Mobility restriction as a consequence of a fall is a major issue of care in assisted-living facilities. These facilities are along the continuum between living in one’s own private home in a community setting and living in a nursing home. They provide limited assistance for activities of daily living (ADLs) but do not provide the more advanced skilled care one may receive in a nursing home (Towne, Lee, Li, & Smith, 2016). Studies in Germany assessing incidence in community-dwelling women and men aged 65–90 show rates of 38.7% and 29.7% fallers/year and 13.7% and 10.9% recurrent fallers/year, respectively (Rapp et al., 2014). However, fall incident rates differ when comparing across settings (nursing homes vs. assisted-living facilities). Studies suggest that 15–28% of older adults residing in assisted-living facilities had one or more falls in the past 90 days (Klingelhöfer-Noe, Dassen, & Lahmann, 2015; Stock, Amuah, Lapane, Hogan, & Maxwell, 2017; Towne, Cho, Smith, & Ory, 2017). Rubenstein and Josephson (2002) found that the rate of falls in nursing homes was nearly twice that of community-dwelling adults and that fallers are more than twice as likely to have a hospitalization (Gimm & Kitsantas, 2016). More troubling are the facts that one third of falls result in serious injuries (Büchele et al., 2014) and that falls can result in significant changes in an individual’s willingness to engage in social activities and can result in a decrease in physical activity (Hoang, Jullamate, Piphatvanitcha, & Rosenberg, 2017; Ruthig, Chipperfield, Newall, Perry, & Hall, 2007; Hall & McAuley, 2011). Increasing inactivity may be due to mobility limitations after an injury or to avoiding activities because of fear of falling (Leung et al., 2017). Lower levels of physical activity may in turn decrease strength and balance and initiate a downward cycle toward losing independence and entering long-term care (Graafmans, Lips, Wijlhuizen, Pluijm, & Bouter, 2003).
To prevent severe consequences caused by falls, one strategy is to identify risk factors and to act on those that are modifiable. Although fall risk is multifactorial (individual characteristics [e.g., sex, education, age, needs of ADL], interpersonal factors [e.g., participation in social activities], institutional factors [e.g., size of institution, ownership type], community factors [e.g., rural, urban]), being female, being in need of assistance with at least one ADL, and residing in a large facility are the most common causes of falls among residents in assisted-living settings (Towne et al., 2017). Moreover, older people living in assisted-living settings experience impaired cognition, reduced mobility, poor balance, and reduced overall physical activity when compared with their peers living in the community (Corcoran et al., 2016; Moran, MacMillan, Smith-Merry, Kilbreath, & Merom, 2015; Nitz & Josephson, 2011; Stenhagen, Ekström, Nordell, & Elmståhl, 2014). Physical inactivity and a sedentary lifestyle (periods of inactivity, typically while sitting or lying down) have also been linked to falls, mainly because of decreasing muscle mass contributing to the development of osteoporosis and osteopenia (Franco, Pereira, & Ferreira, 2014; Gillespie et al., 2012). Low falls efficacy (self-perceived confidence in performing daily physical activities) is associated with an increased risk of subsequent falls, a decline in the ability to carry out ADLs, and reduced quality of life (Cumming, Salkeld, Thomas, & Szonyi, 2000; Hughes, Kneebone, Jones, & Brady, 2015; Schott, 2007).
Although social support has been found to be directly and positively related to various health outcomes and to buffer adverse effects (e.g., stress, depression, anxiety, and physical disability) on health (Wolff, Lindenberger, Brose, & Schmiedek, 2016), only a few studies have explicitly examined social support in relation to fall events (Durbin, Kharrazi, Graber, & Mielenz, 2016; Pin & Spini, 2016). These authors argue that social support could have a direct, primary protective effect on falls, because friends and relatives help older adults to be more aware of risky situations or help prevent such situations by helping the person concerned. Durbin et al. (2016) observed no relationship between social contact and perceived availability of social support and falls in a population of more than 1,000 community-dwelling older adults. But, the authors note that the direction of the odds ratios for social contact changed when adjusted for covariates (e.g., demographics, vision difficulties, presence of health issues, physical activity). However, the close proximity of the odds ratio to 1.00 may indicate that the findings are random. By contrast, Pin and Spini (2016) found, in a longitudinal four-wave study, a negative association between falls and social participation and a positive association with social support. The latter finding indicates the higher involvement of relationships after a fall. Both studies assessed two facets of social support (contact and perception), but they used only one question to evaluate each facet. In addition, the studies used different time intervals; Dubin et al. (2016) focused on the previous month, and Pin and Spini (2016) examined only the previous week.
