Background
Considerable efforts have been made to increase quality of care and quality of life in nursing home residents (NHR) in the past decades [
1,
2], but physical and social inactivity still remain a large concern in modern nursing facilities. As research has shown, the majority of NHR spend their time inactively sitting or lying alone [
3,
4]. A measure that has been positively associated with physical activity as well as social participation in community-dwelling older adults [
5,
6] and NHR [
7] is life-space (LS). LS has been conceptualized as the spatial extension of an individual’s environment that he/she moves in during a specified time period [
5,
8], irrespective of the types of conducted activity, use of walking aids, or other assistance [
9]. In general, greater LS implies that an individual has more opportunity to visit personally meaningful places and to interact socially with others [
10]. To highlight the complexity of factors influencing LS in older adults, Webber et al. [
11] presented a theoretical framework in which LS mobility in old age is assumed to be influenced by cognitive, psychosocial, physical, socio-demographic, financial, and environmental dimensions. The model has found partial empirical support in previous research which has demonstrated associations between LS mobility and physical performance [
8,
12,
13], global cognitive functioning [
14], and psycho-social factors including depression [
12,
13,
15], concerns about falling [
16], and apathy [
13] in community-dwelling older persons.
Applying the LS framework in NHR—an institutionalized group of older persons with multiple impairments—poses the question whether the determinants identified in community-dwelling older subjects can also be verified for LS in NHR. Institutions such as nursing homes (NH) are expected to have pronounced characteristics that strongly determine life and behavior within them [
17]. Such characteristics are, for example, architectural features (e.g., special care units, meeting places); care routines; or institution-dependent organizational schedules including meal times and weekly recurring and highly standardized events [
18,
19]. That said, we assume that the framework of Webber et al. (2010) may largely be valid also for the NH setting, but needs additional qualification in that the institutional factors impacting on behavior may play a key role in determining LS.
LS has mostly been assessed in- and outside of the private home environment [
8,
9,
12,
20]. Self-report measurements as the
Life-Space Diary [
8], the
University of Alabama at Birmingham Study of Aging Life-Space Assessment [
12], or the
Life-Space Questionnaire [
9] were predominantly used, providing a composite score of LS across a defined time period. Regarding NHR, only one measure has been introduced to our knowledge, a proxy-rating titled
Nursing Home Life-Space Diameter (NHLSD) [
19]. Such subjective LS assessments—self- or proxy-ratings—come with multiple weaknesses, e.g., recall/response biases, especially in cognitively impaired subjects [
21‐
23]. Moreover, they are unable to identify changes or events with temporal precision and intra-individual specificity [
24]. Also, predictors of LS may operate differently in more global questionnaire data as compared to high resolution data. This is why increasingly objective, technical LS-related assessments [
13,
20,
25] (e.g., Global Positioning System, infrared motion sensors, or Bluetooth transmitters) are used which also provide a high spatiotemporal resolution of LS not achievable by questionnaire-based assessment. However, such an advanced assessment strategy has so far not been applied in the nursing home setting. Specifically, a continuous, real-time assessment with high spatiotemporal resolution and minimal intrusion of the daily activities of individuals allows a more accurate picture of LS dynamics in daily ecologies and thus higher ecological validity. It may also become important as an endpoint in intervention research or serve diagnostic purposes by adding information to clinical status assessments.
This study provides a new sensor-based LS assessment in an institutional setting, including automated, high-resolution, spatiotemporal recording of residents’ habitual movement behavior within the resident facility across daytime. The aim of this study is threefold: (1) We provide a highly accurate picture of NHR’ LS and movement behavior, not achievable by previously used assessments. (2) Based on the theoretical LS framework of Webber et al. (2010), we examine whether its LS determinants are applicable to the nursing home setting and allow to develop a model explaining variance in NHR’ LS. (3) Given the unique characteristics of institutions described above, we hypothesize that, in addition to the factors described by Webber et al., a large proportion of the variance in NHR’ LS will be attributable to institutional routines, i.e., scheduled mealtimes.
