise differences based on a Tukey test procedure). Multinomial logistic regression analysis was applied to examine which variables were best able to predict group membership (CH, MCI, AD), while simultaneously controlling for potential confounders. Statistical programs used for the analyses were SAS 9.2 and PASW 18.

Results

Group Contrasts Across Levels of Complexity of Out-of-Home Behavior

Table 2 provides information on the out-of-home behavioral indicators compared across the three groups. As expected, no differences between groups appeared regarding the low level of complexity of out-of-home behaviors, that is, the out-of-home walking indicators. However, at the medium level of complexity—that is, global out-of-home mobility indicators—the mean levels of AD were statistically below the means of both other groups. Finally, as also expected, statistically meaningful differences were observed between groups at the high level of complexity regarding out-of-home behaviors (i.e., physically and cognitively demanding out-of-home activities), with statistically meaningful differences between AD and both other groups. There was also a difference between CH and MCI regarding cognitively demanding out-of-home activities, favoring the CH group. Effect sizes were medium to large (Cohen, 1977), with the strongest effect appearing in the number of exerted cognitively demanding out-of-home activities (η² = .16).

Table 2.

Differences Between Groups (CH, MCI, and AD) in Out-of-Home Behavior Indicators.

Variable, N	Total, n = 257	CH, n = 146	MCI, n = 76	AD, n = 35	Statistical test	Effect size
Out-of-home walking indicators
Walking distance (km)	0.7, 0.6, 0.1-3.7	0.7, 0.5, 0.1-2.7	0.7, 0.6, 0.1-3.2	0.8, 0.7, 0.1-3.7	ns	Partial η² = .00
Walking duration (h)	0.2, 0.2, 0.0-3.2	0.2, 0.3, 0.0-3.2	0.2, 0.2, 0.0-0.8	0.2, 0.2, 0.0-1.0	ns	Partial η² = .01
Walking speed (km/h)	3.7, 0.6, 1.6-5.7	3.8, 0.7, 1.6-5.7	3.7, 0.6, 2.4-5.1	3.7, 0.7, 1.7-4.9	ns	Partial η² = .00
Global out-of-home mobility indicators
Time out-of-home (hours)	4.0, 2.3, 0.0-12.8	4.2, 2.3, 0.7-12.8	4.3, 2.3, 0.3-9.9	2.7, 1.5, 0.0-7.3	** (CH, MCI > AD)	Partial η² = .05
Number of nodes visited	4.5, 1.4, 1.1-9.4	4.7, 1.3, 1.9-9.4	4.6, 1.4, 1.4-8.2	3.6, 1.2, 1.1-6.6	*** (CH, MCI > AD)	Partial η² = .08
Out-of-home activities
Number of physically demanding activities	2.1, 0.8, 0-3	2.1, 0.7, 0-3	2.1, 0.7, 1-3	1.7, 0.8, 0-3	*(CH, MCI > AD)	Partial η² = .03
Number of cognitively demanding activities	1.8, 1.0, 0-5	2.1, 1.0, 0-5	1.7, 1.0, 0-4	0.9, 0.7, 0-2	*** (CH > MCI > AD)	Partial η² = .16
Frequency of physically demanding activities^a	4.2, 1.3, 1.0-7.0	4.3, 1.3, 1.0-7.0	4.3, 1.3, 1.3-6.7	3.6, 1.2, 1.0-6.5	** (CH, MCI > AD)	Partial η² = .04
Frequency of cognitively demanding activities^a	2.2, 0.8, 1.0-5.5	2.4, 0.8, 1.0-5.5	2.2, 0.8, 1.0-4.6	1.7, 0.6, 1.0-2.8	*** (CH, MCI > AD)	Partial η² = .10

1 = never, 2 = less than once/month, 3 = once/month, 4 = more than once/month, 5 = once/week, 6 = more than once/week, 7 = daily.

ns=not significant. *

p < .05. **p < .01. ***p < .001.

Simultaneous Examination of the Relationship Between Cognitive Impairment Status and Various Indicators of Out-of-Home Behavior in the Context of Potentially Confounding Variables

Table 3 depicts the results of the multinomial logistic regression analyses, with group membership seen as the outcome variable. To avoid multicollinearity, we reduced the set of out-of-home behavior predictors by excluding walking duration, which was highly correlated with walking distance; time out-of-home, because of its high correlation with the number of visited nodes; and frequency of out-of-home activities, which was highly associated with the number of out-of-home activities (all r > .70). As can be seen in Table 3, regarding the contrast between CH and MCI, the number of exerted cognitively demanding out-of-home activities was the only indicator of out-of-home behavior that significantly predicted group membership at the .10 p level. More cognitively demanding out-of-home activities were associated with a higher likelihood of belonging to the CH group. Moreover, a higher probability of belonging to the MCI group was related to living in Germany, lower education level, and more persons in household.