The nature of the relationship between social support and falls remains somewhat unclear, because the many factors they share may provide confounding information but have not been evaluated in the aforementioned studies. Few empirical studies have been carried out to examine psychosocial factors affected by falls among older adults in assisted-living homes. However, none of these previous studies have examined the relationships among the relevant factors simultaneously (e.g., perceived social support, falls efficacy, and physical activity). In addition, to the best of our knowledge, there have been no studies investigating the relationship between falls and physical activity through mediating factors such as social support or falls efficacy (i.e., confidence or belief in one’s ability to perform activities without losing balance or falling). Therefore, the primary purpose of this study was to propose and test a model that explains the interrelationships among these factors using structural equation modeling. The secondary purpose was to examine the mediating roles of social support and falls efficacy on the relationship between falls and physical activity. The rationale for this mediation model is justified by four aspects. First, studies show a nonlinear relationship between fall risk and physical activity, with a protective effect for highly active older adults in assisted-living but more likely falls in individuals with low or intermediate physical activity levels (Graafmans et al., 2003). Second, falls do not directly lead to behavioral changes in physical activities; rather, affective–cognitive variables, such as falls efficacy, mediate this process. It is suggested that individuals with higher falls efficacy are better able to appraise the experience of a fall and will adjust their activity accordingly (Delbaere, Crombez, van Haastregt, & Vlaeyen, 2009). Third, according to Smith, Banting, Eime, O’Sullivan, and van Uffelen (2017), social support is an important factor in assisting older adults to be physically active. Fourth, recent studies indicate that cognitive factors, such as self-efficacy, mediate the association of social support and physical activity. This means that social–environmental factors, such as social support, positively influence cognitive factors, such as self-efficacy, which in turn positively influence physical activity (Aghdam et al., 2013; Ishii, Shibata, & Oka, 2010). Consequently, the current study used a mediation method to understand (a) the direct effects of falls on the physical activity of older adults, (b) the indirect effect of falls on the physical activity of older adults through falls efficacy, (c) the indirect effect of falls on the physical activity of older adults through social support, and (d) the indirect effect of falls on the physical activity of older adults through both falls efficacy and social support.
Methods
Study Design and Participants
A cross-sectional design was used to examine the relationship between social support, physical activity, falls efficacy, and falls.
The power analysis for structural equation modeling suggested by MacCallum, Browne, and Sugawara (1996) yielded 40 participants for .80 power, df = 14, root mean square error of approximation (RMSEA)-H0 = .20 (high misfit), and RMSEA-H1 = .05 (close to good fit) (p = .05). We maximized power by recruiting 81 participants (56 women and 25 men) aged 66–94 (Mage = 79.8, SD = 6.75). The participants were residents of different assisted-living facilities in the area of Münster, Germany. Subjects were excluded if they had dementia (as documented in the residents’ files) or were not able to walk with or without an assistive device. Informed written consent was obtained prior to the beginning of testing. Participants were told that they could opt out at any time. All procedures were in accordance with the Declaration of Helsinki ethical standards, legal requirements, and international norms; all study procedures were reviewed and received review board approval by the University of Münster.
Materials and Procedures
Participants were provided with a standardized set of instructions regarding questionnaire completion and for the use of the pedometer. Questionnaires were administered by trained research assistants at the homes of the participants. Three-day physical activity behavior was determined using a piezoelectric pedometer.
Social support
Social support was assessed by the German Social Support Questionnaire-Short Form 22 (Fragebogen zur Sozialen Unterstützung, F-SozU-K-22; Fydrich, Geyer, Hessel, Sommer, & Brähler, 1999). This 22-item, written, self-reported measure (5-point scale ranged from “relevant” to “not relevant,” with higher scores indicating better social support) is a widely used instrument. It has been validated for ages 16–96 in Germany. It investigates the perception of social support using three main scales: emotional support (“I have friends or family members who listen to me when I want to talk about a problem”), tangible support (“When I am sick, I can ask friends/relatives to handle important things for me without hesitation”), and social integration (“I often see myself as an outsider”). Cronbach’s alphas of the scales were .93, .88, .81, and .78 for the total score, emotional support, tangible support, and social integration, respectively (intercorrelations r = .643–.776; p < .001). Fydrich et al. (1999) recommended using the total score because of the strength of the emotional support factor in the exploratory factor analysis and high intercorrelations between emotional support, tangible support, and social integration.
Physical activity
Three-day physical activity behavior was determined using a piezoelectric pedometer (NL-2000; New LifeStyles Inc., MT). These were accelerometers with pedometer-like features that were able to record daily step counts and daily distance for up to 7 days. They have been shown to be one of the most reliable and valid activity monitors available (Grant, Dall, Mitchell, & Granat, 2008; Schneider, Crouter, & Bassett, 2004). Participants were instructed to wear the activity monitors for all waking hours except during bathing, swimming, and some contact sports.
Falls efficacy
Participants were assessed with the Activities-specific Balance Confidence scale (ABC: Powell & Myers, 1995; German short and long version: Schott, 2014), a 16-item self-reported questionnaire that asks people to score their perceived level of fall efficacy when performing common indoor and outdoor ADL (e.g., walking up and down stairs, standing on tiptoes, and reaching for something overhead). The six-item short form includes the six most challenging items on the confidence scale. Each item was scored on an 11-level rating scale, from 0% (no confidence) to 100% (full confidence in performing the activity without losing balance). The total Activities-specific Balance Confidence score is the mean sum of the individual item scores. Validity is suggested because the Activities-specific Balance Confidence scale discriminates between fallers and nonfallers (Schott, 2014). In this study, Cronbach’s alpha was .96 for the 16-item scale and .90 for the 6-item scale.
Fall history
A fall was defined as an event during which a subject comes to rest on the ground or at some lower level not as the result of a major intrinsic event or overwhelming hazard (Tinetti & Williams, 1998). Participants reported the number of falls experienced in the previous year; one or more falls classified as a positive fall history.
Demographic information
Participants provided information regarding their age, sex, and the number of people living in their household.
Data Analysis
Statistical analysis was performed using SPSS 23 for Windows (SPSS Inc., Chicago, IL) in conjunction with Analysis of Moment Structures Graphics (AMOS; version 23; Small Waters Corp., Chicago, IL).
Analyses of variance ascertained whether there were differences between the male and female fallers and nonfallers groups for the demographics and main risk factors measures. Cohen’s d was employed to compute the effect sizes when comparing the two groups.
