Determinants of out-of-home activity in older adults with cognitive impairment following inpatient rehabilitation: A secondary cross-sectional analysis of a randomised controlled trial using location-based and accelerometer-derived measures

Abstract

Objective

To examine physical activity and its spatial context after geriatric rehabilitation and to identify determinants of out-of-home activity in older adults with cognitive impairment.

Design

Cross-sectional observational study.

Setting

Community, following discharge from geriatric rehabilitation.

Participants

113 geriatric patients (mean age 82.1 ± 6.0 years; 76.1% female) with cognitive impairment following geriatric rehabilitation.

Main Measures

Physical activity was assessed over 48 h using a body-worn accelerometer (PamSys^®) and location-based tracking derived from the Global Positioning System (Qstarz BT-Q1000XT). Group differences between participants with and without out-of-home activity were analysed. Multivariable logistic regression identified predictors across physical, psycho-social, cognitive, environmental, financial, and personal domains.

Results

Only 36 (31.9%) of the 113 participants (82.1 ± 6.0 years; 76.1% female) left their homes. Those with out-of-home activity accumulated significantly more daily steps, walking episodes, active time, and less sedentary time than those staying in-home (all p < .01). Receiver operating characteristic analyses identified thresholds of 2140 steps/day (area under the curve 0.81) and 29.6 s mean walking episode duration (area under the curve 0.73) for discriminating between participants with and without out-of-home activity. Physical capacity (OR: 1.30, 95% CI: 1.02–1.65) and cognitive function (OR: 1.29, 95% CI: 1.02–1.62) were independent predictors of out-of-home activity.

Conclusion

Out-of-home activity is infrequent after rehabilitation in older adults with cognitive impairment and closely reflects overall physical activity. It may serve as a practical marker of mobility and highlights the need for targeted interventions to support community mobility during this vulnerable phase.

Keywords

Cognitive impairment motion sensors older adults physical activity rehabilitation

Introduction

Mobility in older age is a multifaceted construct, shaped by a dynamic interplay of personal, environmental, and contextual factors.¹ It encompasses not only physical activity ‒ defined as any bodily movement that results in energy expenditure ‒ but also spatial expansion, with reference to the concept of life-space.² In this context, out-of-home activity ‒ regularly leaving the home for various purposes ‒ is a facilitator for maintaining a physically and socially active lifestyle,³ and is also associated with substantial physical and mental health benefits. Older adults engaging more frequently in out-of-home activities are at lower risk for nursing home admission⁴ and mortality.⁵ Furthermore, studies have linked out-of-home activity with higher levels of physical activity across various populations,⁶ including older adults.⁷ Declining neurological, physical, musculoskeletal, metabolic, and sensory status as part of the ageing process profoundly impacts mobility behaviour.⁸ Out-of-home mobility frequently declines in older adults, (partly) leading to an increased focus on in-home activities.⁹ This decline can be attributed to the complexity of mobility and the intricate interplay of the different influencing factors.¹⁰ As out-of-home activity requires physical, psychological, and cognitive engagement, it is inherently more demanding than in-home activity.¹¹ Accordingly, cognitive impairment, which is highly prevalent in geriatric rehabilitation patients,¹² is associated with reduced out-of-home activity,^11,13 reduced life-space,¹⁴ and shorter average daily walking time,¹⁵ contributing to the ‘shrinking world phenomenon’ in older adults with cognitive impairment.¹⁶ However, it remains unclear whether the findings apply to older adults with cognitive impairment during the highly vulnerable transition phase following ward-based geriatric rehabilitation, a period marked by substantially increased risks of deconditioning, functional decline, and loss of independence.¹⁷ Importantly, it is unknown whether engaging in out-of-home activity after discharge in this vulnerable population increases physical activity levels or is offset by compensatory rest or inactivity at home as suggested for discussion earlier.¹⁸

The introduction of small, lightweight Global Navigation Satellite System devices, such as Global Positioning System trackers, has substantially improved the objective assessment of out-of-home activity by enabling continuous, non-invasive location tracking of real-world mobility.¹⁹ Combined use of accelerometer and Global Positioning System measures reduces recall bias and enhances data quality compared to self-report and has demonstrated feasibility in community-dwelling adults with cognitive impairment.²⁰ Despite increasing recognition of mobility metrics as valuable health biomarkers, studies objectively examining determinants of out-of-home activity among adults with cognitive impairment in the critical post-discharge phase remain scarce, although such activity may represent a key marker of overall mobility and health status. Therefore, the present study aims to analyse overall physical activity in relation to spatial location, and to identify determinants of out-of-home activity across personal/biographical, cognitive, physical, psycho-social, environmental and financial domains in adults with cognitive impairment following inpatient rehabilitation.

