Wheeled-mobility correlates of life-space and social participation in manual wheelchair-users aged 50 and older

Abstract

Background

Distance travelled, occupancy time in the wheelchair, and use of the wheelchair to transition between activities (i.e. bouts) are objective measures of wheelchair mobility that may contribute to the explanation of life-space mobility of manual wheelchair-users.

Objectives

To characterize the life-space mobility and social participation of manual wheelchair-users using multiple objective measures of wheeled mobility.

Design

Cross-sectional.

Methods

Individuals (n=49) were included for study if they were 50 years of age or older, community-dwelling, and used their wheelchair on a daily basis for the past 6-months. Life-space mobility and social participation were measured using the Life-Space Assessment and the Late Life Disability Instrument, respectively. The wheeled mobility variables (i.e. distance travelled, occupancy time, and number of bouts) were captured using a custom-built data logger.

Results

After controlling for age and sex, multivariate regression analyses revealed that the wheeled mobility variables accounted for 24% of the life-space variance. The number of bouts variable, however, did not account for any appreciable variance above and beyond the occupancy time and distance travelled variables. Occupancy time, and number of bouts were statistically significant predictors of social participation, and accounted for 23% of the variance after controlling for age and sex.

Limitations

Due to the use of a cross-sectional study design causality cannot be established. In addition, objectively monitoring wheelchair mobility may introduce limitations related to instrumentation failure and data loss.

Conclusion

Wheelchair occupancy time and distance travelled are statistically significant predictors of life-space mobility. Whereas lower occupancy time may be indicative of travel to more distant life-spaces, distance travelled is likely a better reflection of mobility within each life-space. Occupancy time and number of bouts are significant predictors of participation frequency.

Introduction

Mobility limitations are shown to compromise independence,^1,2 increase the risk of institutionalization,³ and mortality,¹ and decrease quality of life.^2,4 Mobility limitations are a leading cause of disability in community-dwelling individuals,⁵ and are the primary reason for both the onset and persistence of participation restrictions in individuals 50 years of age and older.⁶ Addressing issues with mobility is therefore an important rehabilitation focus, clinically and for research.

The assessment of life-space mobility is often considered a feasible and minimal burden approach to obtain measures of mobility. Life-spaces are defined as the different areas in which people conduct their lives.^7,8 They can range from within the home to beyond a person’s town or geographic region.^7,8 May et al. first conceptualized life-spaces as five concentric zones, beginning in the bedroom and extending through to the area across a traffic-bearing street.⁷ Life-spaces have also been conceptualized as zones within and around long-term care facilities,⁹ as well as from the bedroom to places outside of town.⁸

Mobility through life spaces reflects the spatial extent of travel within a person’s environment. In older, ambulatory populations more life-space mobility has been shown to be associated with better physical performance, activities of daily living (basic and instrumental), health-related quality of life, self-reported health and fewer depressive symptoms.⁸ Moreover, evidence suggests life-space mobility has a fair to moderate association with objective, component measures of mobility such as faster gait speed, smaller sway path, and better balancing ability in ambulatory populations.^7,8

Life-space mobility is increasingly being used as an indicator of wheelchair mobility.^10–16 Studies are providing evidence in support of hypothesized associations between life-space mobility of wheelchair-users and measures of social and community participation.^14,15 For example, Meyers et al. report that individuals with higher life-space scores reach significantly more desired community destinations, and overcome more barriers to reach the destinations than individuals with lower life-space scores.¹⁵ More recently, through the use of multivariable analyses, we demonstrated a positive association between life-space mobility and participation frequency in adult wheelchair-users.¹⁶

Life-space mobility models exist in the wheelchair-use literature, however, at present are largely limited to psychological, functional, and environmental variables. This body of research has established that self-efficacy,¹⁶ depression,¹² functional ability,¹¹ self-reported wheelchair skills,^11,12,16 need for a seating intervention,¹⁰ environmental barriers,¹² and social support¹¹ are all associated with life-space mobility. Although this research provides a foundational understanding of life-space mobility among wheelchair-users the extent to which life-space mobility is associated with objective measures of wheelchair mobility is unknown.

Wheelchair mobility is complex and single objective measures provide an incomplete picture.¹⁸ Distance travelled, occupancy time in the wheelchair, and use of the wheelchair to transition between activities (i.e. bouts) are component measures of wheelchair mobility that have successfully been used to develop a greater understanding of powered wheelchair mobility.¹⁸ For example, distance travelled may represent the amount of wheeling to get to work or school, or participation in social activities such as exercising or going for wheels/walks.¹⁸ Occupancy time is likely a reflection of the use of the wheelchair as a support while engaging in a desired activity, whether it be for moving between places, sitting in a restaurant, or at home performing various activities of daily living.¹⁸ A bout is conceptualized as the transition between stationary activities,¹⁸ and number of bouts has previously provided information into the number of activities performed in different locations among people using powered wheelchair.¹⁸ Knowledge of the associations between these objective measures of mobility and life-space will provide researchers and clinicians with important information that may be used to develop and evaluate interventions to increase life-space mobility in people who use manual wheelchairs.