Table 3.

Multinomial Logistic Regression Model: Predicting Cognitive Impairment Status by Out-of-Home Behavior Indicators.

	Predictors (odds ratio) [95% confidence interval]
CH versus MCI
Country	2.94** [1.32, 6.56]
Age	1.01 [0.95, 1.08]
Education	0.89** [0.82, 0.97]
Physical functioning	0.99 [0.97, 1.01]
Income satisfaction	0.90 [0.78, 1.04]
Number of persons in household	1.67(*) [0.97, 2.88]
Walking distance	1.43 [0.73, 2.80]
Walking speed	1.02 [0.55, 1.89]
Mean number of nodes	1.16 [0.89, 1.53]
Number of cognitively demanding activities	0.73 (*) [0.51, 1.05]
Number of physically demanding activities	1.29 [0.78, 2.16]
CH versus AD
Country	1.11* [0.29, 4.09]
Age	1.02 [0.93, 1.12]
Education	0.98* [0.87, 1.11]
Physical functioning	1.01 [0.98, 1.03]
Income satisfaction	0.99 [0.80, 1.23]
Number of persons in household	1.68 [0.78, 3.59]
Walking distance	1.63 [0.70, 3.79]
Walking speed	1.02 [0.42, 2.46]
Mean number of nodes	0.52** [0.31, 0.88]
Number of cognitively demanding activities	0.27*** [0.14, 0.51]
Number of physically demanding activities	0.92 [0.48, 1.77]
R² Cox & Snell	.30
R² Nagelkerke	.35

Note: N = 257. CH = cognitively healthy control persons; MCI = persons with mild cognitive impairment; AD = persons with early-stage Alzheimer’s disease.

(*)

p < .10. * p < .05. **p < .01. ***p < .001.

In respect to the contrast between CH and AD, we found a significant effect of cognitively demanding out-of-home activities, with a higher chance of belonging to the CH group when exerting more cognitively demanding activities. Furthermore, a higher number of visited nodes out-of-home was related with a higher likelihood of CH group membership. In addition, country (due to sample composition; see above) and a higher education level were also shown to be significant in contributing to CH membership.

Pseudo R² values ranged from .30 to .35, depending on different calculation modalities indicating an appropriate fit of the multinomial logistic regression model.

Discussion

Although out-of-home behavior of older adults is closely linked with quality of life and likely associated with cognitive status, research in this area has been limited. Specifically, previous research has focused frequently on older adults in long-term care settings, tended to neglect adults with mild cognitive impairment, relied heavily on experimental methodology with the consequence of lowered external validity, and has mostly used a rather narrow operationalization of out-of-home behavior, solely based on questionnaire data. In the current study, we aimed to fill this gap in the literature by considering various forms of cognitive impairment in community-dwelling older adults. In addition, we have argued for the analysis of out-of-home behavior and cognitive status in natural settings based on a multidimensional understanding of out-of-home behavior, various levels of complexity, as well as a multimethods approach, including GPS and questionnaire assessment.

