Fear of falling and its association with life-space mobility of older adults: a cross-sectional analysis using data from five international sites

Abstract

Background

fear of falling (FOF) is a major health concern among community-dwelling older adults that could restrict mobility.

Objective

to examine the association of FOF with life-space mobility (i.e. the spatial area a person moves through in daily life) of community-dwelling older adults from five diverse sites.

Methods

in total, 1,841 older adults (65–74 years) were recruited from Kingston, Canada; Saint-Hyacinthe, Canada; Tirana, Albania; Manizales, Colombia and Natal, Brazil. FOF was assessed using the Fall Efficacy Scale-International (FES-I total score), and the life space was quantified using the Life-Space Assessment (LSA), a scale that runs from 0 (minimum life space) to 120 (maximum life space)

Results

the overall average LSA total score was 68.7 (SD: 21.2). Multiple-linear regression analysis demonstrated a significant relationship of FOF with life-space mobility, even after adjusting for functional, clinical and sociodemographic confounders (B = −0.15, 95% confidence interval (CI) −0.26 to −0.04). The FOF × site interaction term was significant with a stronger linear relationship found in the Canadian sites and Tirana compared with the South American sites. After adjusting for all confounders, the association between FOF with LSA remained significant at Kingston (B = −0.32, 95% CI −0.62 to −0.01), Saint-Hyacinthe (B = −0.81, 95% CI −1.31 to −0.32) and Tirana (B = −0.57, 95% CI −0.89 to −0.24).

Conclusion

FOF is an important psychological factor that is associated with reduction in life space of older adults in different social and cultural contexts, and the strength of this association is site specific. Addressing FOF among older adults would help improve their mobility in local communities, which in turn would improve social participation and health-related quality of life.

Background

Fear of falling (FOF) has been defined as enduring concern about falling that leads a person to avoid daily activities that he/she could be capable of performing [1]. FOF has been associated with deconditioning, social isolation, more falls, greater frailty, decline in mobility and increased mortality [2–4]. Furthermore, it has been reported that FOF can change gait characteristics and result in overtly cautious adaptations of walking (e.g. reduced gait speed and increased step width) [5], which may further influence mobility.

There is an increasing interest in studying mobility as it relates to the surrounding environment, which is referred to as life-space mobility. Life-space mobility is the spatial area a person moves through in daily life [6], and reflects an individual's physical activity status [7] within the scope of sociocultural and environmental factors. Overall, restriction in life-space mobility is associated with cognitive impairment [8], frailty [9], poor physical functional health, disability, as well as mortality [10]. Measures of life space are unique and different from traditional measures of mobility in that in addition to measuring physical mobility, they take into consideration the interaction between an individual's functional ability and the greater sociocultural, physical and economical environments [11, 12].

Despite the established association between FOF and traditional mobility outcome measures, the association between FOF and life-space mobility has seldom been examined, and the limited available evidence is restricted to specific environments and populations (e.g. Japanese patients with mild cognitive impairment) [7, 13]. Therefore, ‘our research question’ was to what extent there is an association between FOF and life-space mobility in community-dwelling older adults and whether this association differs between socioculturally diverse sites. We ‘hypothesised’ that FOF is linearly but inversely related to life-space mobility. Furthermore, because life space is influenced by sociocultural/environmental factors, we further hypothesized that the strength of this relationship will differ between culturally diverse sites. Ultimately, the findings from this study will improve our understanding of the factors that hamper the independent life-space mobility of older adults living in diverse communities and will guide the design of enhanced strategies to expand life-space mobility.

Methods

Participants and sampling strategy

This study is part of the International Mobility In Aging Study (IMIAS), a longitudinal multisite research project examining the various factors related to mobility of community-dwelling older adults at five sites that are different in culture, economy and physical environment. Both ‘Kingston’ (population: 120,000) and ‘Saint-Hyacinthe’ (population: 50,000) are Canadian cities with flat topography. While Kingston is an Anglophone city, Saint-Hyacinthe is Francophone. ‘Tirana’ (Albania), in a central valley, has approximately 700,000 inhabitants. ‘Manizales’ (Colombia) is found in the Andes Mountains, with steep terrain (population: 400,000). ‘Natal’ is a coastal and flat city in Northeast Brazil (population: 817,590). These cities were particularly selected as the socioeconomic and cultural characteristics between them are significantly different while each has a relatively homogeneous population in culture and general characteristics. A sample of 1,995 community-dwelling participants, ages 65–74 years, was recruited in 2012 forming the IMIAS baseline data used in the present study (approximately 200 women and 200 men per site). Ethics approval was obtained from each site and all participants signed an informed consent form (see published papers for additional details [14]). The Leganes Cognitive Test (LCT) was used to screen mental status during initial recruitment and individuals who were disoriented (≥4 errors on the LCT orientation scale) were excluded. For the current study, we further excluded participants with possible dementia (LCT≤ 22) [15] to avoid potential errors in responses as well as those who did not complete the FOF and/or life-space mobility questionnaires.