Discussion
To the best of our knowledge, this is the first study to explore LS in NHR based on objective, sensor-based assessment with a high spatiotemporal resolution. Key findings of the current study were that (1) LS was very limited in NHR; (2) factors belonging to dimensions included in the framework by Webber et al. (2010) are also applicable to the NH setting; (3) when included in the models, the strongest association was found between overall LS variables and institutionally scheduled routines such as mealtimes.
Following the methodological paths of key studies in the field of LS research [
8,
9,
12,
19], we took a different practical approach by using a sensor-based system to obtain a comprehensive and objective picture of LS in NHR. Unlike subjective assessments used in previous studies, our objective assessment approach is not limited to generating a composite LS score, but also provides data on the chronological order in which LS areas where visited and for how long. This allows investigating LS far more extensively than before, including aspects of daily movement behavior in a sample of highly vulnerable NHR with high prevalence of advanced motor and cognitive impairment.
Despite the rather tight corset of the daily structure in institutions like nursing homes, data analysis revealed a wide spectrum of LS, ranging from residents who permanently stayed in their private room to those who were permanently absent from their room during daytime (see Table
1). Due to the lack of LS-related research in the nursing home setting, there are no results available for comparison with our findings on the duration a subject spent at a certain room or the frequency in which s/he changed zones on a daily basis in NHR. Results from studies using the NHLSD [
7,
19] are hardly comparable as this measure is conceptualized as composite LS score regarding the past 2 weeks. When compared to independent-living seniors, NHR perform considerably less Transits (6.9 per day vs. 10.8 room changes per hour) and spent muss less time out of the house (0.3 h per day vs. 4.0 h per day) [
20]. It has to be taken into account that in these two groups, the different LS zones have a different connotation, e.g., TAFR for a NHR still means staying indoors, whereas time away from home for an independent-living subject means leaving the building.
Our data shows that the main part of NHR’ daily life unfolded on the living units. Only very few individuals left their unit or facility and thus were not engaged in any activity beyond the facility at all. This is in line with Goffman (1961), who pointed out that institutions such as nursing homes are characterized by a “barrier to social intercourse with the outside” (p. 4). Some NHR may be worried by the thought of entering a less controlled, rather unknown and unsafe area beyond their unit. Others may feel drawn out of isolation in their room towards more eventful places. As a result, most NHR mainly stay in the public areas of their living units—a behavior which may also be attributable to motor and cognitive impairment of NHR.
Results from linear regression analyses confirm the LS-related dimensions identified by Webber et al. (2010) in their framework as well as findings in previous studies [
8,
12‐
16]. In line with previous studies that found male sex being associated with larger LS mobility [
40,
41], male sex was associated with more Transits in our sample.
Regarding motor performance, we found conflicting results on the association between LS and gait speed. Whereas more TAFR was associated with higher gait speed, more Transits were associated with lower gait speed. Thus, residents with better walking abilities change LS areas less frequently but stay in zones 2 to 4 for longer periods of time. One explanation may be that NHR with inferior walking abilities and functional capacity need to take rest periods in their private rooms more frequently than those with better physical function. We see this contrasting association of functional performance with both LS parameters as an indicator of different underlying concepts of both parameters requiring further investigation.