Although links between cognitive status and out-of-home behavior have consistently been found in the previous literature (Ble et al., 2005; Crowe et al., 2008; Holtzer et al., 2006; James et al., 2011), we found that differentiation is needed. Consistent with our hypotheses, the CH, MCI, and AD did not significantly differ in out-of-home walking indicators; these indicators reflect highly routinized out-of-home behaviors of a rather low complexity level, typically happening in familiar settings near to one’s home. In some contrast to our findings, however, previous studies have reported substantial relationships between walking speed and cognitive abilities or cognitive status (Ble et al., 2005; Holtzer et al., 2006). One reason for these incongruent findings could be differences in the operationalization of walking speed. In this study, walking speed was not assessed in an experimental setting under controlled context conditions. Therefore, the nonsignificant difference in walking speed between the three groups could be due to a differential choice of walking environments by the groups (see Prohaska et al., 2009), with cognitively impaired persons possibly preferring less complex and more familiar environments where maintaining a normal walking speed was still possible. Global out-of-home mobility indicators with a medium complexity (time spent out of home, number of nodes visited), as well as highly complex behaviors—assessed by physically but particularly cognitively demanding out-of-home activities—were significantly reduced in AD, as compared to both other groups. The number of exerted cognitively demanding out-of-home activities was also lower in MCI, as compared to CH participants. Controlling for potentially confounding variables, such as age, sex, education, number of persons in the household, and physical functioning, did not change the picture that cognitively demanding out-of-home activities best discriminated among the three groups out of all out-of-home indicators used. However, this was less the case for the number of visited out-of-home nodes, a GPS indicator of global out-of-home mobility, and definitely was not the case for out-of-home walking indicators. Results thus rather consistently underscore that older adults with AD engage in less out-of-home behavior of high and medium complexity as compared to MCI and CH, and the only, at least marginally significant difference, between MCI and CH appeared in terms of cognitively demanding out-of-home activities. Specifically, older persons with AD seem to visit fewer places in their environment, possibly avoiding complex and cognitively demanding locations and preferring familiar, less demanding places.

Explanations of the observed differences in out-of-home behavior between AD and the remaining groups are tempted to solely refer to the role of lowered cognitive performance. However, as seen from a wider perspective of human adaptation in case of constraining developmental conditions, older adults with early-stage AD, as included in the present study—in cooperation with their caregivers—may reduce their out-of-home behavior due to an awareness of lowered cognitive resources; thus, they may attempt to avoid risky and potentially dangerous situations. On the other hand, highly routinized behaviors, such as out-of-home walking patterns in safe and familiar settings, are still feasible and executed in a similar magnitude, as in MCI and CH. This would reflect an efficient strategy in terms of compensation via simplification and reduction of out-of-home behavior (Dixon & Bäckman, 1995), which may preserve quality of life for a considerable period of time. Indeed, referring to a classic distinction in lifespan psychology (Baltes, Lindenberger, & Staudinger, 2006), some pronounced loss in terms of a lowered life space would lead to the gain of maintaining out-of-home autonomy to some extent. However, the effectiveness of such compensation strategies may be reduced as cognitive impairment progresses. Specifically, older adults with early-stage AD may still have some awareness of their impaired cognitive capacity, which supports their ability to use adaptational strategies. This sort of awareness may be absent in advanced stages of AD and may lead to out-of-home behaviors of high safety risk, such as wandering and severe spatial disorientation (Colombo et al., 2001), which can result in severe injury or even death due to falls or traffic accidents (Klein et al., 1999; Koester & Stooksbury, 1995). Overall, the increasing lack of fit between lowered cognitive competencies and high demands of the external environment coming along with dementia-related disorders (Lawton & Nahemow, 1973; Wahl & Gitlin, 2007) may lead to increased restriction of life space in these individuals and may indeed also heighten their risk of social isolation.

In contrast, the sole significant difference between CH and MCI with regards to out-of-home behavior was in cognitively demanding out-of-home activities, which MCI participants conducted to a lower extent than cognitively healthy participants. In contrast to those with AD, it seems that cognitive resources in those with MCI are still sufficient to maintain previous patterns of out-of-home behavior. However, this out-of-home competence seems to find its limits in highly cognitively demanding activities, such as banking, accompanying and supervising another person (e.g., a grandchild), or doing educational activities. Indeed, it may be the case that early signals for MCI in the everyday context may be seen in the reduction or relinquishing of cognitively demanding activities; this may even eventually, as future research may show, provide earlier indications of MCI than assessed cognitive performance.

It should be mentioned that we have also checked for differences in intraindividual variability of out-of-home behavior based on our 4-week GPS data, because such variability, for example, in the area of cognitive performance, has been found to predict potential for long-term cognitive trajectories (Hultsch, MacDonald, Hunter, Levy-Bencheton, & Strauss, 2000). However, in our data, the analysis of intraindividual variability, based on a selection of out-of-home mobility indicators, revealed only minor group differences regarding mobility fluctuation measures. Moreover, these fluctuation measures were not relevant for the prediction of cognitive impairment status in a respective multinomial logistic regression model.