Therefore, the objective of this study was to characterize the life-space mobility of manual wheelchair-users using multiple objective measures of wheeled mobility. We hypothesized that similar to studies of life-space in ambulatory individuals,^7,8 objective measures of wheeled mobility (i.e. distance travelled, occupancy time, number of bouts) will each have statistically significant (p≤0.05) positive associations with life-space mobility, after controlling for age and sex, in adult manual wheelchair-users. Because mobility has also been shown to be important for participation, a secondary, exploratory, regression analysis investigated the associations between the measures of wheeled mobility and social participation frequency.

Methods

Participants and recruitment

This study was embedded in a larger cross-sectional study¹⁴. A volunteer sample was recruited from three rehabilitation hospitals in Vancouver (British Columbia), and Quebec City and Montréal (Quebec), Canada, by community-based home-care rehabilitation teams from three regional health authorities in British Columbia, and by advocacy groups. Clinicians provided potential participants with study information and asked if they would like to be contacted by the research team for more detailed information. The research coordinator contacted anyone who expressed interest and agreed to be contacted. Study advertisements were also posted at the rehabilitation hospitals and community centres. Ethics approval was received from all participating institutions.

Individuals were included for study if they were 50 years of age or older, community-dwelling, and used their manual wheelchair daily for the past 6-months. Individuals with a Mini Mental State Examination (MMSE) score ≤23,¹⁹ and/or an acute illness were excluded.

Measures

Dependent variables: Life-space mobility was measured using the 20-item Life-Space Assessment (LSA).⁸ The LSA assesses the frequency (1=less than once a week, to 4=daily), and independence (1=assistance from other persons, 1.5=with equipment, or 2=no assistance) that individuals move in five life spaces over the past month: 1) within the home; 2) around the home; 3) in the neighbourhood; 4) in town; and 5) outside of town.⁸ A composite score is derived by first multiplying each life-space level by the weekly frequency, and then by the level of independence. The products from each life-space level are then summed to derive the composite score that ranges between 0 and 120, or 0 to 90 for individuals who use assistive devices such as those who participated in this study. Higher scores represent more life-space mobility. The composite score has excellent test-retest reliability in both community-dwelling ambulatory older adults (ICC=0.96, 95% CI=0.95,0.97),⁸ and power wheelchair-users (ICC=0.87, 95% CI=0.69, 0.92).¹³ Moderate correlations (r=−0.41 to 0.60) in the expected directions are reported with measures of physical performance, activities of daily living, and depression, in older individuals.⁸

Frequency of social participation was measured using the 16-item Late-Life Disability Instrument (LLDI).²⁰ Item responses range from 1 (never) to 5 (very often). Responses to each item are then summed to derive raw scores, which are then standardized into scores ranging from 0 to 100.²⁰ Higher scores indicate more frequent participation. Recent evidence supports the test-retest reliability of LLDI measurements in adult wheelchair-users (ICC=0.86, 95% CI=0.76–0.93).²¹ The LSA and LLDI are low burden self-report measures, each taking approximately 10 minutes to complete.

Independent variables: Data on distance travelled, wheelchair occupancy time, and number of bouts were collected using a custom-built data-logger, which included two Force Sensitive Resistors (FSRs), a Hall effect magnetic encoder, and a data storage/battery module. Installed data-loggers did not interfere with wheelchair usability (e.g. wheeling) or functioning (e.g. folding), and did not require any modification to the wheelchair.

Distance travelled data (meters) was collected using a magnetic encoder along with a ring of six magnets placed at the centre of the wheel. The encoder tracked the passing magnets, and recorded data every two seconds. Distance traveled was then calculated relative to the diameter of the wheel. Accuracy of the distance measure provided by the encoder has been verified through tests (straight line and turning) conducted at various speeds (~0.5m/s to ~1.5m/s), which have shown a maximal error of 2.5%. Mean distance travelled per day was calculated for each participant and used in the regression analysis.

Wheelchair occupancy time data (minutes) was collected using two FSRs. The FSRs were placed under the seat cushion. A pressure threshold was established for each participant to differentiate between cushion pressure (set to 0) and pressure from seated occupancy (set to 1). At least one FSR had to provide a signal to indicate that a participant was seated. Data-logger data from each participant was visually observed. If a participant’s data was not interpretable, the data set was excluded. Mean occupancy time per day was calculated for each participant and used in the regression analyses.

