Abstract
Objectives
We examined the effects of a walking intervention in older adults residing in long-term care (LTC) homes on gait velocity (primary outcome), and stride length, cadence and heel-to-heel base of support (secondary outcomes) compared to those in an interpersonal interaction control group and a care-as-usual control group at 16-weeks post-intervention.
Methods
These previously unpublished gait data were collected as part of a larger prospective, randomized, three group study. One hundred and sixty-eight participants residing in 12 LTC homes were randomized into: a) a walking group (n=57) — 1:1 supervised, individualized, progressive, 30 minutes, five times a week walking program for 16 weeks; b) an interpersonal interaction group (n=55) — stationary 1:1 conversation time with research personnel; and, c) a care-as-usual control group (n=56). Gait was assessed at baseline and 16-weeks post-intervention using the GAITRite® computerized system. One-way Analysis of Covariance (ANCOVA), controlling for age, sex, cognitive status and baseline gait parameter (velocity, stride length, cadence, heel-to-heel base of support) was used to examine differences among groups for velocity, stride length, cadence, and heel-to-heel base of support at 16-weeks post-intervention.
Results
Ninety-one participants with available data were included in this analysis: walking group (n=31/57, mean age=82.77±6.75 years); interpersonal interaction group (n=31/55, mean age=82.74±9.27 years); care-as-usual control group (n=29/56, mean age=85.40±8.78 years). ANCOVA showed a significant difference in the mean gait velocity at 16-weeks post-intervention [F(2, 84) =6.99, p=0.0006); η2 (95%CI)=0.16 (0.02, 0.27)]. Post hoc comparisons using Sidak test showed that the estimated marginal mean (EMM) for velocity for the walking group [EMM (SE), 0.51m/s (0.03)] was significantly higher compared to the interpersonal interaction group [EMM (SE), 0.38m/s (0.03); t(83)=3.15, p=0.007] and the care-as-usual control group [EMM (SE), 0.38m/s (0.03)]; t(83)=3.32, p=0.004]. No significant difference was observed between groups for stride length, cadence or heel-to-heel base of support.
Conclusion
LTC residents with limited physical functioning showed significant improvement in gait velocity but not in stride length, cadence or heel-to-heel base of support after a 16-week walking intervention.
Key words: Gait, walking, older adults, long-term care
Introduction
Physical activity is of central importance for healthy and active aging, and has been associated with multiple health benefits among long term care (LTC) residents (1). Conversely, the lack of physical activity has several detrimental effects on LTC residents, including negative effects on cognition (2) and physical functioning (3). Despite this, LTC residents are often sedentary for most of the day (4, 5).
Exercise interventions aimed at increasing endurance, balance, flexibility and functional status have provided significant outcomes, including improved physical function, mobility, activity of daily life performance, psychological health, and quality of life, for older adults living in LTC settings (1). However, LTC residents living with severe cognitive impairments and mobility limitations are often not capable of participating in formal exercise interventions because of the complexity of some exercise programs (6). Due to the diversity of the functional health and cognitive status of older adults living in LTC settings, exercise instructors often struggle to adapt exercises to address individual needs (7). Therefore, there is a need to develop a simple physical activity intervention that can facilitate maximum participation for older adults living in LTC settings. Walking has been described as a simple physical activity intervention that could improve mobility and functional independence in institutionalized older adults (1). A walking intervention can be delivered individually or in a small group and can be easily integrated into daily care routines; thus, walking is considered an ideal intervention to reduce sedentary behavior among older adults living in LTC settings (8).
Studies have reported on the effect of walking and talking interventions in older adults with Alzheimer's disease residing in nursing homes (9, 10). Tappen et al. (9) used a pre-test and post-test study design to compare three groups: (a) walking — 30 minutes of self-paced (the participant walked at a pace of his or her choosing for 30 minutes), assisted walking, three times per week for 16 weeks; (b) talking — 30 minutes of conversation, and (c) walking and talking simultaneously within 30-minute sessions. They reported a 20.7% decline in the distance walked in six minutes in the walking group, an 18.8% decline in the talking group and a 2.5% decline in the combined group. Cott et al.‘s (10) three-arm study of older adults that participated in walking and talking in pairs (30 minutes, five times per week for 16 weeks), talking only in pairs (30 minutes, five times per week for 16 weeks) and no intervention found no significant difference in communication, ambulation or functional status among the groups. MacRae et al.‘s study (11) determined the effects of 12 versus 22 weeks of walk training on walking endurance capacity, physical activity level, mobility and quality of life in ambulatory nursing home residents. They found significant improvements in walking endurance capacity but not in walking speed at 12 weeks and no significant changes were found from 12 to 22 weeks in either walking endurance capacity or walking speed.