Methods

Ethical approval for the study was obtained from the Ethics Committee of the Medical Faculty at Heidelberg University (S-252/2015), and the trial was conducted in accordance with the Declaration of Helsinki. We performed a cross-sectional analysis of baseline data from a double-blinded, randomised, placebo-controlled trial designed to improve physical capacity and activity in community-dwelling older adults with cognitive impairment following geriatric rehabilitation (ISRCTN82378327).²¹

Participants were recruited consecutively from geriatric rehabilitation wards of a German geriatric hospital between September 2015 and April 2017. Patients with mild to moderate impairment were included in the study. Further inclusion criteria were: age ≥ 65 years; ability to walk at least 4 m without a walking aid; residence within 30 km of the study centre; community-dwelling status (i.e., discharge to the patients’ own home); no terminal disease; no delirium; no language restrictions; and written informed consent.

The data were collected in the participants’ home after discharge from rehabilitation (except for age, sex, and number of comorbidities that were documented from patient charts during inpatient rehabilitation). All interviews and testing procedures were consistently administered by assessors trained in interviewing and testing geriatric patients with cognitive impairment.

Descriptive and potentially associated variables with out-of-home activity were selected according to a comprehensive mobility framework,¹ covering physical, psycho-social, cognitive, environmental, financial, and personal/biographical domains, complemented by health characteristics relevant in clinical populations, as in the present study, life-space mobility and physical activity assessment.

The measures listed below represent the domains examined in this study. Detailed information and references for all instruments and assessment procedures are available in the published study protocol²¹ and the Supplementary Material (Table A).

The physical domain was represented by physical capacity as measured with the Short Physical Performance Battery, and data for habitual walking speed (metres/seconds) were extracted from the test battery.

The psycho-social domain was covered by the assessment of the social situation (Assessment of the Social Situation, depressive symptoms (Geriatric Depression Scale), apathetic symptoms (Apathy Evaluation Scale-Clinical Version), fall-related concerns (Short Falls Efficacy Scale-International), and related avoidance behaviour (Fear of Falling Avoidance Behavior Questionnaire).

The cognitive domain was analysed via the Montreal Cognitive Assessment scores.

As environmental factors weather data (temperature and precipitation) during the 48 h measurement period, recorded from the weather station closest to each participant's home, were included (Deutscher Wetterdienst, Climate Data Centre).

To represent the financial domain, a self-reported rating of ‘insufficient’ vs. ‘sufficient’ money available was considered (as part of the Assessment of the Social Situation).

Personal/biographical and health characteristics included age, sex, and number of comorbidities. Other sociocultural and health-related factors encompassed living situation (‘alone’ vs. ‘with others’), marital status (‘married’ vs. ‘not married’), educational level (‘< 8 years’ vs. ‘≥ 8 years’), as well as the self-rated health status (European Quality of Life 5 Dimensions ‒ Visual Analog Scale), falls in the previous year, documented following an established classification, and care related aspects (documented as ‘number of physician appointments’ and ‘duration of outpatient care’ in the past 3 months; based on a questionnaire for the use of medical and non-medical services in old age (‘Questionnaire for Health-Related Resource Use in an Elderly Population’).

The ‘Life-Space Assessment for persons with cognitive impairment’ was included as a questionnaire-based measure of comprehensive mobility. The tool provides a composite score and three subdomains, each accounting for different levels of support (by equipment or human).

Physical activity and Global Positioning System measures were performed over 48 h during weekdays only. To ensure consistency, the time-based results are presented as averages over a 24-h period. Participants were provided with basic information about the application and use of the equipment to enhance compliance and facilitate problem detection.

Physical activity was measured by a small, wearable motion sensor (PAMSys™, BioSensics, Cambridge, MA, USA) fixed at the participant's chest by use of adhesive tape. The sensors were calibrated by the manufacturer before being used in the study. The sensor was worn continuously, day and night. As no adhesive was provided to participants, removal or replacement would have been noticed by the assessors. The PAMSys ^TM was used to analyse locomotion outcomes (number of steps, walking episodes (at least three strides within 5 s), and duration (in minutes per day) of sedentary behaviour (including lying and sitting), standing, and walking and other active behaviour, mean walking duration per episode (seconds), and mean cadence (steps/minute)), based on established and validated algorithms for older adults.²²

Global Positioning System is a satellite-based global navigation system for tracking geolocation data. The Global Positioning System tracking device recorded location (i.e., longitude and latitude coordinates), speed, and heading direction with corresponding time and date stamps. The QStarz (QStarz BT1000X, Qstarz International Co., Ltd, Taipei, Taiwan) is a portable, stand-alone wireless device and able to measure the spatial location of a person.²³ Participants were asked to place the Global Positioning System sensor on their belts or pockets or to wear them as a pendant. Global Positioning System data were uploaded to the ‘Personal Activity Location Measurement System’ (available from: http://ucsd-palms-project.wikispaces.com/.), and saved in formats appropriate for Google Earth (Google, Inc., Mountain View, CA). Data of the Global Positioning System and physical activity sensor were analysed using the ‘Personal Activity Location Measurement System’, which has been validated for processing Global Positioning System and physical activity data.²⁴ The outcome variable used in the present study was staying (without out-of-home activity) versus leaving-home (with out-of-home activity) as dichotomous measure, verified by visual inspection of Global Positioning System screenshots.