Analysis of Moment Structures was used to conduct structural equation modeling with maximum likelihood estimation, which allowed for an examination of the hypothesized direct and indirect relationships between social support, falls efficacy, physical activity, and falls. The extent of mediation was calculated as the product of the indirect and direct effect. Data for all indicators met assumptions of univariate normality (i.e., skewness indices ≤3, kurtosis indices ≤ 10; Chou & Bentler, 1995). Regarding multivariate normality, Mardia’s coefficient was 6.26, and two participants emerged as multivariate outliers with Mahalanobis distances significant at p < .001. However, none of these participants exhibited patterns of random responding, and removal of their data did not affect fit indices or parameter estimates. Five measures of model fit were calculated: maximum likelihood chi-square (χ2), comparative fit index, the Tucker–Lewis index, standardized root mean square residual, and the RMSEA. Based on recommendations by Hu and Bentler (1999), models were considered to have adequate fit if the comparative fit index and Tucker–Lewis index were > .90 and the standardized root mean square residual and RMSEA were < .08.
Results
Sample Characteristics
Table 1 summarizes participants’ characteristics by fall. A total of 44 subjects (54.3%) reported having fallen in the previous 12 months; of these, 27 (61.4%) reported having fallen more than once.
M ± SD for Characteristics of the Sample According to Fall Status and Sex
| Fallers | Nonfallers | Statistical Analysis | |||||
|---|---|---|---|---|---|---|---|
| Men (n = 12) | Women (n = 32) | Men (n = 13) | Women (n = 24) | Group | Sex | Group × Sex | |
| Age (years) | 85.1 ± 6.42 | 80.8 ± 6.54 | 77.1 ± 5.39 | 77.3 ± 6.25 | ** | – | – |
| Household size (n) | 1.42 ± 0.51 | 1.59 ± 0.56 | 2.00 ± 0.00 | 1.38 ± 0.58 | – | *** | ** |
| ABC scale | |||||||
| short 6 items | 28.3 ± 26.3 | 37.2 ± 26.4 | 70.4 ± 26.1 | 53.3 ± 23.0 | ** | – | * |
| long 16 items | 42.7 ± 26.0 | 58.7 ± 23.1 | 80.1 ± 20.6 | 71.6 ± 18.8 | ** | – | * |
| FSozU-K-22 | |||||||
| emotional support | 3.60 ± 1.02 | 4.19 ± 0.98 | 4.45 ± 0.62 | 4.44 ± 0.47 | ** | – | – |
| tangible support | 3.58 ± 0.96 | 4.26 ± 0.96 | 4.62 ± 0.44 | 4.73 ± 0.56 | ** | * | – |
| social integration | 3.25 ± 1.05 | 3.84 ± 1.03 | 4.21 ± 0.87 | 4.29 ± 0.51 | ** | – | – |
| total | 3.50 ± 0.89 | 4.03 ± 0.85 | 4.38 ± 0.45 | 4.40 ± 0.46 | ** | – | – |
| Falls (n) | 4.17 ± 3.93 | 2.28 ± 1.84 | 0.00 ± 0.00 | 0.00 ± 0.00 | ** | * | * |
| Physical activity | |||||||
| number of steps/day | 584 ± 522 | 1245 ± 809 | 1562 ± 582 | 1623 ± 731 | ** | * | *** |
| percentilea | 16.2 ± 12.7 | 34.2 ± 20.6 | 28.7 ± 14.3 | 38.5 ± 22.2 | *** | ** | – |
Note. ABC = Activities-specific Balance Confidence; FSozU-K-22 = social support questionnaire-short form 22.
aPercentile calculation after Tudor-Locke et al. (2013).
*p < .05. **p < .01. ***p < .10.
Fallers and nonfallers differed significantly in age, F(1, 77) = 14.5, p < .001,
Interrelationships Between Social Support, Falls Efficacy, Physical Activity, and Falls
Correlations between the predictor variables and physical activity are shown in Table 2. Each of the variables was significantly associated with the number of steps taken (p < .01); falls efficacy and social support demonstrated positive associations, whereas age and falling exhibited negative relationships with physical activity. The strength of the correlations indicates that much of the variance in the number of falls cannot be explained by any one of the predictor variables in isolation.
Correlations Among Predictor Variables and Physical Activity (Number of Steps)
| Sex | Household | ABC-6 | ABC-16 | FSozU_ES | FSozU_TS | FSozU_SI | Steps | Falls | |
|---|---|---|---|---|---|---|---|---|---|
| Age | .114 | −.203 | −.582** | −.656** | −.291** | −.374** | −.387** | −.415** | .415** |
| Sex | – | .187 | .120 | −.019 | −.136 | −.182 | −.142 | −.191 | .135 |
| Household | – | .276* | .252* | .251* | .224* | .258* | .179 | .011 | |
| ABC-6 | – | .945** | .490** | .490** | .554** | .651** | −.501** | ||
| ABC-16 | – | .559** | .575** | .597** | .670** | −.562** | |||
| FSozU_ES | – | .776** | .750** | .490** | −.451** | ||||
| FSozU_TS | – | .643** | .488** | −.516** | |||||
| FSozU_SI | – | .587** | −.497** | ||||||
| Steps | – | −.475** |
Note. ABC = Activities-specific Balance Confidence; FSozU = social support questionnaire; ES = emotional support; TS = tangible support; SI = social integration. ABC-6, ABC-16: falls efficacy.