Reduced gait velocity, stride length (12) and cadence (13) have been reported to be predictors of falls among LTC residents. Falls among older adults living in the LTC setting often lead to fractures (14). Even when there is no severe injury, resultant fear of falling and self- or facility-imposed mobility restrictions can lead to physical inactivity (14). While published studies to date have provided useful information regarding the gait parameters that predict falls in older adults in LTC, the effect of a walking intervention, a simple and easy to implement physical activity intervention, on gait speed is equivocal and has not yet been established on other gait parameters (e.g., stride length and cadence) known to predict falls among LTC residents.
In summary, the effects of walking interventions on outcomes among LTC residents are equivocal. Only MacRae et al.‘s study (11) has evaluated the effects of walking on gait speed among LTC residents; however, their participants’ gait speed was calculated from the participants' maximum walk time performed during a single day of walking, which is prone to human error. Our study addressed the previous studies' methodological limitations by exploring gait parameters (e.g., gait speed) using a computerized system-GAITRite®. This paper examined the effects of a walking intervention for older adults residing in LTC home on gait velocity (primary outcome), and stride length, cadence and heel-to-heel base of support (secondary outcomes) compared to those in the interpersonal interaction group and care-as-usual control group at 16-weeks post-intervention.
Methods
Study design
The data for this study were collected as part of a larger project that involved a 32-week, prospective, randomized, three-group [walking, interpersonal interaction and care-as-usual control group] experimental design that investigated the effects of a 16-week walking intervention on injurious falls and other physical and psychological measures (15, 16). The study methods and design are published in detail elsewhere (15, 16). A central methods centre was not utilized for randomization, but allocation was concealed prior to random assignment. Participants within each LTC facility were randomly assigned to one of the three groups. Random assignment was not blocked, nor was random assignment stratified by facility or demographic variable e.g., sex, age. Research assistants administering the outcome measures and collecting the data were blinded to group assignment (15, 16).
The study was approved by the appropriate institutional research ethics board University of Saskatchewan Biomedical Research Ethics Board (BIO#10-125), the individual LTC facilities, and registered in the clinical trial registry [NCT01277809].
Target population, inclusion and recruitment
Participants from 12 LTC homes located in a mid-size metropolitan area were recruited. In each facility, nursing staff identified potential participants, and the Research Assistant (RA) approached them with a detailed explanation of the study and inquired if they wished to participate. Participants provided consent themselves or expressed assent to participate along with consent from their substitute decision-maker who provided consent (15).
Inclusion criteria included: age ≥60 years, residing in a LTC home, able to follow simple instructions, able to ambulate with or without a walking aid for at least 10 meters, and available for the stipulated days for the research — 5 days per week (Monday to Friday) over four months. Participants were excluded if they experienced a recent cardiovascular event (within the past six months) or fracture (within the past four months); had severe mobility-limiting arthritis, a vestibular disorder, uncontrolled hypertension, uncontrolled epilepsy, or an acute care admission within the past four months. Participants scheduled for surgery or hospitalization within the next six months and those participating in another regular exercise program or who exercised independently for half an hour or more, three or more times per week were also excluded (15, 16).
Intervention procedures
Participants (n=168) were randomly assigned into one of three groups — walking (intervention group), interpersonal interaction (socialization control group), or care-as-usual (control group) (16). Participants in the walking group received an individualized, progressive, 1:1, daily supervised walking program for 16 weeks, lasting for up to half an hour, five days per week (Monday to Friday), facilitated by RAs and supervised by a physiotherapist. The participant walked at a pace of his or her choosing, and the distance walked was initially determined for each individual based on how far the person could walk before visibly becoming fatigued or short of breath, reporting pain or requesting to sit down or rest. The walking pace and walking distance were determined by the participant in each session, and were gradually increased over the 16 weeks as tolerated by the participant. The RA recorded the overall time spent walking each day of the intervention, and the distance walked. Participants in the interpersonal interaction group received approximately 30 minutes of one-on-one interaction time per day with an RA, five days a week, for 16 weeks. Participants were stationary and were either engaged in conversations, picture viewing or playing board games. The interaction time was recorded. The control group participants received care-as-usual (e.g., usual non-physically demanding recreational activities such as bingo, intergenerational visits and dog visits) administered by the LTC staff caring for the individual, including recreational therapists, personal support workers and nursing staff (15). During the second 16-week time period, the control group continued with care-as-usual, and the walking and interpersonal interaction group participants received no active interventions. Participants in all groups completed scheduled assessments at 8 weeks and 16 weeks after completion of the intervention period.
Licensed physiotherapists were available to the RAs during the study to advise on practical and safety issues related to the intervention. The RAs were trained for 3 hours in the administration of the physical assessments by a physiotherapist.
Data Collection
A description of the outcome measures used in the original study has been published in the study protocol (15). All outcome measures reported in the original study (15) were completed at baseline, 8-weeks (mid-trial) and 16-weeks (post-baseline), and at the post-intervention follow-up assessments at 8-weeks and 16-weeks e.g., assessments after the 16-week active intervention period was completed. Participants' demographics, including age, sex, and medication history (e.g., antidepressant and anti-psychotic use) were collected. Information about adverse health events for any reason, hospitalizations, and death (date and cause when available) were collected from the medical record. The following outcome measures were also collected in the original study, and were used in this study to describe the sample at baseline: Geriatric Depression Scale (GDS) (17), Short Form, Short Portable Mental Status Questionnaire (SPMSQ) (18), Six-minute Walk Test (6MWT) (19), Berg Balance Scale (BBS) (20), and Eight-Foot Up and Go Test (21).