Statistical analysis

Descriptive data included frequencies and percentages, medians and interquartile ranges, or means and standard deviations. Independent-samples t-tests, Mann‒Whitney U-tests, Wilcoxon-tests, and chi-square tests were used to analyse the group differences in descriptive characteristics between those with out-of-home activity and those with in-home activity only. To determine the optimal threshold of daily steps and mean walking duration distinguishing out-of-home from only in-home activity, receiver operating characteristic analyses were performed. The state variable was defined as being outdoors (1 = yes, 0 = no), and daily steps and mean walking duration per episode were entered as the test variables. The area under the curve with 95% confidence intervals was computed to evaluate the discriminative ability of steps to predict out-of-home activity. Optimal cut-off values were determined using Youden's index (J = sensitivity + specificity − 1), with higher values indicating better overall diagnostic performance. Logistic regression modelling was conducted to examine the independent predictors of out-of-home activity. Independent variables for model inclusion to represent all domains of the mobility framework¹ were selected based on statistically significant bivariate correlations (p < .05) using Spearman or point-biserial tests. Variables with non-normally distributed residuals were naturally log-transformed before regression analyses. To avoid multicollinearity, variables with intercorrelations > 0.70 were not included simultaneously. Where multiple variables reflected overlapping constructs or domains, the following prioritisation criteria were applied: (a) conceptual comprehensiveness or clinical relevance of the variable, and (b) the strength of the bivariate association with out-of-home activity. Odds ratios with corresponding 95% confidence intervals were reported to indicate the strength and direction of associations. Model fit was evaluated using Nagelkerke's R² to represent the proportion of explained variance. Statistical significance was set at p < .05. All analyses were conducted using IBM SPSS Statistics for Windows, Version 29.0.

Results

Out of 1981 patients screened for eligibility, 113 community-dwelling older adults (mean age = 82.1 ± 6.0 years) following geriatric rehabilitation with mild to moderate cognitive impairment (mean score 16.7 ± 2.8), physical impairment (mean score = 5.3 ± 2.3), and multimorbidity (mean number of diagnoses: 11.4 ± 4.4) were included. Daily step counts were low (median 2468; Inter Quartile Range 1115–4056), reflecting the limited mobility. About two-thirds reported a history of fall in the past year (67%). Most participants were women (76.1% female); 50% lived alone, 30% with a spouse, and 20% were supported by a carer permanently residing in their home. Further characteristics are presented in Table 1.

Table 1.

Participant characteristics.

Variables	n = 113
Age (years)	82.1 ± 6.0 (65–94)
Female (n)	86 (76.1)
Living alone (n)	78 (69.0)
Married (n)	35 (31)
Educational level, < 8 years (n)	36 (31.9)

Number of diagnose (n)	11.4 ± 4.4
VAS (score^a)	55.4 ± 16.1
Care at home, duration past 3 months (min)	120 [0–1350]
Number of physician appointments, past 3 months (n)	4 [2–7.5]
Persons with at least one fall in the previous year (n)	77 (68)

MOCA (score^a)	16.7 ± 2.8

SPPB (score^a)	5.3 ± 2.3
Walking speed (m/s)	0.46 ± 0.20

SOS (score^a)	19 [17–20]
GDS (score^b)	5.3 ± 3.0
AES-C (score^b)	22.0 ± 9.1
FES-I (score^b)	11 [9–14.5]
FFABQ (score^b)	18.4 ± 12.5

Persons with insufficient money available (%)	11 (9.7)

LSA-CI (score^a)	24.1 ± 13.3
Steps per day (n)	2468 [1115–4056]

Data are given as mean ± standard deviation, median [Inter Quartile Range], or number (%). MOCA = Montreal Cognitive Assessment, Scoring 0–30; SOS = Evaluation of the Social Situation, scores 0–25; financial status = subjective rating of insufficient vs. sufficient money available; VAS = Visual Analogue Scale, scores 0–100; GDS = Geriatric Depression Scale, scores 0–15; AES-C = Apathy Evaluation Scale Clinical Version, scores 0–54; FES-I = Falls Efficacy Scale – International, scores 7–28; FFABQ = Fear of falling avoidance behavior questionnaire, scores 0–84; LSA-CI = Life-Space Assessment for Persons with Cognitive Impairment, scores 0–90; SPPB = Short Physical Performance Battery, scores 0–12.

Score: higher = better.

Score lower = better.

Differences between participants with and without out-of-home activity

The majority of participants stayed in-home during the measurement period (n = 77; 68.1%) and only one third (n = 36; 31.9%) left their home at all and engaged in out-of-home activity.

Adults with out-of-home activity differed significantly from those who remained in-home across a range of variables (Table 2). These included not only several physical activity measures, but also almost all domains including the physical, psycho-social, cognitive, and personal/biographical domains, as well as health characteristics and life-space mobility.

Table 2.

Differences between persons with and without out-of-home activity.