*p < .05. **p < .01.
The structural equation model is depicted in Figure 1. Standardized regression weights are shown for associations between each variable. Falls were associated with higher social support, yet falls were not correlated directly with physical activity. Social support was indirectly related to physical activity through falls efficacy. Social support had a positive impact on falls efficacy, which was positively related to physical activity. Age was directly associated with falls, and living in a household with a partner was directly related to perceived social support. Although not statistically significant, there was a trend suggesting that women showed higher falls efficacy and physical activity levels; those with a higher number of falls reported lower falls efficacy. Associations between sex and falls, falls efficacy, or social support did not meet the significance threshold.
—Proposed casual paths of physical activity behavior (standardized solution). All solid lines represent significant effects (*p < .05; **p < .01; ***p < .10); bold lines represent paths of study interest; broken lines represent proposed paths that were nonsignificant. Male: n = 25; female: n = 66.
Citation: Journal of Aging and Physical Activity 27, 1; 10.1123/japa.2017-0378
—Proposed casual paths of physical activity behavior (standardized solution). All solid lines represent significant effects (*p < .05; **p < .01; ***p < .10); bold lines represent paths of study interest; broken lines represent proposed paths that were nonsignificant. Male: n = 25; female: n = 66.
Citation: Journal of Aging and Physical Activity 27, 1; 10.1123/japa.2017-0378
—Proposed casual paths of physical activity behavior (standardized solution). All solid lines represent significant effects (*p < .05; **p < .01; ***p < .10); bold lines represent paths of study interest; broken lines represent proposed paths that were nonsignificant. Male: n = 25; female: n = 66.
Citation: Journal of Aging and Physical Activity 27, 1; 10.1123/japa.2017-0378
Path analysis was performed to evaluate whether the relationship between falls and physical activity was mediated by falls efficacy and social support. A nonsignificant chi-square (χ2 = 18.9, p = .169) and the goodness of fit indicators (comparative fit index = .98, Tucker–Lewis index = .95, standardized root mean square residual = .03, and RMSEA = .07) revealed that this model had a good fit, with a reasonable number of degrees of freedom (df = 14). The magnitude of the direct effect, indirect effect, and total effect of each variable on the number of steps is shown in Table 3. The variable with the largest direct effect on physical activity was falls efficacy, followed by social support and sex, and the variable with the largest indirect effect was age, followed by falls and social support. The variable with the largest overall effect on physical activity was falls efficacy, followed by social support, falls, and age.
Direct Effect, Indirect Effect, and Total Effect of Predictor Variables on Physical Activity as Determined by the Structural Equation Model
| Direct effect | Indirect effect | Total effect | |
|---|---|---|---|
| Falls efficacy | 0.550 | – | 0.550 |
| Social support | 0.164 | 0.225 | 0.389 |
| Falls | −0.080 | −0.290 | −0.370 |
| Age | 0.080 | −0.455 | −0.365 |
| Sex | −0.155 | −0.017 | −0.172 |
| Household | 0.006 | 0.104 | 0.110 |
Discussion
The aim of this study was to examine the interrelation between fall history, physical activity (pedometer-determined ambulatory activity from steps per day), and the mediation effects of social support, falls efficacy, household, sex, and age. As far as we know, no other studies have tested the mediating role of social support and falls efficacy.
As expected, we found our sample of assisted-living facility residents to have lower levels of physical activity compared with their counterparts living in the community. The average number of steps in the group examined was 1,295 ± 786 steps/day. Levels of physical activity (percentiles) were similar in women who did not fall and women who fell at least once. However, compared with male nonfallers, male recurrent fallers had a considerably lower number of steps; these results correspond with reports that self-reported physical activity is curtailed after a fall (Tinetti & Williams, 1998). Only 27.3% of our sample scored average or above average levels of steps when compared with the normative step counts of free-living individuals aged between 60 and 84 from the National Health and Nutrition Examination Survey (Tudor-Locke et al., 2013), further indicating how inactive this population of older adults in assisted-living facilities is. These results are consistent with other studies that have reported a similar number of steps taken per day (1,345 ± 1,100 steps/day; Corcoran et al., 2016) for older adults residing in assisted-living facilities. The total effect of falls on physical activity occurred indirectly, mediated by falls efficacy and social support, with falls having a negative effect on physical activity. Although sex and household did not directly influence falls or falls efficacy, living alone did predict lower social support, which in turn predicted lower falls efficacy in older adults. Lower falls efficacy also predicted lower physical activity levels. Surprisingly, direct relationships between age and sex and social support or physical activity were not detected. The results showed an overall good fit between the proposed model and the data.
Among all variables, falls efficacy was the strongest direct factor related to physical activity among residents of assisted-living facilities. This was in agreement with previous investigations (Jefferis et al., 2014; McAuley, Mihalko, & Rosengren, 1997; Schepens, Sen, Painter, & Murphy, 2012; Schott, 2007) that showed falls efficacy to have a strong positive direct effect on physical activity. For example, in a recent investigation of a sample of independently living older men (n = 1,398; ages 71–93), Jefferis et al. (2014) showed a direct effect of falls efficacy on physical activity. By contrast, assisted-living older adults did not perceive falling as high of a risk, possibly due to the supervision or assistance of staff or due to limiting their physical activity to safer environments confined to the boundaries of the assisted-living facility (Moran et al., 2015). Sex and household only had a small effect on physical activity and falls efficacy, but age had a strong effect on falls efficacy. As explained by Bandura (1991), self-efficacy is an important component in determining decision making, the effort that older adults put into a task, and their stress when presented with a challenge. Self-efficacy is related to falls efficacy, which in turn is a predictor for future falls in community-dwelling older adults (Landers et al., 2016) but not, as we determined, a consequence of falls in the past. Our results support this view, because when our male fallers reported decreased falls efficacy, they were more likely to alter their behavior to avoid activities and situations that could cause falls, perhaps because they believe that if they do not, falls are unavoidable. Furthermore, the results of the current study were consistent with previous studies that reported that the oldest older adults exhibited lower falls efficacy than younger older adults (Kocić et al., 2017; Schott, 2014). In our study, we observed that older adults between the ages of 66 and 84 exhibited an obviously higher level of falls efficacy than individuals of ages 85 and older.