For this paper, of primary interest were gait parameters collected using the GAITRite® system (CIR Systems Inc., Clifton, NJ), a portable electronic walkway system 6 meters (m) in length and 0.64m in width, with embedded sensors. The data collected by this system have not been reported previously for this study. The GAITRite® system is valid and reliable for measuring temporal and spatial gait parameters, permits parallel videotaping of ambulation through a video camera interface (22, 23), and has been used with LTC residents (12). Participants were asked to walk at their usual walking speed over 6 m on the sensor mat of the GAITRite® system. Participants' acceleration and deceleration were accounted for during the testing by adding an extra two meters of walking at both the start and endpoints of the GAITRite® system. Each participant undertook three trials on the GAITRite® mat, with a two-minute rest in between each trial. The first trial was a practice trial, and participants' gait velocity, the primary outcome (meters/second, m/s), was calculated by averaging the last two trials. The secondary outcomes were stride length (cm), cadence (number of steps/minute) and heel-to-heel base of support (cm). Heel-to-heel base of support is defined as the vertical distance from the heel center of one footprint to the line of progression formed by two footprints of the opposite foot (24). GaitRite® data, collected in cm or cm/s, were converted to m or m/s. These variables were selected because a decrease in gait velocity, stride length, cadence and heel-to-heel base of support are predictors for falls among LTC residents (12, 13). Gait assessments were completed at baseline and 16-weeks post-intervention.
Research assistants recorded if and when consent or assent was withdrawn by participants or if and when contraindications to the study protocol developed.
Data analysis
Descriptive analysis, with means and standard deviations for continuous variables and frequency and percentage for non-continuous variables, was conducted. Little's test (25) was used to test the mechanism of missing data. All continuous data were initially checked for normality prior to analysis. Non-normally distributed data were log-transformed for ANOVA or ANCOVA. Baseline differences were compared between groups with ANOVA for continuous variables and normally distributed data or Kruskal-Wallis analysis of variance for categorical variables. Four one-way ANCOVAs were used to examine differences between groups for velocity, stride length, cadence, and heel-to-heel base of support at 16-weeks post-intervention. Age, sex, cognitive status, baseline gait parameter (velocity, stride length, cadence, heel-to-heel base of support) were used as covariates and groups (walking, interpersonal interaction and care-as-usual control) as the fixed factor. Sidak's multiple-comparison test was used post-hoc to examine means when the F-ratio was significant. Effect size, a proportion of variance that a variable explains that is not explained by other variables in the analysis (26), was reported using partial eta-squared (η2).
To further explore our findings for gait velocity (27), we conducted a sub-group analysis using two different velocity cut-offs (< 0.50m/s and ≥ 0.50m/s). A gait velocity of 0.5m/s represents the usual pace of gait velocity noted among LTC residents in a systematic review (28). Data were analyzed using STATA/IC (v14), with a p-value for significance set at <0.05.
Results
Figure 1 presents the CONSORT flow diagram of the original RCT and the present analysis, and the reported reasons for loss to follow-up. Age, sex, 8-Foot Up and Go time, SPMSQ, 6MWT distance, GDS, BBS scores, baseline scores of velocity, stride length, cadence, and heel-to-heel base of support did not differ between participants who had no missing values and those with missing data values. Little's test further indicated that data were missing completely at random (MCAR), [χ2(4, n=168) = 1.27, p= 0.866]. The proportion of missing data was large (45.8%), and as a result, we did not perform imputation (29). We present only available case analysis (n=91).
Figure 1.
Consort diagram of original randomized control trial and secondary analysis
Demographics and baseline characteristics
Ninety-one participants, 31 males and 60 females, ranging in age from 60 to 96 years, had gait parameter variables available for analysis: walking group (n=31), interpersonal interaction group (n=31), care-as-usual group (n=29). Participants had moderate cognitive impairment as indicated by the SPMSQ, had limited physical functioning as indicated by 8-foot Up and Go Test, had limited functional capacity as measured by 6MWT and were at high risk of falls as indicated by BBS (Table 1). Participants had no significant depressive symptoms as measured by GDS.
Table 1.