Measurement tools

The Falls Efficacy Scale-International (FES-I) questionnaire was used to measure FOF. FES-I measures the level of concern about falling during 16 social/physical activities in the home and community, using a 4-point Likert scale [16]. FES-I score ranges from 16 to 64, with higher scores indicating greater concern. FES-I has been translated and validated in several languages and countries to measure FOF [16, 17].

The Life-Space Assessment (LSA) [6] measures an individual's pattern of mobility, across five levels of living space (from within the home to out of town), during the month prior to assessment. For each level of living space, respondents were asked about their frequency of travel (from once a week to daily), as well as their reliance on people (yes or no) or assistive mobility devices (yes or no). The LSA total scores ranges from 0 (complete restriction to bed) to 120 points (complete independence, including ability to travel frequently out of town). The LSA has been validated for use in several populations, including those in North and South America [6, 12].

Several functional, clinical and sociodemographic variables may influence and potentially confound the relationship of FOF with life-space mobility. Therefore, these confounders were grouped into two subgroups and accounted for in the analysis (a full description of the confounder variables and how they were recorded is provided in Supplementary data, Appendix 1, available in Age and Ageing online):

• ‘Functional and clinical variables’: grip strength, physical performance (measured using the short physical performance battery [SPPB]) [18], vision acuity [19], depressive symptoms (measured by the Center for Epidemiological Studies Depression [CES-D] scale) [20], global cognition (using LCT) [15], number of comorbidities and number of falls.

• ‘Sociodemographic variables’: age, sex, years of formal education, income sufficiency and social support (assessed using IMIAS-Social Networks and Social Support [SNSS] scale) [14, 21].

Statistical analyses

Distributions of baseline characteristics were compared between sites using one-way analysis of variance (ANOVA) for continuous variables and chi-square tests for categorical variables. Multiple-linear regression was performed to estimate mean change and 95% confidence intervals (CIs) of the outcome variable (LSA total score) in relation to change in the explanatory variable (FES-I total score). A P-value < 0.05 was considered statistically significant. All assumptions for linear regression models were tested (linearity, normality, independence and equal variance) and fulfilled. Additionally, correlations between variables were checked for possible multicollinearity problems (but no evidence was found). Potential confounding variables were identified using linear regression adjusting for age to test for any association with the outcome variable. Those with P < 0.1 were included in the multivariate regressions.

In order to test the unique effect of each confounder group on the relationship between FOF and life space, they were added one at a time—Model type 1: a basic model (without adjusting for any confounders); Model type 2: adjusting for functional and clinical outcomes (grip strength, SPPB, vision acuity, depressive symptoms, cognition, comorbidities and number of falls); Model type 3: adjusting for sociodemographic factors (age, sex, education, income sufficiency and social support provided); Model type 4: fully adjusted (all variables from Model types 2 + 3 together). The interaction terms (FOF total score × site and FOF × sex) were then added and tested in Model type 4. No significant interaction was found for FOF × sex (P = 0.45) but the interaction was significant for FOF × site (P < 0.0001). Hence, all the above Model types 1–4 were then fitted for each site separately. Statistical analyses were performed using SAS Enterprise Guide version 6.1 (SAS Institute Inc., Cary, NC).

Results

Out of 1,995 IMIAS participants, 117 were excluded from the present analysis because of possible dementia (LCT ≤ 22) [15], 3 participants were excluded because they did not complete the FES-I questionnaire and 37 because they did not complete the LSA questionnaire. Therefore, 1,841 participants were included in the final analysis (Kingston, Canada: 394; Saint-Hyacinthe, Canada: 380; Tirana, Albania: 354; Manizales, Colombia: 329 and Natal, Brazil: 384).