Variables	Without out-of-home activity (n = 77)	With out-of-home activity (n = 36)	p-value
Age (years)^b	83.2 ± 6.0	80.0 ± 5.6	.008*
Female (n)^a	65 (84)	21 (58)	.002*
Living alone, (n)^a	59 (77)	19 (53)	.010*
Married (n)^a	18 (23)	17 (47)	.010*
Educational level, < 8 years (n)^a	28 (36)	8 (22)	.135
Number of diagnoses (n)^b	11.5 ± 4.6	11.0 ± 3.9	.545
VAS (score)^b	55.9 ± 15.5	54.3 ± 17.5	.640
Care at home, duration past 3 months (min)^c	450 [0–1800]	0 [0–652]	.015*
Number of physician appointments, past 3 months (n)^c	4 [2–6]	6 [3–9]	.060
Persons with at least one fall in the previous year (n%)^a	54 (70)	23 (64)	.521
MOCA (score) ^b	16.2 ± 2.8	17.7 ± 2.7	.008*
SPPB (score)^b	4.8 ± 2.2	6.4 ± 2.1	<.001**
Walking speed (m/s)^b	0.42 ± 0.19	0.56 ± 0.18	<.001**
SOS (score)^c	18 [17–20]	20 [18–21]	033*
GDS (score)^b	5.3 ± 3.1	5.3 ± 2.8	.953
AES-C (score)^b	23.1 ± 9.3	19.6 ± 8.2	.048*
FES-I (score)^c	12 [9–15]	10 [8–13]	.414
FFABQ (score)^b	20.4 ± 12.7	13.9 ± 10.9	.006*
Persons with insufficient money available (n)^a	7 (9)	4 (11)	.748
Life-Space Mobility (LSA-CI score)^b	19.8 ± 11.4	33.2 ± 12.7	<.001**
Maximal Life-Space	3.4 ± 1.3	4.3 ± 0.7	<.001**
Equipment-assisted Life-Space	2.1 ± 1.0	3.4 ± 1.1	<.001**
Independent Life-Space	0.9 ± 1.1	1.8 ± 1.8	.002*
Physical activity measures^b,c
Steps per day (n)	1758 [708–2926]	4240 [2715–5719]	<.001**
Episodes of walking per day (n)	101 [51–175]	197 [108–263]	.002*
Walking and other activities per day (min)	38 [18–63]	72 [54–101]	<.001**
Standing per day (min)	147 ± 66	206 ± 68	<.001**
Sedentary behaviour (lying and sitting) per day (min)	1250 ± 95	1152 ± 95	<.001**
Mean walking duration per episode (seconds)	16.8 [13.9–27.4]	24.5 [16.9–33.9]	.060
Mean cadence (steps/minute)	72.6 ± 11.1	74.6 ± 10.3	.357

Chi-squared-tests as appropriate.

Independent-samples t-tests.

Mann–Whitney U-tests.

*Significance at a 0.05 level, **Significance at a 0.01 level (marked bold).

Date were given as mean ± standard deviation, median [Inter Quartile Range], or number (%);MOCA = Montreal Cognitive Assessment, scores 0–30; SOS = Evaluation of the Social Situation, scores 0–25; VAS = Visual Analogue Scale, scores 0–100; GDS = Geriatric Depression Scale, scores 0–15; AES-C = Apathy Evaluation Scale Clinical Version, scores 0–54; FES-I = Falls Efficacy Scale – International, scores7–28; FFABQ = Fear of falling avoidance behavior questionnaire, scores 0–84; LSA-CI = Life-Space Assessment for Persons with Cognitive Impairment, scores 0–90; SPPB = Short Physical Performance Battery, scores 0–12.

As expected, adults with out-home activity showed higher levels of physical activity, covering more steps and walking episodes, longer durations of walking, other active behaviour, and standing, and less sedentary behaviour (lying or sitting) (p = .001–.002). In contrast, measures not directly related to activity volume, including mean walking duration per episode and mean cadence, did not differ between participants with and without out-of-home activity. Participants with out-of-home activity also reported a higher subjective mobility, as indicated by higher life-space mobility (mean score 33.2 ± 12.7 vs. 19.8 ± 11.8; p < .001; and subdomains: Maximal life-space (4.3 ± 0.7 vs. 3.4 ± 1.3; p < .001), equipment-assisted life-space (3.4 ± 1.1 vs. 2.1 ± 1.0; p < .001) and independent life-space (1.8 ± 1.8 vs. 0.9 ± 1.1; p = .002)).

Physical capacity in Short Physical Performance Battery score and walking speed (both p < .001), was greater in participants with out-of-home activity, with clinically meaningful between-group differences (Short Physical Performance Battery >1 point; walking speed >0.1 m/s).

Significant group differences emerged for most variables of the psycho-social domain. Participants with out-of-home activity reported a more favorable social situation, less fear of falling-related avoidance behaviour, and fewer apathetic symptoms (p = .002‒.029). No differences were observed for depressive symptoms or concerns about falling.

Cognitive function was significantly higher among those with out-of-home activity (p = .005).

Environmental variables (e.g., mean temperature and precipitation) and financial domain did not differ between the groups (not shown in the table; p = .132‒.497).

In the personal/biographical domain, participants with out-of-home activity were significantly younger, less often female, and more frequently married, while living alone was less common (p = .004‒.011). Indicators of health-related care suggested a better status in this group, including fewer physician appointments (p = .026) and less outpatient care (p = .005) in the past three months. No significant differences were found for educational level, multimorbidity, self-rated health status, or fall history.