Although social support is a less commonly studied correlate of the relationship between falls and physical activity in older adults residing in assisted-living facilities, it is considered a key facilitator of physical activity (Durbin et al., 2016; Moran et al., 2015). Overall, we found a strong mediating effect for social support; it influences the relationship between falls and physical activity through falls efficacy. In the current study, falls history and falls efficacy were found to be strongly associated with social support with older adults who reported falls in the previous 12 months; they exhibited lower scores for emotional and tangible support as well as social integration. Furthermore, older adults who perceived higher availability of social support reported higher falls efficacy. These results differed from the work of Pin and Spini (2016) and Malini, Lourenço, and Lopes (2016). The difference could be explained by the operationalization of the construct of social support, whereas Pin and Spini (2016) asked “Now, please think about the past twelve months. Has any family member from outside the household, friend, or neighbor given you any type of help?” and Malini et al. (2016) asked “If you need or will need help performing any of these activities of daily living, do you have anyone to turn to?” and “Do you have a relative, friend, or neighbor who could take care of you if necessary?” In this study, we used a more elaborate construct of social support comprising emotional support, tangible support, and social integration. Based on our findings using the F-SozU, we argue that fallers are less able to actively integrate into a social network and therefore experience less emotional and tangible support as their network gets smaller (Loprinzi & Crush, 2018). Furthermore, our results support the deterioration model (Ensel & Lin, 1991), which postulates that stressors such as falls exhaust social support, and thus support cannot continuously suppress the effects of falls on physical activity. In a positive way of thinking, three mechanisms can conceivably account for the effect of social support on physical activity through falls efficacy: (a) physical activation through living a socially active lifestyle (e.g., leaving the room more often to meet friends), (b) cognitive stimulation through social interaction, and (c) positive emotions caused by perceived social support, which may increase falls efficacy and physical activity.
Limitations
The present study has some limitations. The first limitation is related to the study’s methodology. Because of the study’s cross-sectional design, the causal paths in the model are still based on hypothetical relationships. Hence, a longitudinal study is required to ascertain causal relationships among variables.
Second, another limitation of this study is the small sample size in relation to the complexity of the model (participants:parameters ratio; Bentler & Chou, 1987; Wolf, Harrington, Clark, & Miller, 2013). Although controversy exists surrounding sample sizes appropriate for use in structural equation modeling (Wolf et al., 2013), this study’s relatively small sample size of n = 81 meets previously published guidelines for conducting such analyses. Two of the most recent simulation studies empirically supported minimum sample sizes ranging from 30 to 80 (Sideridis, Simos, Papanicolaou, & Fletcher, 2014; Wolf et al., 2013). This is because sample size adequacy depends not only on model complexity but also on many other factors, such as missing data, statistical power, bias in parameter estimates, and overall solution propriety. Because our model was not very complex (only nine observed variables), the results are interpretable.
Third, this study assessed physical activity by means of the pedometer (NL-2000; New LifeStyles, MT). The accuracy of the specific pedometer model has already been demonstrated (Schneider, Crouter, Lukajic, & Bassett Jr., 2003; Schneider et al., 2004). However, some studies have reported reduced accuracy for slower walking (Beevi, Miranda, Pedersen, & Wagner, 2016). This could have affected physical activity measurement in the sample of older adults living in assisted-living facilities. Therefore, an underestimation of actual steps cannot be excluded and step totals should be interpreted with caution. However, it may be assumed that although a measurement error of this nature would affect the absolute level of physical activity, it would not affect our questions addressing the effects of falls on physical activity mediated by falls efficacy and social support. Although the total physical activity levels reported in this study should therefore be interpreted with caution, the diurnal profiles of physical activity may be regarded as a true reflection of older adults’ daily physical activity rhythm.
Fourth, we only assessed perceived social support, not actual social support, and no differentiation between qualitative and quantitative support was made. In the setting of this study, it could be interesting to assess social contact with other older adults and with representatives/caregivers from the health system. In addition, sources of support could be interesting to look at, especially because females have a lesser probability of living in a household with a companion requiring other support givers to step into place. In this study, we differentiated between emotional support, tangible support, and social integration; nevertheless, it could be interesting to examine which kind of social support should be fostered. These factors could help to draw a clearer picture of an effective social support profile for older people in order to facilitate well-being and health. This is especially true bearing in mind that with age, social network decline, and a pool of social support is needed to prevent loneliness and is an indicator for being institutionalized (Hoogendijk et al., 2016).
Fifth, in this study, we looked at a very specific sample—participants residing in assisted-living facilities—which makes it difficult to compare the results to other studies, which looked, for example, at older adults receiving home care. The transition between different living conditions is somewhat blurred but has an impact on active living possibilities. Future studies need to differentiate between living conditions in order to identify specific intervention steps and actions for each setting.