Baseline differences of participants by group
| Variables | Walking group (n=31) | Interpersonal interaction group (n=31) | Control group (n=29) | Total (n=91) | Baseline differences |
|---|---|---|---|---|---|
| Age (years), mean (SD) | 82.77 (6.75) | 82.74 (9.27) | 82.40 (8.78) | 82.20 (8.23) | F(2,88)=0.02, p=0.98 |
| Female, n (%) | 22 (70.67) | 17 (54.84) | 21 (72.41) | 60 (66.00) | χ2(2)=2.59, p=0.27 |
| GDS, mean (SD) | 3.59 (2.92) | 3.29 (3.35) | 3.28 (3.01) | 3.39 (3.07) | F(2,88)=0.09, p=0.92 |
| Six-minute Walk (meters), mean (SD) | 113.85 (89.02) | 109.59 (89.34) | 87.94 (71.27) | 104.15(83.72) | F(2,88)=0.82, p=0.45 |
| 8-Foot Up and Go Test (sec), mean (SD) | 45.43 (28.67) | 54.26 (36.63) | 60.93 (52.36) | 53.50 (40.36) | F(2,88)=1.10, p=0.34 |
| SPMSQ, mean (SD) | 7.07 (2.37) | 5.10 (2.47) | 6.28 (2.53) | 6.14 (2.56) | F(2,88)=5.04, p=0.0085* |
| Berg Balance Scale, mean (SD) | 31.18 (12.66) | 27.68 (15.53) | 24.93 (16.15) | 28.00 (14.88) | F(2,88)=1.34, p=0.27 |
Berg Balance Scale [scores range from 0–54]; higher score indicates better balance (20). GDS: Geriatric Depression Scale [scores range from 0–15]; Scores of 0–4 are considered normal, depending on age, education, and complaints; 5–8 indicate mild depression; 9–11 indicate moderate depression; and 12–15 indicate severe depression (17)]. SPMSQ: Short Portable Mental Status Questionnaire [scores range from 0–10]; 0–2 errors: normal mental functioning 3–4 errors: mild cognitive impairment 5–7 errors: moderate cognitive impairment 8 or more errors: severe cognitive impairment higher score indicates poorer cognitive impairment (18); *= significant difference between interpersonal interaction group and walking group. Sidak test indicated mean SPMSQ score for the walking group was significantly higher than the interpersonal interaction group, t(88)=3.16, p=0.007; but no difference between walking vs control or control vs interpersonal interaction group. F-ratio test is ANOVA and χ2 is Kruskal Wallis test.
The walking group's total distance walked over the 16-weeks of walking intervention ranged from 2909.55 m to 63,538.20 m, with a mean of 6684.85 m (SD=13,246.81 m). The mean gait velocity (SD) for all participants at baseline was 0.40 m/s (SD=0.24; 95% Cl=0.35, 0.45]. At baseline, there was no significant difference in age, sex, 8-foot Up and Go time, GDS and BBS scores, gait velocity, cadence, stride length, and heel-to-heel base of support between groups (See Appendix). There was a significant difference in SPMSQ score between groups, with higher scores (more impairment) in the walking group compared to the interpersonal interaction group (See Table 1).
Changes in gait parameters at 16 weeks post-intervention
After adjusting for age, sex, SPMSQ score and baseline velocity, the ANCOVA showed a significant difference in the mean gait velocity at 16 weeks post-intervention [F(2,84)=8.12, p=0.0006; η2 (95% CI)= 0.16 (0.02, 0.27)].
Means, standard deviation (SD), 95% Confidence Interval (CI) for gait velocity, stride length, cadence and heel-to-heel base at baseline and 16 weeks post-intervention, and within group change findings are reported in Table 2. There were no significant differences in mean stride length (F(2,84)=0.46, p=0.630), cadence (F(2,84)=0.77, p=0.445) and heel-to-heel base of support (F(2,84)=0.63, p=0.537) at 16 weeks post-intervention across the groups. Table 3 reports the complete case post-hoc analysis for gait velocity and sub-group analysis for participants with a baseline velocity <0.5m/s. Post hoc comparison using Sidak test showed that the estimated marginal mean (EMM) for gait velocity for the walking group [EMM (SE), 0.51m/s (0.03)] was significantly higher compared to the interpersonal interaction group [EMM (SE), 0.38m/s (0.03); t(83)=3.15, p=0.007] and the care-as-usual control group [EMM (SE), 0.38m/s (0.03)]; t(83)=3.32, p=0.004].
Table 2.