A Receiver Operating Curve analysis identified the optimal cut-off for steps predicting out-of-home activity. The area under the curve was 0.81 (95% CI: 0.74–0.88), indicating good discrimination. The optimal threshold of 2140 steps/day yielded a sensitivity of 0.92 and a specificity of 0.38 (Supplemental Figure A). When evaluating walking pattern rather than volume, a second Receiver Operating Curve analysis identified an optimal mean walking duration cut-off of 29.6 s (sensitivity: 0.94, specificity: 0.58). The area under the curve was 0.73 (95% CI: 0.63–0.83), reflecting acceptable discrimination (Supplemental Figure B).

Determinants of out-of-home activity

Correlational Spearman and point-biserial analyses revealed many associated variables out of the domains outlined by Webber et al.¹ with out-of-home activity (Supplemental Table B), serving as a basis for selecting independent variables for the regression analysis. The subsequent logistic regression model revealed only higher cognitive function (OR: 95% 1.24; CI: 1.02–1.50) and physical capacity (Short Physical Performance Battery OR: 95% 1.30; CI: 1.02–1.66; Table 3) to be independently predictive of out-of-home activity. The model demonstrated a significant fit to the data (χ²⁸ = 38.738, p < .001), with a Nagelkerke R² of 0.407, indicating that the model explained 40.7% of the variance in the outcome variable.

Table 3.

Logistic regression analysis for out-of-home activity.

Logistic regression model	Without vs. with out-of-home activity
Nagelkerkes R²	.407
	Odd´s ratio (95% Confidence Interval)
Age (year)	0.97 (0.89–1.05)
Sex, female (%)	2.72 (0.86–8.59)
Living situation, living alone (%)	1.50 (0.52–4.34)

Duration of outpatient care^a	1.00 (1.00–1.00)

MOCA (score)	1.24 (1.02–1.50)*

SPPB (score)	1.30 (1.02–1.66)*

SOS (score)^a	1.25 (0.97–1.61)
FFABQ (score)	1.01 (0.96–1.05)

Natural log-transformed.

*Significance at a 0.05 level marked bold.

The table shows the regression analysis to identify variables with significant influences on leaving home as dichotomized variable in bivariate analysis. MOCA = Montreal Cognitive Assessment, scores 0–30; SOS = Evaluation of the Social Situation, scores 0–25; FFABQ = Fear of falling avoidance behavior questionnaire, scores 0–84; SPPB = Short Physical Performance Battery, scores 0–12.

Discussion

The present study provides a comprehensive comparison of multimorbid older adults with and without out-of-home activity following geriatric rehabilitation and identifies key determinants of out-of-home activity among persons with cognitive impairment. The findings reveal a markedly restricted mobility profile, with more than two-thirds of participants remaining exclusively at home, exceeding prevalence rates reported in previous studies of older populations.¹³ This highlights a pronounced risk of a ‘shrinking world’ in this particularly vulnerable group,¹⁶ likely driven by the accumulation of physical, cognitive, and psycho-social limitations.

Participants engaging in out-of-home activity demonstrated consistently higher activity levels across multiple dimensions, including step counts, walking episodes and duration, other active behaviour, standing time, and life-space mobility. Importantly, greater out-of-home activity was not compensated by reduction in in-home activity, suggesting that individuals remaining at home are not reallocating activity indoors but are instead constrained by limited physical capacity, exhaustion, or reduced energy.²⁵ These findings are consistent with previous studies in community-dwelling older adults,⁶ including those with cognitive impairment but not undergoing post-rehabilitation transition,⁷ and underline out-of-home activity as a marker of overall mobility rather than simple displacement of activity location.

Out-of-home activity was strongly associated with physical capacity, reflected by a clinically meaningful 1.6-point difference in Short Physical Performance Battery scores.²⁶ Physical and cognitive limitations are known to interact and may form a ‘vicious cycle’ accelerating functional decline and loss of independence.²⁷ In line with this, cognitive performance was lower among participants who remained homebound, consistent with previous findings in less vulnerable older adults with cognitive impairment.²⁸ Progressing cognitive deficits, particularly impairments in executive function, memory, and spatial navigation, likely limit planning abilities and reduce engagement in complex behaviours such as out-of-home activity.²⁹ In the present study, among psycho-social factors, apathetic symptoms were more prevalent in those staying at home, supporting apathy, alongside cognitive impairment, as a relevant barrier to out-of-home activity.¹⁴ In contrast, depressive symptoms did not differ significantly between groups, possibly reflecting the presence of meaningful in-home activities or social interactions that buffer mood despite reduced participation. Fear-of-falling-related avoidance behaviour, but not fall-related self-efficacy, was higher among homebound participants, suggesting that avoidance tendencies may more directly capture restrictions in out-of-home activity than self-efficacy measures. This emphasises the need to address motivational and behavioural avoidance in addition to physical capacity and self-efficacy.^11,29