Sixth, we have only examined the impact of falls on physical activity; however, it seems necessary to also look at the predictive impact of social support and falls efficacy not only on physical activity but also on falls. Future studies need to look within a reciprocal effect model at the relationship between fall history and physical activity as well as social support and falls efficacy. Fall history has an effect on self-perception, which influences physical activity; by contrast, motor competence itself has an impact on the number of falls. Longitudinal studies that cover more than a couple of weeks are necessary in order to investigate the interaction between these multiple variables.
Conclusion
This is the first study that has looked at the relationship between falls history, social support, falls efficacy, and physical activity in an assisted-living facility and tested the mediating effect of sex, age, social support, and falls efficacy on the relationship between falls history and physical activity. The development of a structural model illustrating the mediating effects of social support and falls efficacy on the relationship between falls and physical activity will help health care professionals to predict risk factors of falls that can be compromised by residing in an assisted-living facility. This understanding provides support for creating tailored intervention programs in such settings.
Acknowledgments
The authors would like to acknowledge the work of Lisa Ahl, who collected the data as part of her master’s thesis at the University of Münster.
References
Aghdam, F.B., Moghaddam, M.H.B., Jafarabadi, M.A., Allahverdipour, H., Nikookheslat, S.D., & Noorizadeh, R. (2013). Explaining the role of personal, social and physical environment factors on employed women’s physical activity: A structural equation analysis. Global Journal of Health Science, 5(4), 189–199. PubMed ID: 23777735 doi:10.5539/gjhs.v5n4p189
Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational behavior and Human Decision Processes, 50, 248–287.
Beevi, F.H.A., Miranda, J., Pedersen, C.F., & Wagner, S. (2016). An evaluation of commercial pedometers for monitoring slow walking speed populations. Telemedicine Journal and e-Health, 22, 441–449. PubMed ID: 26451900 doi:10.1089/tmj.2015.0120
Bentler, P.M., & Chou, C.H. (1987). Practical issues in structural equation modeling. Sociological Methods & Research, 16, 78–117. doi:10.1177/0049124187016001004
Büchele, G., Becker, C., Cameron, I.D., König, H.H., Robinovitch, S., & Rapp, K. (2014). Predictors of serious consequences of falls in residential aged care: Analysis of more than 70,000 falls from residents of Bavarian nursing homes. Journal of American Medical Directors Association, 15, 559–563. doi:10.1016/j.jamda.2014.03.015
Chou, C., & Bentler, P.M. (1995). Estimates and test in structural equation modeling. In R.H. Hoyle (Ed.), Structural equation modeling: Concepts, issues and applications (pp. 37–59). Thousand Oaks, CA: Sage.
Corcoran, M.P., Chui, K.K., White, D.K., Reid, K.F., Kirn, D., Nelson, M.E., . . . Fielding, R.A. (2016). Accelerometer assessment of physical activity and its association with physical function in older adults residing at assisted care facilities. Journal of Nutrition, Health & Aging, 20(7), 752–758. PubMed ID: 27499309 doi:10.1007/s12603-015-0640-7
Cumming, R.G., Salkeld, G., Thomas, M., & Szonyi, G. (2000). Prospective study of the impact of fear of falling on activities of daily living, SF-36 scores, and nursing home admission. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 55(5), 299–305. doi:10.1093/gerona/55.5.M299
Delbaere, K., Crombez, G., van Haastregt, J.C.M., & Vlaeyen, J.W.S. (2009). Falls and catastrophic thoughts about falls predict mobility restriction in community-dwelling older people: A structural equation modelling approach. Aging & Mental Health, 13(4), 587–592. PubMed ID: 19629784 doi:10.1080/13607860902774444
Durbin, L., Kharrazi, R.J., Graber, R., & Mielenz, T.J. (2016). Social support and older adult falls. Injury Epidemiology, 3(1), 4. PubMed ID: 27747541 doi:10.1186/s40621-016-0070-y
Ensel, W.M., & Lin, N. (1991). The life stress paradigm and psychological distress. Journal of Health and Social Behavior, 32, 321–341. PubMed ID: 1765624 doi:10.2307/2137101
Franco, M.R., Pereira, L.S., & Ferreira, P.H. (2014). Exercise interventions for preventing falls in older people living in the community. British Journal of Sports Medicine, 48, 867–868. PubMed ID: 23314888 doi:10.1136/bjsports-2012-092065
Fydrich, T., Geyer, M., Hessel, A., Sommer, G., & Brähler, E. (1999). Fragebogen zur Sozialen Unterstützung (F-SozU): Normierung an einer repräsentativen Stichprobe. Diagnostica, 45, 212–216. doi:10.1026//0012-1924.45.4.212
Gillespie, L.D., Robertson, M.C., Gillespie, W.J., Sherrington, C., Gates, S., Clemson, L.M., & Lamb, S.E. (2012). Interventions for preventing falls in older people living in the community. Cochrane Database Systematic Review, 9, CD007146.