Gait Parameter means (SD, 95%Cl) at baseline, 16 weeks post-intervention, and within group change
| Gait parameter | Group | Baseline | 16-weeks post-intervention | Within group change | |||
|---|---|---|---|---|---|---|---|
| Mean (SD) | 95%CI | Mean (SD) | 95%CI | Mean (SD) | 95%CI | ||
| Velocity (m/s) | Walking | 0.42 (0.26) | 0.32, 0.51 | 0.52 (0.22) | 0.44, 0.60 | 0.10 (0.19) | 0.03, 0.17 |
| Interpersonal Interaction | 0.38 (0.24) | 0.30, 0.47 | 0.37 (0.24) | 0.28, 0.46 | −0.01 (0.12) | −0.06, 0.03 | |
| Control | 0.40 (0.21) | 0.31, 0.48 | 0.38 (0.16) | 0.32, 0.44 | −0.02 (0.17) | −0.08, 0.05 | |
| Stride length (m) | Walking | 0.69 (0.26) | 0.56, 0.75 | 0.68 (0.30) | 0.57, 0.79 | 0.02 (0.14) | −0.03, 0.07 |
| Interpersonal interaction | 0.65 (0.30) | 0.55, 0.74 | 0.60 (0.25) | 0.51, 0.69 | −0.53 (0.24) | −0.14, 0.03 | |
| Control | 0.63 (0.25) | 0.53, 0.73 | 0.58 (0.17) | 0.51, 0.69 | −0.05 (0.21) | −0.13, 0.03 | |
| Cadence (steps/min) | Walking | 78.30 (22.39) | 70.08, 86.51 | 77.98 (22.75) | 69.64, 86.33 | −0.31 (21.37) | −8.15, 7.53 |
| Interpersonal Interaction | 76.13 (22.42) | 67.76, 84.51 | 76.84 (27.38) | 66.80, 86.89 | 1.20 (16.20) | −4.85, 7.25 | |
| Control | 80.39 (22.84) | 71.70, 89.08 | 84.72 (25.12) | 75.17, 94.28 | 4.33 (25.74) | −5.46, 14.13 | |
| Heel-to-heel base of support (m) | Walking | 0.12 (0.04) | 0.11, 0.14 | 0.12 (0.04) | 0.11, 0.14 | −0.00 (0.03) | −0.01, 0.01 |
| Interpersonal Interaction | 0.13 (0.06) | 0.10, 0.15 | 0.13 (0.06) | 0.11, 0.15 | −0.00 (0.06) | −0.02, 0.02 | |
| Control | 0.11 (0.04) | 0.09, 0.12 | 0.13 (0.07) | 0.010, 0.15 | 0.20 (0.07) | −0.00, 0.05 | |
Walking intervention group, n=31; Interpersonal interaction group, n=31; Control group, n=29; CI=Confidence interval, m=meter, min=minute, s=second, SD=standard deviation
Table 3.
Post-hoc analysis for gait velocity — complete case analysis and sub-group analysis
| Analysis (n) | Gait Parameter | Groups | Adjusted mean difference (95% CI) | p-valuea |
|---|---|---|---|---|
| Complete Case Analysis (n=91) | Velocity (m/s) | Walking vs Interpersonal interaction | 0.12 (0.03, 0.21) | 0.007* |
| Walking vs Control | 0.12 (0.03, 0.22) | 0.004* | ||
| Interpersonal interaction vs Control | 0.00 (−0.9, 0.09) | 1.000 | ||
| Sub-group Analysis (n=68) | Velocity <0.5m/s | Walking vs Interpersonal interaction | 0.12 (0.03, 0.22) | 0.007* |
| Walking vs Control | 0.10 (0.01, 1.19) | 0.033* | ||
| Interpersonal interaction vs Control | −0.02 (−0.11, 0.07) | 0.885 | ||
There was no statistically significant difference between groups for cadence, stride length and heel-to-heel base support — see Changes in gait parameters at 16 weeks post-intervention Section, post-hoc analysis was not completed; Sub-group analysis was performed using baseline gait velocity <0.5m/s and ≥ 0.5m/s as cut-offs (23); Significant difference in gait velocity was noted for participants with a baseline gait velocity of <0.5m/s (n=68) but not for individuals with ≥ 0.5m/s baseline gait velocity (n= 23); a Sidak's test; *= significant difference
Sub-group analysis for gait velocity
There was no significant difference between groups for baseline velocity ≥0.5m/s (n=23), [F(2,16)=0.25, p=0.779)]. There was a significant difference between groups in participants with a baseline velocity <0.5 m/s (n=68) [F(2,61)=5.57, p=0.006), η2 (95% CI)= 0.19 (0.03, 0.34)]. The EMM for the walking group [EMM (SE), 0.42m/s (0.03)] was significantly higher compared to the interpersonal interaction group [EMM (SE), 0.30m/s (0.03); t(61)=3.16, p=0.007] and the care-as-usual control group [EMM (SE), 0.32m/s (0.03); t(61)=2.62, p=0.033). See Table 3 for adjusted mean difference across groups for complete case analysis and sub-group analysis.
Discussion
We found that participants in LTC who engaged in a 16-week individualized walking intervention had improved gait velocity 16-weeks post-intervention compared to the interpersonal interaction group and the care-as-usual control group. This finding differs from MacRae et al.'s study (11), one of the first studies that explored the effect of a walking intervention on multiple outcome measures among LTC residents. They reported that the walking group participants significantly improved in walking endurance capacity (both maximal time and distance walked), but not in walking endurance speed. These differences could be because of their walking group were mostly female with a mean age of 91.7 years and had mild depressive symptoms and a baseline gait velocity of 0.28m/s. While we measured gait parameters, including gait speed using a computerized GAITRite® system, MacRae et al.'s study (11) calculated their participants' gait speed using the participants' maximum walk time performed during a single day walking.