Our finding that lower cognitive status and less depressive symptoms were associated with TAFR finds support in results on community-dwelling subjects regarding the time out of home [
20,
42]. Lower cognitive status was associated with more TAFR and less Transits in our sample. Cognitively impaired subjects probably feel drawn to public areas due to certain aspects of these areas that draw attention (e.g., noise or conversation) [
43]. Due to diminished wayfinding abilities (i.e., not finding their own private room and staying where they presently are instead), or a high prevalence of apathy, they are often bound to stay in such public areas or other already determined locations. Cognitive performance may also reflect staffs’ reaction to these symptoms of dementia, that is, to keep residents in sight in the public area, and thus a larger amount of time is spent in public areas away from the own room, and less Transits are performed [
44]. However, beta weights were not significant for MMSE scores in the Transits regression model. The same applies to beta weights for apathy, with lower apathy being associated with more Transits, as previously reported in community-dwelling subjects [
13]. To explore this insignificance, we examined these linear regression results more closely. We found that AES-D scores were significantly correlated with MMSE scores (Pearson’s
r = −.558;
p < .001). When eliminating one of both measures from the regression model, the other factor became significant (AES-D: standardized beta = −.280;
p = .005; MMSE: standardized beta = .253;
p = .045), indicating that AES-D and MMSE have a considerable proportion of shared variance explanation. This is not surprising as apathy is a key symptom of dementia [
45].
As expected, the strong association between institutional factors and LS became very clear. When controlling for the variance of Transits and TAFR during institutionally scheduled mealtimes in separate models, it explained 67% (Transits) and 80% (TAFR) of the variance of Transits and TAFR during overall daytime, even though it only stood for one third of the overall measurement time. In the presence of these control variables, all other predictors included in the final models became insignificant, which demonstrates the high association between the variance of LS parameters during mealtimes and overall variance, adding valuable information regarding the structure of the Webber et al. framework when applied in the nursing home setting. That is, when comparing institutionally scheduled mealtimes with unscheduled daytime, considerable differences in LS parameters were found. During institutionally scheduled mealtimes, there were twice as many Transits per hour and TAFR was more than 20% higher. This has several implications: First, institutionally scheduled time is a rather “active” time, as it requires the majority of otherwise rather sedentary residents to move (or be moved) to the dining area and to be around others in a social context. Compared to this active time, NHR actually tend to be less active when they can freely decide what to do, e.g., participate in optional social group activities. Second, it implies a restriction of LS in terms of its range, as NHR have to be inside and within the living units during these institutionally scheduled mealtimes if they want to be served their meal—unless they are invited and picked up for a meal by friends or relatives.
Several limitations of the study have to be noted. Although the sensor-based assessment provides an objective documentation of LS, this technological approach comes with some technical limitations, especially regarding gaps in data transmission. Due to the systems localization frequency of one per 30 s, Transits within this time frame could have been missed if more than one had occurred. However, due to the low gait speed and motor function of our sample, more than one Transit within 30 s is a rather unlikely event. As a relatively high number of participants had to be excluded from analyses, the study sample was limited and thus potentially underpowered for certain research questions. Some of the independent variables are based on self-report measures, which may have been affected by recall and response bias due to cognitive impairment or other factors such as depressive symptoms. The study design was intentionally inclusive, also including a minor group of persons being unable to move independently. However, we see LS as an objective reality, irrespective of its active or passive occurrence.
Some valuable practical implications arose from our study. With a view to the associated factors found in this study, and by identifying the individual movement patterns of each resident during the day, our assessment approach may also be suitable for documenting deterioration in motor function and development of depression or behavioral symptoms related to dementia such as apathy (manifesting as ‘never leaving the private room’) or wandering/restlessness (manifesting as ‘moving around constantly’). The fact that NHR are particularly inactive between meals shows the good occasion in the daily schedule for implementing physical activity and LS enhancing interventions. These should be focused on associated factors that are susceptible to intervention (especially gait, apathy, and depressive symptoms), and be carried out on the living units in order to be within reach of the majority of residents who do not go beyond their living unit. Overall, the sensor-based LS assessment is a good example of how new assessment strategies may provide new and more comprehensive insights into the movement behavior of NHR. As it is still undergoing further development, the sensor-based LS assessment promises to capture more complex parameters that may be derived from raw data in the future, e.g., distance travelled within the facility as a measure of physical activity. Our approach may also help in identifying architectural and environmental characteristics of NHs such as dangerous, fall-provoking spots or important meeting places, and unfrequented, deserted areas in the facility, allowing enrichment of the environment and further stimulation of NHR’ social participation in daily life.