Gimm, G.W., & Kitsantas, P. (2016). Falls, depression, and other hospitalization risk factors for adults in residential care facilities. International Journal of Aging and Human Development, 83, 44–62. PubMed ID: 27147680 doi:10.1177/0091415016645347
Graafmans, W.C., Lips, P., Wijlhuizen, G.J., Pluijm, S.M., & Bouter, L.M. (2003). Daily physical activity and the use of a walking aid in relation to falls in elderly people in a residential care setting. Zeitschrift fur Gerontologie und Geriatrie, 36(1), 23–28. PubMed ID: 12616404 doi:10.1007/s00391-003-0143-8
Grant, P.M., Dall, P.M., Mitchell, S.L., & Granat, M.H. (2008). Activity-monitor accuracy in measuring step number and cadence in community-dwelling older adults. Journal of Aging and Physical Activity, 16(2), 201–214. PubMed ID: 18483442 doi:10.1123/japa.16.2.201
Hall, K.S., & McAuley, E. (2011). Examining indirect associations between physical activity, function, and disability in independent- and assisted-living residents. Journal of Physical Activity & Health, 8, 716–723. PubMed ID: 21734318 doi:10.1123/jpah.8.5.716
Hoang, O.T.T., Jullamate, P., Piphatvanitcha, N., & Rosenberg, E. (2017). Factors related to fear of falling among community-dwelling older adults. Journal of Clinical Nursing, 26, 68–76. PubMed ID: 27723217 doi:10.1111/jocn.13337
Hoogendijk, E.O., Deeg, D.J., Poppelaars, J., van der Horst, M., Broese van Groenou, M.I., Comijs, H.C., . . . Huisman, M. (2016). The Longitudinal Aging Study Amsterdam: Cohort update 2016 and major findings. European Journal of Epidemiology, 31, 927–945. PubMed ID: 27544533 doi:10.1007/s10654-016-0192-0
Hu, L., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1–55. doi:10.1080/10705519909540118
Hughes, C.C., Kneebone, I.I., Jones, F., & Brady, B. (2015). A theoretical and empirical review of psychological factors associated with falls-related psychological concerns in community-dwelling older people. International Psychogeriatrics, 27, 1071–1087. PubMed ID: 25633917 doi:10.1017/S1041610214002701
Ishii, K., Shibata, A., & Oka, K. (2010). Environmental, psychological, and social influences on physical activity among Japanese adults: Structural equation modeling analysis. The International Journal of Behavioral Nutrition and Physical Activity, 7, 61. doi:10.1186/1479-5868-7-61
Jefferis, B.J., Iliffe, S., Kendrick, D., Kerse, N., Trost, S., Lennon, L.T., . . . Whincup, P.H. (2014). How are falls and fear of falling associated with objectively measured physical activity in a cohort of community-dwelling older men? BMC Geriatrics, 14, 114. PubMed ID: 25348492 doi:10.1186/1471-2318-14-114
Klingelhöfer-Noe, J., Dassen, T., & Lahmann, N. (2015). Vollstationäre Pflegeeinrichtungen vs. „betreutes Wohnen mit ambulanter Versorgung“. Zeitschrift für Gerontologie & Geriatrie, 48, 263–269. doi:10.1007/s00391-014-0655-4
Kocić, M., Stojanovic, Z., Lazović, M., Nikolić, D.V., Zivkovic, V.D., Milenkovic, M.R., & Lazarević, K.K. (2017). Relationship between fear of falling and functional status in nursing home residents aged older than 65 years. Geriatrics & Gerontology International, 17(10), 1470–1476. doi:10.1111/ggi.12897
Landers, M.R., Oscar, S, Sasaoka, J., & Vaughn, K. (2016). Balance confidence and fear of falling avoidance behavior are most predictive of falling in older adults: Prospective analysis. Physical Therapy, 96(4), 433–442. doi:10.2522/ptj.20150184
Leung, P.-M., Ejupi, A., van Schooten, K.S., Aziz, O., Feldman, F., Mackey, D.C., . . . Robinovitch, S.N. (2017). Association between sedentary behaviour and physical, cognitive, and psychosocial status among older adults in assisted living. BioMed Research International, 2017, 1–7. PubMed ID: 28913360 doi:10.1155/2017/9160504
Loprinzi, P.D., & Crush, E.A. (2018). Source and size of social support network on sedentary behavior among older adults. American Journal of Health Promotion, 32(1), 28–31. PubMed ID: 29214834 doi:10.1177/0890117116686888
MacCallum, R.C., Browne, M.W., & Sugawara, H.M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130–149.
Malini, F.M., Lourenço, R.A., & Lopes, C.S. (2016). Prevalence of fear of falling in older adults, and its associations with clinical, functional and psychosocial factors: The Frailty in Brazilian Older People—Rio de Janeiro Study. Geriatrics & Gerontology International, 16, 336–344. PubMed ID: 25869919 doi:10.1111/ggi.12477
McAuley, E., Mihalko, S.L., & Rosengren, K.S. (1997). Self-efficacy and balance correlates of fear of falling in the elderly. Journal of Aging and Physical Activity, 5, 329–340. doi:10.1123/japa.5.4.329
Moran, F., MacMillan, F., Smith-Merry, J., Kilbreath, S., & Merom, D. (2015). Perceived barriers, facilitators and patterns of physical activity of older old adults living in assisted retirement accommodation: A qualitative and quantitative pilot research. Journal of Gerontology & Geriatric Research, 4, 255. doi:10.4172/2167-7182.1000255
Nitz, J.C., & Josephson, D.L. (2011). Enhancing functional balance and mobility among older people living in long-term care facilities. Geriatrics Nursing, 32(2), 106–113. doi:10.1016/j.gerinurse.2010.11.004
Pin, S., & Spini, D. (2016). Impact of falling on social participation and social support trajectories in a middle-aged and elderly European sample. SSM - Population Health, 2, 382–389. doi:10.1016/j.ssmph.2016.05.004
Powell, L., & Myers, A. (1995). The activities-specific balance (ABC) scale. Journal of Gerontology, Series A: Biological Sciences and Medical Sciences, 50A, M28–M34.