The increase in gait velocity in the walking group in our study suggests potential clinical benefits of a walking intervention among LTC residents. Although not examined in the LTC population, Perera et al. (30) reported that the most meaningful change estimates of gait speed ranged from 0.08 to 0.14m/s for community-dwelling older adults and subacute stroke survivors. In our study, the change in gait velocity among participants in the walking intervention between baseline and 16 weeks post-intervention was 0.10m/s. This change could be considered clinically meaningful for LTC residents, as it is within the clinical meaningful estimate of 0.08 to 0.14m/s reported by Perera et al. (30). However, a further study is needed to estimate a clinical meaningful change of walking intervention among LTC residents. The estimated mean difference (0.12m/s, 95% CI=0.03, 0.22) noted in our study was higher than the estimated mean difference reported in a 2013 Cochrane systematic review of the fixed pooled effects of physical rehabilitation (e.g. exercise) on LTC residents' walking speed (0.03m/s, 95% CI=0.00, 0.07, p=0.02) (1). This difference may be due to the heterogeneity of the exercise interventions included in the review.
The sub-group analysis showed individuals with a baseline gait velocity of <0.50m/s improved significantly with the walking intervention compared to individuals with a baseline gait velocity of ≥0.5m/s. We could argue that LTC residents whose walking speed is slower might benefit more from a self-selected pace and distance walking intervention, while those with a faster walking speed might need a more challenging walking intervention. This sub-group finding is similar to the findings of a study of debilitated older adults in nursing homes and geriatric evaluation and management units — those with more impaired function made the most gains from an exercise intervention (31). It is possible that walking interventions may be more beneficial for LTC residents with slower gait velocity, e.g., <0.50m/s, and could help reduce the sedentary behaviour observed among LTC residents (5). Therefore, a self-selected pace and distance walking intervention could be an alternative for low physical functioning LTC residents who may not be able to participate in or adhere to more complex or multifactorial exercise programs.
We examined stride length, cadence and heel-to-heel base of support as secondary outcomes because some of these variables are associated with fall risk in the LTC population (12, 13). We found an absence of a significant effect of the walking intervention on these variables. This result is not entirely surprising, as our walking intervention was self-selected in terms of pace and distance; the RAs did not provide any specific instructions regarding altering gait patterns during the intervention, and the intervention was not multifaceted nor was it task oriented. Lord et al.'s (32) study reported a positive effect of a multifaceted, task-oriented, motor learning intervention on stride length and cadence among community-dwelling older women. A self-selected pace and distance walking intervention may not improve stride length, cadence, and heel-to-heel base support of LTC residents. Multifaceted walking interventions and task-specific walking interventions should be explored to determine which combinations of interventions may improve the known gait variables that predict falls among LTC residents.
While outcome measures of the original study (15) were completed at baseline, 8-weeks and 16-weeks post-baseline and 8-weeks and 16-weeks follow up, GaitRite ® gait variables were collected at baseline and 16-weeks post-intervention only. Thus, we were only able to examine gait variables at these two-time points, and we were not able to determine the immediate effect of walking intervention on our study outcomes. We cannot account for participants' mobility behavior after the conclusion of the walking intervention. However, it is implausible that individuals in any of study groups continued walking after the intervention. The walking intervention required the assistance of a RA, and LTC residents have been noted to spend most of their time in sedentary behaviours (e.g. sitting and lying down) (5).
The limitations of our study highlight some of the challenges associated with secondary analysis of existing data that may affect internal and external validity — having access to the data that are available, but lack of involvement in, and control over, recruitment, randomization, and data collection processes and related nuances. The range and variability of some of the data e.g., total distance walked in the intervention group may have introduced a Type II error. Another limitation of our study was that the proportion of missing gait parameters data was considerable (45.8%). Jakobsen et al. (29) suggested that multiple imputations should not be used to handle missing data if the proportion of missing data is too large (for example, more than 40%). As a result, we did not perform multiple imputation but used available case analysis. This approach may have reduced the statistical power due to the reduction in the study sample size, and Type II error may be introduced for the outcomes that did not reach significance. While we followed Jakobsen et al's (29) recommendations, the extent of missing data and analyzing available cases may have also introduced selection and confounding bias. Further studies using larger sample sizes with gait velocity as the primary outcome measure should be conducted to examine the generalizability of this finding. Selection bias may have also been introduced through the various elelements of the original study — recruitment procedures, not using a central methods centre for randomization, lack of stratification by LTC facility and demographic characteristics such as sex to randomly assign participants.
Walking is simple, cost-effective, has no reported adverse effects, and can easily be incorporated into the daily activities of LTC residents (8). The study findings suggest that a simple walking intervention is a promising intervention for LTC residents, particularly those with slower gait speed. Although walking did not have a significant effect on stride length, cadence and heel-to-heel base of support, walking may have a ‘protective effect’ on gait velocity, e.g., walking is more beneficial than “not walking” for older adults residing in LTC. We believe that walking can be used as a strategy for reducing sedentary behaviors among LTC residents.