Sociodemographic and environmental findings were largely consistent with previous literature. Older age and female sex were associated with staying home more frequently.^9,30 Living with others and a better social situation were linked to higher out-of-home activity, highlighting the importance of the immediate social environment and support, in line with previous research that identifies social isolation as a major barrier to mobility.^10,30 No differences were observed for financial sufficiency or weather conditions, likely reflecting the region's relatively homogeneous socioeconomic context and moderate climate. Increased healthcare use and support among homebound participants aligns with prior findings on growing dependency with increasing physical and cognitive impairment.^31,32 Notably, multimorbidity and self-rated health did not differ between groups, reinforcing the importance of functional limitations over disease counts, as emphasised by the World Health Organisation's International Classification of Functioning, Disability and Health framework.³³

Receiver operating characteristic analysis identified a threshold of 2140 steps/day with good discriminative performance, high sensitivity, and low specificity, indicating suitability for screening rather than diagnostic purposes. These findings correspond with prior work linking daily step counts to life-space mobility in cognitively intact older adults³⁴ and support step count as a quantitative mobility indicator. Beyond activity volume, walking pattern characteristics also showed discriminative potential. A mean walking episode duration threshold of 29.6 s showed acceptable discrimination, suggesting that the pattern of walking behaviour may contribute meaningful contextual information beyond overall activity quantity. This aligns with prior work emphasising the relevance of everyday walking pattern characteristics in older adults.³⁵ Longer uninterrupted walking episodes may reflect mobility patterns more commonly associated with activity across broader environments and greater spatial navigation demands, whereas shorter and more fragmented walking episodes may characterise more restricted mobility behaviour. Thus, walking episode duration may provide an ecologically relevant indicator of functional mobility pattern in everyday life. Importantly, distinguishing between out-of-home and in-home activity may provide a more contextually meaningful characterisation of mobility behaviour, as out-of-home activity has been associated with higher physical capacity,^1,6,7 better cognitive function,⁴ and greater social engagement.^1,3 The present study extends these findings to a highly vulnerable population, supporting out-of-home activity as a simple, multidimensional marker of mobility and vulnerability.

Among a broad range of potential determinants, as indicated by univariate associations to physical activity, only two major predictors were identified: physical capacity emerged as the most consistent determinant of out-of-home activity, in line with previous research,¹⁰ highlighting its role as a prerequisite for real-life mobility. Cognitive status also remained as an independent determinant, consistent with findings in cognitively unimpaired adults,³⁶ and reinforcing the complexity of out-of-home activity compared to in-home activity.¹¹ Evidence from prior studies supports the feasibility of improving physical capacity even in comparable vulnerable populations,^37–39 underscoring its relevance as a modifiable intervention target.

A key limitation of this study is the 48-h sensor monitoring period. Out-of-home activity may have been misclassified, as rare excursions could be missed, while a single event could lead to overestimation. Although short-term monitoring captures certain activity metrics reliably,⁴⁰ longer wear periods (e.g., five days) may improve reliability. Additionally, as the location tracking sensor was not body-worn, non-compliance cannot be fully excluded. Although several steps were implemented to minimise this risk, including clear placement instructions, repeated reminders, and a follow-up visit after 24 h, non-compliance of body-worn sensors cannot be fully excluded.

Overall, the findings highlight out-of-home activity as a multidimensional indicator of mobility, health, and autonomy in older adults with physical and cognitive impairment following geriatric rehabilitation. Whether assessed by sensor-based methods or a simple single-item question, out-of-home activity may offer a pragmatic and scalable approach for risk stratification and targeted intervention. The high prevalence of homebound individuals underscores the need for tailored strategies to support out-of-home mobility.⁴¹ Determinant analyses suggest intervention targets at multiple levels, including physical training,^21,39 cognitive training,⁴² and support for driving or alternative transportation.⁴³ Multimodal approaches may be particularly effective in mitigating mobility restrictions, reducing dependency, and preventing institutionalisation, especially among cognitively impaired adults who remain underserved by post-acute and intersectoral rehabilitation pathways despite retaining substantial rehabilitation potential.^38,44

Clinical messages

Out-of-home activity is infrequent in older adults with cognitive impairment after rehabilitation, highlighting a critical period of mobility vulnerability.

Sensor-derived activity metrics clearly distinguish adults who leave the home from those who remain indoors.

Physical capacity and cognitive function independently predict out-of-home activity this vulnerable, multimorbid population.

Supplemental Material

sj-docx-1-cre-10.1177_02692155261462457 - Supplemental material for Determinants of out-of-home activity in older adults with cognitive impairment following inpatient rehabilitation: A secondary cross-sectional analysis of a randomised controlled trial using location-based and accelerometer-derived measures

Supplemental material, sj-docx-1-cre-10.1177_02692155261462457 for Determinants of out-of-home activity in older adults with cognitive impairment following inpatient rehabilitation: A secondary cross-sectional analysis of a randomised controlled trial using location-based and accelerometer-derived measures by Phoebe Ullrich, Theresa Buchner, Martin Bongartz, Tobias Eckert, Carl-Philipp Jansen, Jürgen M. Bauer, Christian Werner and Klaus Hauer in Clinical Rehabilitation

Footnotes

Acknowledgements

We thank all individuals who contributed to this study, including the participants.