In terms of practical implications of our results, we see these at different levels. First, the finding that older adults with AD engage in less out-of-home behavior is an important insight for city planners and stakeholders at large; public places, spaces, and buildings may undergo optimization—such as including orientation tools, reducing environmental hazards, and alerting professionals (such as police, cab drivers, professionals of public transport)—to accommodate the needs and problems of older adults with dementia. That is, there is a still little planning for creating “age-friendly” livable community settings based on a clear idea of the needs of those with cognitive impairment, not only in terms of objective barriers but also when it comes to social participation and “normal” navigation through one’s environments as a senior citizen. Technology such as GPS technology, used in our study only as a research tool, may also become a major practical aid here in the future (Topo, 2009). The societal relevance of such strategies is obvious, given that a high number of people in the community are currently living with cognitive impairment, and even more people are expected to do so in the future. Thus, there is a pressing need to move toward what some have coined the “open city” of the future, with more investment in safety and leisure activity options in public areas that better accommodate the needs of those with various cognitive impairments (Blackman et al., 2003). Second, because studies have shown that caregivers of people with dementia often regard the problems with out-of-home behavior as a major burden and stressor (Colombo et al., 2001), advanced tracking technologies may help caregivers to locate (via the Internet) lost cognitively impaired older adults in the quickest time possible; therefore, they may serve as an efficient emergency- and harm-preventing device (e.g., Brooks & Scarfo, 2009). However, the ethical issues associated with the use of such equipment deserve serious attention, and additional findings from SenTra (Landau et al., 2009) support the view that the perspectives of family caregivers may be different from professional caregivers. In particular, the possibility and risk of “Big Brother” scenarios that accompany lost privacy and “total control” generally seems to be more emphasized by professional caregivers. Third, at the measurement level, an important implication of our findings is that activity and everyday competence-oriented assessment scales able to differentiate between MCI and cognitively healthy adults should put major emphasis on a differentiated assessment of highly cognitively demanding out-of-home activities. Such insights may lead to a new generation of Activities of Daily Living–like scales, able to add to the early prediction and differentiation of MCI and thus serve early intervention.

The results of the current research need to be considered relative to a number of limitations. Our research was based on rather limited sample sizes, particularly regarding MCI and AD groups, limiting the statistical power of our analyses. Furthermore, the cross-sectional nature of our data prevents any causal conclusions. For example, although we have implied that cognitive impairment affects out-of-home behavior, it seems to also be the case that certain out-of-home behaviors, such as those connected with social participation, impact the course of cognitive impairment (Lövdén, Ghisletta, & Lindenberger, 2005). Indeed, some studies have demonstrated that out-of-home behavior is protective against cognitive decline and dementia (e.g., Wilson & Bennett, 2003) or is able to predict cognitive impairment at a later point in time (James et al., 2011; Prohaska et al., 2009). Moreover, we could not assess for every single conducted out-of-home activity of each person whether the activity had been exerted alone or accompanied, which probably makes a difference in terms of the amount of cognitive demands posed by the activity. However, we did take into account possible influences of social partners by controlling for the number of household members when predicting group membership based on cognitive predictors. Furthermore, cultural and geographical differences between Germany and Israel may also have been underrated in our study—in merging both samples—although we have controlled for country in our multivariate analyses. Finally, caution is necessary to generalize our findings to those living alone because the majority of our participants lived together with another person.

There is, nevertheless, reason to assume that we have addressed in our findings rather fundamental relationships, which may be largely independent from cultural and geographic characteristics. Finally, our measurement of out-of-home behavioral variables, including the GPS tracking, may be still too rough (exactness of 4-5 meters) to detect differences in out-of-home behavior among CH, MCI, and AD, both at the mean and intraindividual variability level.

Despite these limitations, the current study contributes to the understanding of out-of-home behavior and cognitive impairment. In particular, our findings lend support to the view that research regarding out-of-home behavior of older adults with various degrees of cognitive performance can gain from the use of a multimethod assessment of out-of-home behavior, as well as from an interdisciplinary approach including psychological, geographical, and psychiatric expertise.

Footnotes

Acknowledgements

The Study “The Use of Advanced Tracking Technologies for the Analysis of Mobility in Alzheimer’s Disease and Related Cognitive Diseases” (abbreviated as Senior Tracking/SenTra) was supported by the German Research Foundation between 2008 and 2011, based on a grant to Hans-Werner Wahl (WA809/11-1). We would like to thank our geography partners in Germany—in particular, Dr. Tim Freytag—for their valuable support regarding the GPS/GIS data collection. Heike Hercher has provided the expert rating for distinguishing cognitively demanding activities in her Master’s thesis in Psychology. Additionally, we would like to thank Katharina Hager, Heike Hercher, Hannah Schmidt-Friderichs, and Johanna Martinez-Slebi for great support in collecting and processing the data of the project. We are also very thankful to the older adults who have participated in the study.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Markus Wettstein is post-doc and a psychologist with interest in cognitive and sensory aging, and out-of-home mobility in later life.