Funding
This work was supported by the Saskatchewan Health Research Foundation Health Research (SHRF) Team Grant. Trial registration: ClinicalTrials.gov NCT01277809.
Conflict of interests
No potential conflict of interest was reported by the authors.
Ethical standards
The study was approved by the appropriate institutional research ethics board [University of Saskatchewan Biomedical Research Ethics Board, BIO#10-125], and the individual LTC facilities.
Electronic supplementary material
Supplementary material is available for this article at https://doi.org/10.1007/s12603-021-1683-6 and is accessible for authorized users.
Appendix: Baseline gait parameters differences by group
References
- 1.Shakeel S, Newhouse I, Malik A, Heckman G. Identifying feasible physical activity programs for long-term care homes in the Ontario context. Can Geriatr J. 2015;18(2):73–104. doi: 10.5770/cgj.18.158. 10.5770/cgj.18.158 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Volkers KM, Scherder EJA. Impoverished environment, cognition, aging and dementia. Rev Neurosci. 2011;3:259–266. doi: 10.1515/RNS.2011.026. [DOI] [PubMed] [Google Scholar]
- 3.Marshall SC, Berg K. Cessation of exercise in the institutionalized elderly: Effects on physical function. Physiother Canada. 2010;62(3):254–260. doi: 10.3138/physio.62.3.254. 10.3138/physio.62.3.254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.MacRae PG, Schnelle JF, Simmons SF, Ouslander JG. Physical activity levels of ambulatory nursing home residents. J Aging Phys Act. 1996;4(3):264–278. 10.1123/japa.4.3.264 [Google Scholar]
- 5.Ikezoe T, Asakawa Y, Shima H, Kishibuchi K, Ichihashi N. Daytime physical activity patterns and physical fitness in institutionalized elderly women: An exploratory study. Arch Gerontol Geriatr. 2013;57(2):221–225. doi: 10.1016/j.archger.2013.04.004. 10.1016/j.archger.2013.04.004 [DOI] [PubMed] [Google Scholar]
- 6.Shin C-N, Lee Y-S, Belyea M. Physical activity, benefits, and barriers across the aging continuum. Appl Nurs Res. 2018;44:107–112. doi: 10.1016/j.apnr.2018.10.003. 10.1016/j.apnr.2018.10.003 [DOI] [PubMed] [Google Scholar]
- 7.Benjamin K, Edwards N, Ploeg J, Legault F. Barriers to physical activity and restorative care for residents in long-term care: A review of the literature. J Aging Phys Act. 2014;22(1):154–165. doi: 10.1123/japa.2012-0139. 10.1123/japa.2012-0139 [DOI] [PubMed] [Google Scholar]
- 8.de Souto Barreto P, Morley JE, Chodzko-Zajko W H, Pitkala K, Weening-Djiksterhuis E, Rodriguez-Mañas L, et al. Recommendations on physical activity and exercise for older adults living in long-term care facilities: A taskforce report. J Am Med Dir Assoc. 2016;17(5):381–392. doi: 10.1016/j.jamda.2016.01.021. 10.1016/j.jamda.2016.01.021 [DOI] [PubMed] [Google Scholar]
- 9.Tappen RM, Roach KE, Applegate EB, Stowell P. Effect of a combined walking and conversation intervention on functional mobility of nursing home residents with Alzheimer disease. Alzheimer Dis Assoc Disord. 2000;14(4):196–201. doi: 10.1097/00002093-200010000-00002. 10.1097/00002093-200010000-00002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cott CA, Dawson P, Sidani S, Wells D. The effects of a walking/talking program on communication, ambulation, and functional status in residents with Alzheimer disease. Alzheimer Dis Assoc Disord. 2002;16(2):81–87. doi: 10.1097/00002093-200204000-00005. 10.1097/00002093-200204000-00005 [DOI] [PubMed] [Google Scholar]
- 11.MacRae PG, Asplund LA, Schnelle JF, Ouslander JG, Abrahamse A, Morris C. A walking program for nursing home residents: Effects on walk endurance, physical activity, mobility, and quality of life. J Am Geriatr Soc. 1966;44(2):175–180. doi: 10.1111/j.1532-5415.1996.tb02435.x. 10.1111/j.1532-5415.1996.tb02435.x [DOI] [PubMed] [Google Scholar]
- 12.Sterke CS, van Beeck EF, Looman CWN, Kressig RW, van der Cammen TJM. An electronic walkway can predict short-term fall risk in nursing home residents with dementia. Gait Posture. 2013;36(1):95–101. doi: 10.1016/j.gaitpost.2012.01.012. 10.1016/j.gaitpost.2012.01.012 [DOI] [PubMed] [Google Scholar]
- 13.Camicioli R, Licis L. Motor impairment predicts falls in specialized Alzheimer care units. Alzheimer Dis Assoc Disord. 2004;18(4):214–218. PubMed PMID: 15592133. [PubMed] [Google Scholar]