ORCID iDs

Phoebe Ullrich

Theresa Buchner

Carl-Philipp Jansen

Christian Werner

Ethical approval

The RCT was approved by the ethics committee of the Medical Department of Heidelberg University (S-252/2015) in accordance with the Declaration of Helsinki and was registered at (ISRCTN82378327).

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors gratefully acknowledge funding from the social and private long-term care insurance (soziale und private Pflegeversicherung) and the Municipal Association for Youth and Social Affairs in Baden-Württemberg (Kommunalverband für Jugend und Soziales Baden-Württemberg) (grant number 80221-208-009-01-01).

Declaration of conflicting interests

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

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Informed consent statement

Written informed consent by the participants was obtained.

Authors’ contributions

Phoebe Ullrich: conceptualization, data curation, formal analysis, investigation, methodology, validation, writing–original draft, and writing–review and editing.

Theresa Buchner: writing–review and editing

Martin Bongartz: data curation, writing–review and editing

Tobias Eckert: data curation, writing–review and editing

Carl-Philipp Jansen: writing–review and editing

Jürgen M. Bauer: resources, supervision, writing–review and editing

Christian Werner: writing–review and editing

Klaus Hauer: conceptualization, funding acquisition, investigation, methodology, supervision, validation, writing–review and editing.

Supplemental material

Supplemental material for this article is available online.

References

Webber

Porter

Menec

. Mobility in older adults: a comprehensive framework. Gerontologist 2010; 50: 443–450.

Ullrich

Werner

Abel

, et al. Assessing life-space mobility: a systematic review of questionnaires and their psychometric properties. Z Gerontol Geriatr 2022; 55: 660–666.

Mollenkopf

Marcellini

Ruoppila

, et al. Outdoor mobility and social relationships of elderly people. Arch Gerontol Geriatr 1997; 24: 295–310.

Hajek

Brettschneider

Lange

, et al. Longitudinal predictors of institutionalization in old age. PloS one 2015; 10: e0144203.

Inoue

Shono

Matsumoto

. Absence of outdoor activity and mortality risk in older adults living at home. J Aging Phys Act 2006; 14: 203–211.

Beyer

KMM

Szabo

Hoormann

, et al. Time spent outdoors, activity levels, and chronic disease among American adults. J Behav Med 2018; 41: 494–503.

Portegijs

Tsai

Rantanen

, et al. Moving through life-space areas and objectively measured physical activity of older people. PloS one 2015; 10: e0135308.

Ferrucci

Cooper

Shardell

, et al. Age-Related change in mobility: perspectives from life course epidemiology and geroscience. J Gerontol Series A, Biolog Sci Med Sci 2016; 71: 1184–1194.

Gagliardi

Spazzafumo

Marcellini

, et al. The outdoor mobility and leisure activities of older people in five European countries. Ageing Soc 2007; 27: 683–700.

10.

Kilgour

AHM

Rutherford

Higson

, et al. Barriers and motivators to undertaking physical activity in adults over 70—a systematic review of the quantitative literature. Age Ageing 2024; 53(4), afae080.

11.

Wettstein

Wahl

Shoval

, et al. Out-of-home behavior and cognitive impairment in older adults: findings of the SenTra project. J Appl Gerontol 2015; 34: 3–25.

12.

Dutzi

Schwenk

Kirchner

, et al. Influence of cognitive impairment on rehabilitation received and its mediating effect on functional recovery. J Alzheimer's Dis : JAD 2021; 84: 745–756.

13.

Harada

Lee

Park

, et al. Going outdoors and cognitive function among community-dwelling older adults: moderating role of physical function. Geriatr Gerontol Int 2016; 16: 65–73.

14.

Tung

Rose

Gammada

, et al. Measuring life space in older adults with mild-to-moderate Alzheimer's disease using mobile phone GPS. Gerontology 2014; 60: 154–162.

15.

Bayat

Naglie

Rapoport

, et al. A GPS-based framework for understanding outdoor mobility patterns of older adults with dementia: an exploratory study. Gerontology 2021; 68: 106–120.

16.

Duggan

Blackman

Martyr

, et al.

The impact of early dementia on outdoor life: a ‘shrinking world'?

Dementia 2008; 7(2): 191–204.

17.

Krumholz

. Post-hospital syndrome–an acquired, transient condition of generalized risk. N Engl J Med 2013; 368: 100–102.

18.

Sattari

Weitkamp

Meijering

. What happens behind doors? Exploring everyday indoor activities when ageing in place. J Aging Stud 2023; 64: 101109.

19.

Suri

VanSwearingen

Dunlap

, et al. Facilitators and barriers to real-life mobility in community-dwelling older adults: a narrative review of accelerometry- and global positioning system-based studies. Aging Clin Exp Res 2022; 34: 1733–1746.

20.

Breasail

Biswas

Smith

, et al. Wearable GPS and accelerometer technologies for monitoring mobility and physical activity in neurodegenerative disorders: a systematic review. Sensors (Basel, Switzerland) 2021; 21(24): 8261.