Hans-Werner Wahl is professor and a psychologist with interest in person–environment interrelations in later life, sensory aging, and the awareness of age-related change.

Noam Shoval is professor and a geographer with interest in the relationship between mobility and cognitive impairment in old age.

Frank Oswald is professor and a psychologist with interest in person–environment interrelations in later life including out-of-home mobility.

Elke Voss is a research associate with interest in out-of-home mobility and cognitive impairment as well as depression in old age.

Ulrich Seidl is a psychiatrist with interest in cognitive impairment and mobility in later life as well as autobiographical memory and apathy.

Lutz Frölich is professor and a psychiatrist with interest in dementia, mild cognitive impairment, and mobility in later life, including pharmacological and behavioral interventions.

Gail Auslander is professor and a social worker with interest in caregiving for demented older adults.

Jeremia Heinik is a psychiatrist with interest in dementia, mild cognitive impairment, and mobility in later life, including diagnostic and intervention issues.

Ruth Landau is professor and social worker with interest in ethical issues related to the use of technology in demented older adults.

References

Baltes

P. B.

Lindenberger

Staudinger

U. M.

(2006). Life-span theory in developmental psychology. In Damon

Lerner

R. M.

(Eds.), Handbook of child psychology. Vol. 1: Theoretical models of human development (6th ed., pp. 569-664). New York, NY: John Wiley.

Blackman

Mitchell

Burton

Jenks

Parsons

Raman

Williams

(2003). The accessibility of public spaces for people with dementia: A new priority for the “open city.” Disability & Society, 18(3), 357-371.

Ble

Volpato

Zuliani

Guralnik

J. M.

Bandinelli

Lauretani

. (2005). Executive function correlates with walking speed in older persons: The InCHIANTI Study. Journal of the American Geriatrics Society, 53(3), 410-415.

Brooks

K. R.

Scarfo

(2009). Geographic information systems: Health and aging as a spatial construction. In Abbott

P. S.

Carman

Scarfo

(Eds.), Re-creating neighborhoods for successful aging (pp. 33-52). Baltimore, MD: Health Professional Press.

Bullinger

Kirchberger

(1998). SF-36 Fragebogen zum Gesundheitszustand—Handanweisung [SF-36 questionnaire on health—manual]. Göttingen, Germany: Hogrefe.

Cohen

(1977). Statistical power analysis for the behavioral sciences. New York, NY: Academic Press.

Colombo

Vitali

Cairati

Perelli-Cippo

Bessi

Gioia

Guaita

(2001). Wanderers: features, findings, issues. Archives of Gerontology and Geriatrics, 33(Suppl. 1), 99-106.

Crowe

Andel

Wadley

V. G.

Okonkwo

O. C.

Sawyer

Allman

R. M.

(2008). Life-space and cognitive decline in a community-based sample of African American and Caucasian older adults. Journal of Gerontology: Medical Sciences, 63(11), 1241-1245.

Dixon

R. A.

Bäckman

(Eds.). (1995). Compensating for psychological deficits and declines. London: Routledge.

10.

Folstein

M. F.

Folstein

S. E.

McHugh

P. R.

(1975). Mini-Mental State: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), 189-198.

11.

Holtzer

Verghese

Xue

Lipton

R. B.

(2006). Cognitive processes related to gait velocity: Results from the Einstein aging study. Neuropsychology, 20(2), 215-223.

12.

Horgas

A. L.

Wilms

H.-U.

Baltes

M. M.

(1998). Daily life in very old age: Everyday activities as expression of successful living. The Gerontologist, 38(5), 556-568.

13.

Hort

Laczó

Vyhnálek

Bojar

Bureš

Vlček

(2007). Spatial navigation deficit in amnestic mild cognitive impairment. Proceedings of the National Academy of Sciences, 104(10), 4042-4047. doi:10.1073/pnas.0611314104

14.

Hultsch

D. F.

MacDonald

S. W. S.

Hunter

M. A.

Levy-Bencheton

Strauss

(2000). Intraindividual variability in cognitive performance in older adults: Comparison of adults with mild dementia, adults with arthritis, and healthy adults. Neuropsychology, 14, 588-598.