For the overall sample, the average LSA total score was 68.7 (SD: 21.2); its distribution was significantly different at the five study sites (P < 0.001) and with averages ranging from 56.6 in Natal, Brazil to 82.6 in Kingston, Canada (Table 1). The average LSA total score was significantly higher in men for the entire sample (P < 0.001) as well as at each site, and a similar pattern was noticed with regard to FOF (Figure 1). Overall, the average FES-I total score was 23.3 (SD: 8.8), and was lowest in Saint-Hyacinthe, Canada and highest in Manizales, Colombia (overall P < 0.001) (Table 1).

Table 1.

Differences in the characteristics of participants among the five sites. Results reported in means (SD) unless otherwise specified.

Variable	Overall sample (n = 1,841)	Kingston (n = 394)	Saint-Hyacinthe (n = 380)	Tirana (n = 354)	Manizales (n = 329)	Natal (n = 384)	Pa
LSA total score	68.7 (21.2)	82.6 (16.7)	78.3 (19.3)	60.4 (22.4)	64.2 (16.3)	56.6 (16.2)	<0.001
FES-I score	23.3 (8.8)	21.1 (6.4)	19.3 (4.9)	22.6 (8.6)	28.8 (10.4)	25.6 (9.5)	<0.001
Age (years)	69.0 (2.8)	69.1 (2.7)	68.5 (2.6)	69.3 (3.1)	69.2 (3.0)	69.1 (2.7)	0.009
Sex (women %)	51.9	53.0	52.1	50.9	52.0	51.6	0.971
Years of education	10 (5.7)	15.8 (4.0)	12.2 (4.0)	10.5 (3.9)	6.1 (4.6)	4.8 (3.9)	<0.001
Depressive symptoms	9.4 (9.0)	6.6 (7.7)	6.7 (7.0)	14.4 (11.2)	9.8 (8.2)	10.3 (8.6)	<0.001
Global cognition	29.6 (2.4)	30.6 (1.3)	29.1 (1.8)	28.1 (2.3)	27.0 (2.7)	28.0 (2.3)	<0.001
Grip strength	28.0 (9.3)	29.6 (9.9)	31.4 (9.5)	28.3 (9.3)	25.5 (7.5)	25.4 (8.7)	<0.001
SPPB	9.8 (2.0)	10.3 (1.8)	10.3 (1.6)	9.3 (2.6)	9.7 (1.8)	9.2 (2.3)	<0.001
Visual acuity	44.7 (10.6)	51.3 (7.7)	48.5 (8.1)	39.4 (9.2)	38.2 (10.7)	44.3 (10.6)	<0.001
Comorbidity	1.9 (1.3)	1.8 (1.3)	1.7 (1.3)	2.3 (1.3)	1.5 (1.3)	2.0 (1.3)	<0.001
Number of falls in the last year	0.6 (2.0)	1 (3.4)	0.4 (0.8)	0.3 (0.8)	0.7 (2)	2 (1.3)	<0.001
Income sufficiency % (income does not meet needs)	42.0	5.3	7.6	61.5	68.3	74.0	<0.001
Social support from friends	13.1 (6.8)	15.1 (3.9)	13.6 (4.5)	14.7 (6.0)	13.8 (7.8)	8.5 (8.5)	<0.001
Social support from partner	12.0 (8.8)	11.6 (9.1)	12.5 (8.3)	15.0 (8.1)	9.5 (9.5)	11.4 (8.3)	<0.001

Visual acuity was recorded as the number of correctly identified Es; depressive symptoms were assessed using the CES-D scale; comorbidity was recorded as a total number of chronic conditions (0–8); global cognition was evaluated using LCT score; income sufficiency was categorised into whether income meets needs or not; social support from friends and partner variables were measured using IMIAS-SNSS.

^aSignificant test between all sites using ANOVA for continuous variables and chi-square for categorical variables.

Figure 1.

(A) The average total score of the FES-I at the five sites for men and women, aged 65–74 years old and 95% CI and (B) the average total score of LSA at the five sites for men and women, aged 65–74 years old and 95% CI.

The between-site differences in all the confounding variables and characteristics were statistically significant except for sex distribution (Table 1). Canadian participants had lower depression symptoms, better cognitive and physical function, better visual acuity, and they were more satisfied with their income (Table 1). All confounding variables discussed previously were included in the relevant regression models (Model types 2–4) except support provided by children and family variables, as they were not associated with the outcome (P ≥ 0.1).