21.

Bongartz

Kiss

Ullrich

, et al. Development of a home-based training program for post-ward geriatric rehabilitation patients with cognitive impairment: study protocol of a randomized-controlled trial. BMC Geriatr 2017; 17: 214.

22.

Najafi

Aminian

Paraschiv-Ionescu

, et al. Ambulatory system for human motion analysis using a kinematic sensor: monitoring of daily physical activity in the elderly. IEEE Trans Biomed Eng 2003; 50: 711–723.

23.

Chung

Sargent

Brown

, et al. GPS Tracking technologies to measure mobility-related behaviors in community-dwelling older adults: a systematic review. J Appl Gerontol 2021; 40: 547–557.

24.

Carlson

Jankowska

Meseck

, et al. Validity of PALMS GPS scoring of active and passive travel compared with SenseCam. Med Sci Sports Exercise 2015; 47: 662–667.

25.

Søvde

Sandvoll

Natvik

, et al. In the borderland of the body: how home-dwelling older people experience frailty. Scand J Caring Sci 2022; 36: 255–264.

26.

Perera

Mody

Woodman

, et al. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc 2006; 54: 743–749.

27.

Giannouli

Fillekes

Mellone

, et al. Predictors of real-life mobility in community-dwelling older adults: an exploration based on a comprehensive framework for analyzing mobility. Europ Rev Ag Physic Activity 2019; 16: 19.

28.

Harada

Lee

, et al. Objectively-measured outdoor time and physical and psychological function among older adults. Geriatr Gerontol Int 2017; 17: 1455–1462.

29.

Wahl

Wettstein

Shoval

, et al. Interplay of cognitive and motivational resources for out-of-home behavior in a sample of cognitively heterogeneous older adults: findings of the SenTra project. J Gerontol Series B, Psychol Sci Soc Sci 2013; 68: 691–702.

30.

Harada

Lee

, et al. Environmental predictors of objectively measured out-of-home time among older adults with cognitive decline. Arch Gerontol Geriatr 2019; 82: 259–265.

31.

Vaarst

Boyle

Vestergaard

, et al.

Does physical performance and muscle strength predict future personal and nursing care services in community-dwelling older adults aged 75+?

Scand J Public Health 2021; 49: 441–448.

32.

Burgdorf

Wolff

. The impact of cognitive impairment on resource utilization during medicare home health care. Innovation Aging 2020; 4: 260.

33.

Ustün

Chatterji

Bickenbach

, et al. The international classification of functioning, disability and health: a new tool for understanding disability and health. Disabil Rehabil 2003; 25: 565–571.

34.

Tsai

Portegijs

Rantakokko

, et al. The association between objectively measured physical activity and life-space mobility among older people. Scand J Med Sci Sports 2015; 25: e368–e373.

35.

Hausdorff

Hillel

Shustak

, et al.

Everyday stepping quantity and quality among older adult fallers with and without mild cognitive impairment: initial evidence for new motor markers of cognitive deficits?

J Gerontol Series A, Biologic Sci Med Sci 2018; 73: 1078–1082.

36.

McInnes

Jones

Rochester

, et al. Mobility in community dwelling older adults: predicting successful mobility using an instrumented battery of novel measures. J Frailty Aging 2020; 9: 68–73.

37.

Seinsche

Jansen

Roth

, et al. Multidimensional interventions to increase life-space mobility in older adults ranging from nursing home residents to community-dwelling: a systematic scoping review. BMC Geriatr 2023; 23: 412.

38.

Ullrich

Werner

Bongartz

, et al. Increasing life-space mobility in community-dwelling older persons with cognitive impairment following rehabilitation: A randomized controlled trial. J Gerontol Series A 2021; 76(11): 1988–1996.

39.

Ullrich

Werner

Schönstein

, et al. Effects of a home-based physical training and activity promotion program in community-dwelling older persons with cognitive impairment after discharge from rehabilitation: a randomized controlled trial. J Gerontol Series A, Biological Sci Med Sci 2022; 77: 2435–2444.

40.

Abel

Pomiersky

Werner

, et al. Day-to-day variability of multiple sensor-based physical activity parameters in older persons with dementia. Arch Gerontol Geriatr 2019; Nov-Dec(85): 103911.

41.

Ringsten

Ivanic

Iwarsson

, et al. Interventions to improve outdoor mobility among people living with disabilities: a systematic review. Campbell Syst Rev 2024; 20: e1407.

42.

Kang

Kim

Lee

, et al. The beneficial effects of cognitive walking program on improving cognitive function and physical fitness in older adults. Healthcare (Basel, Switzerland) 2021; 9.

43.

Kizony

Schreuer

Rotenberg

, et al. Participation in out-of-home activities among older adults: the role of mobility, attitudes and travel behaviors. J Transp Health 2020; 17: 100846.

44.

Chenoweth

Kable

Pond

. Research in hospital discharge procedures addresses gaps in care continuity in the community, but leaves gaping holes for people with dementia: a review of the literature. Australas J Ageing 2015; 34: 9–14.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.07 MB