15.

James

B. D.

Boyle

P. A.

Buchman

A. S.

Barnes

L. L.

Bennett

D. A.

(2011). Life space and risk of Alzheimer Disease, mild cognitive impairment, and cognitive decline in old age. American Journal of Geriatric Psychiatry. doi:10.1097/JGP.1090b1013e318211c318219

16.

Klein

D. A.

Steinberg

Galik

Steele

Sheppard

J.-M.

Warren

. (1999). Wandering behaviour in community-residing persons with dementia. International Journal of Geriatric Psychiatry, 14(4), 272-279.

17.

Koester

R. J.

Stooksbury

D. E.

(1995). Behavioral profile of possible Alzheimer’s disease patients in Virginia search and rescue incidents. Wilderness & Environmental Medicine, 6(1), 34-43.

18.

Landau

Werner

Auslander

G. K.

Shoval

Heinik

(2009). Attitudes of family and professional caregivers towards the use of GPS for tracking patients with dementia: an exploratory study. The British Journal of Social Work, 39, 670-692.

19.

Lawton

M. P.

Nahemow

(1973). Ecology and the aging process. In Eisdorfer

Lawton

M. P.

(Eds.), Psychology of adult development and aging (pp. 619-674). Washington. DC: American Psychological Association.

20.

Levy

(1994). Aging-associated cognitive decline. Working Party of the International Psychogeriatric Association in collaboration with the World Health Organization. International Psychogeriatrics, 6, 63-68.

21.

Lövdén

Ghisletta

Lindenberger

(2005). Social participation attenuates decline in perceptual speed in old and very old age. Psychology and Aging, 20, 423-434.

22.

Metz

D. H.

(2000). Mobility of older people and their quality of life. Transport Policy, 7(2), 149-152.

23.

Miskelly

(2005). Electronic tracking of patients with dementia and wandering using mobile phone technology. Age and Ageing, 34(5), 497-499.

24.

Mollenkopf

Hieber

Wahl

H.-W.

(2011). Continuity and change in older adults’ perceptions of out-of-home mobility over ten years: A qualitative-quantitative approach. Ageing & Society, 31, 782-802.

25.

Mollenkopf

Marcellini

Ruoppila

Széman

Tacken

Wahl

H. W.

(2004). Social and behavioural science perspectives on out-of-home mobility in later life: Findings from the European project MOBILATE. European Journal of Ageing, 1(1), 45-53.

26.

Morris

J. C.

Heyman

Mohs

R. C.

Hughes

J. P.

(1989). The consortium to establish a registry for Alzheimer’s disease (CERAD): I. Clinical and neuropsychological assessment of Alzheimer’s disease. Neurology, 39(9), 1159-1165.

27.

Murakami

Wagner

D. P.

(1999). Can using global positioning system (GPS) improve trip reporting? Transportation Research Part C: Emerging Technologies, 7(2-3), 149-165.

28.

Murata

Kondo

Tamakoshi

Yatsuya

Toyoshima

(2006). Factors associated with life space among community-living rural elders in Japan. Public Health Nursing, 23(4), 324-331.

29.

Oswald

Wahl

H.-W.

Voss

Schilling

Freytag

Auslander

G. K.

. (2010). The use of tracking technologies for the analysis of outdoor mobility in the face of dementia: First steps into a project and some illustrative findings from Germany. Journal of Housing for the Elderly, 24, 55-73.

30.

Patla

Shumway-Cook

(1999). Dimensions of mobility: Defining the complexity and difficulty associated with community mobility. Journal of Aging and Physical Activity, 7(1), 7-19.

31.

Petersen

R. C.

Doody

Kurz

Mohs

R. C.

Morris

J. C.

Rabins

P. V.

. (2001). Current concepts in mild cognitive impairment. Archives of Neurology, 58(12), 1985-1992.

32.

Petersen

R. C.

Morris

J. C.

(2003). Clinical features. In Petersen

R. C.

(Ed.), Mild cognitive impairment: Aging to Alzheimer’s disease (pp. 15-40). New York, NY: Oxford University Press.

33.

Prohaska

T. R.

Eisenstein

A. R.

Satariano

W. A.

Hunter

Bayles

C. M.

Kurtovich

. (2009). Walking and the preservation of cognitive function in older populations. The Gerontologist, 49(Suppl. 1), 86-93.

34.