The results of the linear regression testing the association of participant's FOF (FES-I score) with life-space mobility (the LSA total score) are presented in Table 2. A significant linear relationship was found in Model type 1 for the overall sample between FOF and LSA where the Pearson correlation was −0.43, and one-point increase in FOF was associated with one-point decrease in the LSA score. This relative change in the LSA score that is associated with the FES-I total score remained significant after accounting for the functional and clinical variables (Model type 2; parameter estimate (B) = −0.25, P < 0.001), for the sociodemographic variables (Model type 3; B = −0.64, P < 0.001) and even after fully adjusting for all the confounding variables in the previous models together (which does not include site) (Model type 4; B = −0.15, P < 0.001).

Table 2.

Multiple-linear regression models for the change in the LSA total score in relation to change in the FOF (FES-I total score) for the overall sample and stratified by site.

Variable	Model type 1		Model type 2		Model type 3		Model type 4
	Parameter estimate (B) 95% CI	R2	Parameter estimate (B) 95% CI	Adjusted R2	Parameter estimate (B) 95% CI	Adjusted R2	Parameter estimate (B) 95% CI	Adjusted R2
Overall	−1.01	0.18	−0.25	0.37	−0.64	0.32	−0.15	0.42
	−1.11 to −0. 91		−0.36 to −0.13		−0.74 to −0.54		−0.26 to −0.04
Kingston	−1.07	0.17	−0.35	0.28	−0.94	0.19	−0.32	0.28
	−1.31 to −0.84		−0.66 to −0.04		−1.18 to −0.69		−0.62 to −0.01
Saint-Hyacinthe	−1.58	0.16	−0.91	0.24	−1.15	0.24	−0.81	0.27
	−1.94 to −1.22		−1.41 to −0.4		−1.50 to −0.77		−1.31 to −0.32
Tirana	−1.34	0.26	−0.56	0.37	−1.12	0.32	−0.57	0.40
	−1.57 to −1.11		−0.88 to −0.24		−1.36 to −0.87		−0.89 to −0.24
Manizales	−0.46	0.07	−0.01	0.27	−0.36	0.13	−0.03	0.27
	−0.63 to −0.28		−0.20 to 0.17		−0.53 to −0.18		−0.18 to 0.21
Natal	−0.49	0.08	−0.07	0.23	−0.31	0.14	−0.05	0.24
	−0.65 to −0.32		−0.27 to 0.12		−0.49 to −0.14		−0.24 to 0.14

Model type 1: a basic model (without adjustment); Model type 2: adjusting for functional and clinical outcomes (grip strength, SPPB total score, vision acuity, depressive symptoms, global cognition, number of comorbidities and number of falls in the last year); Model type 3: adjusting for sociodemographic factors (age, sex, education, income sufficiency and social support provided by friends and partner); Model type 4: fully adjusted (Model types 2 + 3 together); Model type 4 with the interaction term ‘FOF × site’ was also tested but it is not shown here as the conditional estimate does not have meaning now. Each parameter estimate in this table was derived from a separate model.

In the full model, an interaction term for FES-I × site was significant so relationships in each site were examined independently (Table 2). The association between the FES-I and LSA in Model type 1 was strongest at Saint-Hyacinthe (B= −1.58, P < 0.001) followed by Tirana (B = −1.34, P < 0.001) and then by Kingston (B = −1.07, P < 0.001). For Saint-Hyacinthe, Tirana and Kingston, this relative change in the LSA as a result of change in the FES-I total score remained significant even after fully adjusting for all the confounding variables in all models together (Model type 4). In Manizales and Natal, the relationship between the FES-I and LSA total scores was weak and seemed to be primarily explained by the functional and clinical factors, since the relationship between the FES-I and LSA scores became non-significant in Model type 2.

Discussion

Overall, the study findings support our hypothesis that FOF is related to life-space mobility, and that the strength of this relationship differs between sites. Before adjusting for confounders, the relationship between FOF and life space was significant for the overall sample and for all the sites. After adjustment, this relationship remained significant for the overall sample and for three out of the five sites.