- 14.Fonad E, Wahlin T-B, Winblad B, Emami A, Sandmark H. Falls and fall risk among nursing home residents. J Clin Nurs. 2007;17(1):126–134. doi: 10.1111/j.1365-2702.2007.02005.x. 10.1111/j.1365-2702.2007.02005.x [DOI] [PubMed] [Google Scholar]
- 15.Dal Bello-Haas V, Thorpe LU, Lix LM, Scudds R, Hadjistavropoulos T. The effects of a long-term care walking program on balance, falls and well-being. BMC Geriatr. 2012;12(1):76. doi: 10.1186/1471-2318-12-76. 10.1186/1471-2318-12-76 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Thorpe LU, Whiting SJ, Dal B-H, Hadjistavropoulos T. An evaluation of a walking and socialization program in long-term care: Impact on injurious falls. Int J Aging Res. 2019;2(2):31. [Google Scholar]
- 17.Yesavage JA, Sheikh JI. Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. Clin Gerontol. 1986;5(1):165–173. 10.1300/J018v05n01_09 2. [Google Scholar]
- 18.Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc. 1975;23(10):433–441. doi: 10.1111/j.1532-5415.1975.tb00927.x. 10.1111/j.1532-5415.1975.tb00927.x [DOI] [PubMed] [Google Scholar]
- 19.ATS Committee on Proficiency Standards for Clinical Pulmonary Function Laboratories ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med. 2002;166(1):111–117. doi: 10.1164/ajrccm.166.1.at1102. 10.1164/ajrccm.166.1.at1102 [DOI] [PubMed] [Google Scholar]
- 20.Berg K, Wood-Dauphine S, Williams JI, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Canada. 1989;41(6):304–311. 10.3138/ptc.41.6.304 [Google Scholar]
- 21.Rose DJ, Jones CJ, Lucchese N. Predicting the probability of falls in community residing older adults using the 8-foot up-and-go: A new measure of functional mobility. J Aging Phys Act. 2002;10(4):466–475. 10.1123/japa.10.4.466 [Google Scholar]
- 22.McDonough AL, Batavia M, Chen FC, Kwon S, Ziai J. The validity and reliability of the GAITRite system’s measurements: A preliminary evaluation. Arch Phys Med Rehabil. 2001;82(3):419–425. doi: 10.1053/apmr.2001.19778. 10.1053/apmr.2001.19778 [DOI] [PubMed] [Google Scholar]
- 23.van Uden CJ, Besser MP. Test-retest reliability of temporal and spatial gait characteristics measured with an instrumented walkway system (GAITRite®) BMC Musculoskelet Disord. 2004;5(1):13. doi: 10.1186/1471-2474-5-13. 10.1186/1471-2474-5-13 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.GAITRite®. https://www.gaitrite.com/. Accessed 12 December 2018.
- 25.Little RJA, Rubin DB. Statistical analysis with missing data. John Wiley & Sons; New York: 1987. [Google Scholar]
- 26.Field A. Discovering statistics using IBM SPSS statistics. 4th edn. Sage Publications; London: 2012. [Google Scholar]
- 27.Thabane L, Mbuagbaw L, Zhang S, Samaan Z, Marcucci M, Ye C, et al. A tutorial on sensitivity analyses in clinical trials: The what, why, when and how. BMC Med Res Methodol. 2013;13:92. doi: 10.1186/1471-2288-13-92. 10.1186/1471-2288-13-92 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kuys SS, Peel NM, Klein K, Slater A, Hubbard RE. Gait speed in ambulant older people in long term care: A systematic review and meta-analysis. J Am Med Dir Assoc. 2014;15:194–200. doi: 10.1016/j.jamda.2013.10.015. 10.1016/j.jamda.2013.10.015 [DOI] [PubMed] [Google Scholar]
- 29.Jakobsen JC, Gluud C, Wetterslev J, Winkel P. When and how should multiple imputation be used for handling missing data in randomised clinical trials — A practical guide with flowcharts. BMC Med Res Methodol. 2017;17(1):162. doi: 10.1186/s12874-017-0442-1. 10.1186/s12874-017-0442-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc. 2006;54(5):743–749. doi: 10.1111/j.1532-5415.2006.00701.x. 10.1111/j.1532-5415.2006.00701.x [DOI] [PubMed] [Google Scholar]
- 31.Meuleman JR, Brechue WF, Kubilis PS, Lowenthal DT. Exercise training in the debilitated aged: Strength and functional outcomes. Arch Phys Med Rehabil. 2000;81(3):312–318. doi: 10.1016/s0003-9993(00)90077-7. 10.1016/S0003-9993(00)90077-7 [DOI] [PubMed] [Google Scholar]
- 32.Lord SR, Lloyd DG, Nirui M, Raymond J, Williams P, Stewart RA. The effect of exercise on gait patterns in older women: A randomized controlled trial. J Gerontol Ser A Biol Sci Med Sci. 1996;51A(2):M64–M70. doi: 10.1093/gerona/51a.2.m64. 10.1093/gerona/51A.2.M64 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix: Baseline gait parameters differences by group