Reisberg

Ferris

S. H.

de Leon

M. J.

Crook

(1982). The Global Deterioration Scale for assessment of primary degenerative dementia. The American Journal of Psychiatry, 139, 1136-1139.

35.

Reitan

R. M.

(1958). Validity of the Trail Making Test as an indicator of organic brain damage. Perceptual and Motor Skills, 8, 271-276.

36.

Roth

Huppert

F. A.

Tym

Mountjoy

C. Q.

(1998). Camdex-R. The Cambridge Examination for Mental Disorders of the Elderly—Revised. Cambridge, UK: Cambridge University Press.

37.

Schaie

K.-W.

Wahl

H.-W.

Mollenkopf

Oswald

(Eds.). (2003). Aging independently: Living arrangements and mobility. New York, NY: Springer.

38.

Shoval

Auslander

Cohen-Shalom

Isaacson

Landau

Heinik

(2010). What can we learn about the mobility of the elderly in the GPS era? Journal of Transport Geography, 18, 603-612.

39.

Shoval

Auslander

G. K.

Freytag

Landau

Oswald

Seidl

. (2008). The use of advanced tracking technologies for the analysis of mobility in Alzheimer’s disease and related cognitive diseases. BMC Geriatrics, 8, 7.

40.

Shoval

Isaacson

(2006). Application of tracking technologies to the study of pedestrian spatial behavior. The Professional Geographer, 58, 172-183.

41.

Shoval

Wahl

H.-W.

Auslander

G. K.

Isaacson

Oswald

Edry

. (2011). Use of GPS technology for measuring timing and distance of out-of-home mobility of older adults with different levels of cognitive functioning: First findings of the SenTra Project. Ageing & Society, 31, 849-869.

42.

Shumway-Cook

Guralnik

J. M.

Phillips

C. L.

Coppin

A. K.

Ciol

M. A.

Bandinelli

. (2007). Age-associated declines in complex walking task performance: The Walking InCHIANTI Toolkit. Journal of the American Geriatrics Society, 55(1), 58-65.

43.

Silverstein

N. M.

Flaherty

Tobin

T. S.

(2002). Dementia and wandering behavior: Concern for the lost elder. New York, NY: Springer.

44.

Snowdon

D. A.

(2003). Healthy aging and dementia: Findings from the Nun Study. Annals of Internal Medicine, 139, 450-454.

45.

Spreen

Strauss

(1991). A compendium of neuropsychological tests administration, norms, and commentary (3rd ed.). New York, NY: Oxford University Press.

46.

Stern

(2003). The concept of cognitive reserve: A catalyst for research. Journal of Clinical and Experimental Neuropsychology, 25(5), 589-593.

47.

Thalmann

Monsch

Schneitter

Bernasconi

Aebi

Camachova-Davet

. (2000). The CERAD neuropsychological assessment battery (CERAD-NAB)—A minimal data set as a common tool for German-speaking Europe. Neurobiology of Aging, 21(Suppl. 1), 30.

48.

Tippett

W. J.

Lee

J.-H.

Zakzanis

K. K.

Black

S. E.

Mraz

Graham

S. J.

(2009). Visually navigating a virtual world with real-world impairments: A study of visually and spatially guided performance in individuals with mild cognitive impairments. Journal of Clinical and Experimental Neuropsychology, 31(4), 447-454. doi:10.1080/13803390802251360

49.

Topo

(2009). Technology studies to meet the needs of people with dementia and their caregivers. A literature review. Journal of Applied Gerontology, 28, 5-37.

50.

Wahl

H.-W.

Gitlin

L. N.

(2007). Environmental gerontology. In Birren

J. E.

(Ed.), Encyclopedia of gerontology: Age, aging, and the aged (2nd ed., pp. 494-501). Oxford, UK: Elsevier.

51.

Webber

S. C.

Porter

M. M.

Menec

V. H.

(2010). Mobility in older adults: A comprehensive framework. The Gerontologist, 50(4), 443-450.

52.

Wilson

R. S.

Bennett

D. A.

(2003). Cognitive activity and risk of Alzheimer’s disease. Current Directions in Psychological Science, 12, 87-91.

53.

Winblad

Palmer

Kivipelto

Jelic

Fratiglioni

Wahlund

L. O.

. (2004). Mild cognitive impairment—beyond controversies, towards a consensus: Report of the International Working Group on Mild Cognitive Impairment. Journal of Internal Medicine, 256, 240-246.