Published work on the relationship between FOF and life space is very limited. Only two previous Japanese studies tested the relationship between FOF and life space [7, 13]. One of the two earlier studies included older adults with amnestic cognitive impairment and the other one included functional older adults; both had relatively small samples. The findings from these studies are consistent with our overall findings.

Our previous work found a cross-sectional relationship between FOF and mobility disability, a precursor for end-stage disability, across the five sites in this study [22]. FOF could cause significant reduction in physical activities, thus leading to poor balance and musculoskeletal health, which in turn could cause less mobility, decreased social participation and reduced life space. Additionally, FOF may be a marker of health conditions that go on to cause disability.

While we found evidence to support the relationship between FOF and life-space, the strength of this relationship varied between sites, which shows the importance of site-specific approach in research. Life space is related to several factors. Therefore, the way FOF and life-space mobility interact and influence each other would likely depend on site-specific factors including cultural and environmental differences.

The relationship between FOF and life space was significant but weak at two sites (Manizales, Colombia and Natal, Brazil) and explained very little of the variance in LSA scores even before accounting for confounders. After accounting for the confounders, the relationship became non-significant and the coefficient in the regression models was substantially smaller than the coefficient for other sites.. This finding is not surprising since most of the other factors potentially influencing the LSA scores (included as potential confounders in this study) were significantly worse at these two sites (see Table 1). Furthermore, physical environmental barriers are more common at these two sites. Older adults at these two sites require high functional reserves to overcome these physical barriers and keep their life space from shrinking as they age [23]. Additionally, participants at Manizales and Natal lack the financial means to offset the effect of the environmental challenges on their life space, so they need to depend mainly on their functional reserves. Finally, personal safety is a major factor that can restrict life-space mobility due to high crime rate at these two sites.

This is an international study of the relationship between FOF and life-space mobility in older adults that includes culturally diverse cities from middle- and high-income countries. The use of standardised measurement tools and rigorous data collection methodology across all five sites minimised sources of variability. On the other hand, the results should be interpreted taking the following limitations into consideration. The cross-sectional design of this study prevents causal interpretation of the relationships found. Furthermore, the sample that participated at the Canadian sites might not be fully representative since we were unable to randomly sample participants from the general population like other sites due to Canadian ethics committees’ requirements [14]. Finally, the LSA scale asks about life space in the last month and, therefore, could be influenced by the weather at locations with drastic differences in temperatures [24]. However, the LSA has been found to be fairly reliable across seasons for the same population [24, 25].

Clinical and research implications

The association of FOF with life space that was shown in this study is important for clinical practice. Site wise, FOF explained up to one-fourth of the change in the LSA score. FOF had a graded association with life space, which could be substantial depending on the site. For instance, holding all confounders stable at Saint-Hyacinthe (where the strongest relationship was found), a one-point increase in the FES-I score was associated with 0.8 point decrease on the LSA score. This could translate into losing up to 39 points on the LSA scale on average for those individuals with the maximum FOF as measured by the FES-I, almost one-third of their total life space. However, FOF is modifiable and amenable to interventions [26]. Identifying and addressing FOF would help improve life-space mobility, which in turn would improve participation and quality of life [27].

Emerging evidence supports multifactorial interventions for geriatric patients [28], which could also address FOF. Healthcare professionals are encouraged to evaluate and manage FOF in older persons with mobility challenges. Furthermore, the findings demonstrate the importance of including the assessment and management of psychological factors (specifically FOF) in training curriculums for healthcare professionals, specifically for those who directly deal with mobility challenges in older adults. This training in the management of psychological factors for healthcare professionals seems to be inadequate to date [29].

In conclusion, to the best of our knowledge, this is the first international study to find evidence to support the relationship between FOF and life-space mobility and that the strength of this relationship differs between sites. Identifying and addressing FOF may help improve life-space mobility, which consequently would improve social participation and health-related quality of life.

Key points.

• The association between fear of falling (FOF) and mobility as it relates to the surrounding environment has seldom been examined.

• In this study, we found that FOF is related to life-space mobility, and the strength of this relation differs by site.

• The way FOF and life-space mobility interact and influence each other would depend on site-specific factors.

• The findings could improve our understanding of the factors that hamper older adults’ independent mobility in the community.

Supplementary data

Supplementary data are available at Age and Ageing online.

Conflicts of interest

None declared.

Funding

This project is supported by a grant from the Canadian Institutes of Health Research (CIHR) [grant number 108751].