Rapp, K., Freiberger, E., Todd, C., Klenk, J., Becker, C., Denkinger, M., . . . Fuchs, J. (2014). Fall incidence in Germany: Results of two population-based studies, and comparison of retrospective and prospective falls data collection methods. BMC Geriatrics, 14, 105. PubMed ID: 25241278 doi:10.1186/1471-2318-14-105
Rubenstein, L., & Josephson, K. (2002). The epidemiology of falls and syncope. Clinics in Geriatric Medicine, 18, 141–158. doi:10.1016/S0749-0690(02)00002-2
Ruthig, J.C., Chipperfield, J.G., Newall, N.E., Perry, R.P., & Hall, N.C. (2007). Detrimental effects of falling on health and well-being in later life: The mediating roles of perceived control and optimism. Journal of Health Psychology, 12(2), 231–248. doi:10.1177/1359105307074250
Schepens, S., Sen, A., Painter, J.A., & Murphy, S.L. (2012). Relationship between fall-related efficacy and activity engagement in community-dwelling older adults: A meta-analytic review. The American Journal of Occupational Therapy, 66(2), 137–148. PubMed ID: 22394523 doi:10.5014/ajot.2012.001156
Schneider, P.L., Crouter, S.E., & Bassett, D.R., Jr. (2004). Pedometer measures of free-living physical activity: Comparison of 13 models. Medicine & Science in Sports & Exercise, 36(2), 331–335. PubMed ID: 14767259 doi:10.1249/01.MSS.0000113486.60548.E9
Schneider, P.L., Crouter, S.E., Lukajic, O., & Bassett, D.R., Jr. (2003). Accuracy and reliability of 10 pedometers for measuring steps over a 400-m walk. Medicine & Science in Sports & Exercise, 35, 1779–1784. PubMed ID: 14523320 doi:10.1249/01.MSS.0000089342.96098.C4
Schott, N. (2007). Korrelate der Sturzangst im Alter. Zeitschrift für Sportpsychologie, 14(2), 74–86. doi:10.1026/1612-5010.14.2.74
Schott, N. (2014). Reliability and validity of the German short version of the Activities specific Balance Confidence (ABC-D6) scale in older adults. Archives of Gerontology and Geriatrics, 59(2), 272–279. PubMed ID: 24962236 doi:10.1016/j.archger.2014.05.003
Sideridis, G., Simos, P., Papanicolaou, A., & Fletcher, J. (2014). Using structural equation modeling to assess functional connectivity in the brain: Power and sample size considerations. Educational and Psychological Measurement, 74, 733–758. PubMed ID: 25435589 doi:10.1177/0013164414525397
Smith, G.L., Banting, L., Eime, R., O’sullivan, G., & van Uffelen, J.G.Z. (2017). The association between social support and physical activity in older adults: A systematic review. International Journal of Behavioral Nutrition and Physical Activity, 14, 56. doi:10.1186/s12966-017-0509-8
Stenhagen, M., Ekström, H., Nordell, E., & Elmståhl, S. (2014). Both deterioration and improvement in activities of daily living are related to falls: A 6-year follow-up of the general elderly population study Good Aging in Skåne. Clinical Interventions in Aging, 9, 1839–1846. PubMed ID: 25378916 doi:10.2147/CIA.S70075
Stock, K.J., Amuah, J.E., Lapane, K.L., Hogan, D.B., & Maxwell, C.J. (2017). Prevalence of, and resident and facility characteristics associated with antipsychotic use in assisted living vs. long-term care facilities: A cross-sectional analysis from Alberta, Canada. Drugs & Aging, 34(1), 39–53. PubMed ID: 27830567 doi:10.1007/s40266-016-0411-0
Tinetti, M.E., & Williams, C.S. (1998). The effect of falls and fall injuries on functioning in community-dwelling older persons. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 53, M112–M119. doi:10.1093/gerona/53A.2.M112
Towne, S.D., Jr., Cho, J., Smith, M.L., & Ory, M.G. (2017). Factors associated with injurious falls in residential care facilities. Journal of Aging Health, 29(4), 669–687. PubMed ID: 27107006 doi:10.1177/0898264316641083
Towne, S.D., Jr., Lee, S., Li, Y., & Smith, M.L. (2016). Assessment of eHealth capabilities and utilization in residential care settings. Health Informatics Journal, 22(4), 1063–1075. PubMed ID: 26512032 doi:10.1177/1460458215610895
Tudor-Locke, C., Schuna, J.M., Jr., Barreira, T.V., Mire, E.F., Broyles, S.T., Katzmarzyk, P.T., & Johnson, W.D. (2013). Normative steps/day values for older adults: NHANES 2005–2006. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 68(11), 1426–1432. doi:10.1093/gerona/glt116
Wolf, E.J., Harrington, K.M., Clark, S.L., & Miller, M.W. (2013). Sample size requirements for structural equation models: An evaluation of power, bias, and solution propriety. Educational and Psychological Measurement, 73(6), 913–934. doi:10.1177/0013164413495237
Wolff, J.K., Lindenberger, U., Brose, A., & Schmiedek, F. (2016). Is available support always helpful for older adults? Exploring the buffering effects of state and trait social support. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 71, 23–34. PubMed ID: 25134455 doi:10.1093/geronb/gbu085