Number of bouts was calculated using the distance and occupancy time data. A bout started when the wheelchair traveled at least 0.72 meters in 6 seconds and stopped when the travel was less than 0.71 meters in 14 seconds. The bout start/stop criteria is similar to that used in previous research,²² and takes into account our system’s data acquisition rate. Because bouts are based on a distance measure, accuracy of bouts is dependent on the accuracy of the distance measure accuracy. Mean bouts per day was calculated and used in the regression analyses.

Demographic information was collected using the socio-demographic information form. Data on the number of comorbidities (Functional Comorbidity Index;²³ an 18-item index used to assess the number of health conditions (e.g. heart disease, diabetes, chronic obstructive pulmonary disease, arthritis) people have), need for a seating intervention (Seating Identification Tool;²⁴ an 11-item measure with total scores ranging between 0 and 15; a score of 2 or more is indicative of a need for a seating intervention), perceived social support (Interpersonal Support and Evaluation List;²⁵ a 6-item measure with total scores ranging between 0 and18; higher scores indicate more perceived social support), shoulder pain due to wheelchair use (Wheelchair User Shoulder Pain Index;^26,27 a 15-item index with total scores ranging between 0 and 150; higher scores indicate more intense shoulder pain); and functional independence (Barthel Index;²⁸ 10-item Barthel Index-postal version with total score ranging from 0 to 20; higher scores indicate greater functional independence) were also collected for descriptive purposes.

Study protocol

Between October 2011 and September 2012, eligible study volunteers met with a researcher for data collection, who was trained in the administration of all measures. After completing the socio-demographic information form, participants were administered the MMSE, and then the rest of measures in a random sequence to minimize an ordering effect response bias. Upon completing the standardized measures, the data-logger was installed on each participant’s wheelchair, and collected data for the next nine days (for data analysis the first and last days of data-logger data were removed because they were not full days of data collection). After the data-collecting period, participants met with the researcher for data-logger removal.

Sample size

Using an effect size of 0.45, based on a conservative squared multiple correlation coefficient (based on five independent variables: age,^10,16 sex,¹⁰ wheelchair skills,¹⁶ experience (years),¹⁶ and daily use,¹⁶ each with an estimated average, fair correlation magnitude of 0.25 with the LSA) of 0.31, and an alpha of 0.05, a sample size of 43 was determined, using G*Power version 3.1 (available at http://www.psycho.uniduesseldorf.de/abteilungen/aap/gpower3/), to have a power of 0.90 using regression analyses.

Data analysis

Descriptive statistics are presented as frequencies and percentages, and as means and standard deviations. Our life-space mobility model was powered based on the entry of five variables, including the three wheelchair mobility variables. Age and sex were selected for entry because previous research has demonstrated their confounding effects.^14,16 Pearson correlations were performed to determine the bivariable correlations between the mobility parameters and dependent variables. We also looked at possible differences in the LSA and LLDI scores by higher and lower mobility metrics using the non-parametric Mann-Whitney U test. We stratified the sample into higher and lower mobility cohorts using the median score as the cutoff for distance travelled and bouts, and 6 hours of occupancy time (to differentiate between part time and full time wheelchair users). Hierarchical multiple regression analyses were used to test the study hypotheses.²⁹ Age and sex were first entered into the model, followed by the wheelchair mobility variables (i.e. distance travelled, occupancy time, and number of bouts) in the second step to determine their relative statistical importance with the dependent variable. Similar procedures were used to develop the social participation frequency model.

Two indicators were used to assess for multicolinearity, including a correlation between independent variables of r≥0.70, and with a variation inflation factor value greater than 10.²⁹ To minimize multicolinearity all continuous variables entered into the regression model were mean centered, however, in the event of excessive multicolinearity between two independent variables, we only modeled the variable that had the highest correlation with the dependent variable. All other regression assumptions were also assessed.²⁹

Results

In this sample of 49 individuals, the median age was 57 years, and 32 (65%) were male. Twenty-six (53%) individuals reported having a spinal cord injury, and the mean number of comorbidities was 2.3. The median years of experience with using a wheelchair was 20, 10 (20%) individuals reported receiving training to use their wheelchair, and 6 (12%) required assistance with their wheelchair-use. Twenty-one people had a Seating Identification Tool score above 2, and the sample reported little shoulder pain resulting from wheelchair use, and moderate levels of disability according to the Barthel Index. Both the mean LSA and LLDI scores where in the lower half of all possible scores. Due to issues with the data-logger technology (e.g. loose wires) data were not collected for the entire nine days for all participants. Moreover, data was collected for more than nine days for some participants due to scheduling conflicts with a researcher to remove the logger. After removing the first and last day of data-logger data for each participant, the mobility variables were collected for a median of 7 days (range = 6 – 11). The sample’s median wheeled distance was 1430 meters per day, the median occupancy time was 740 minutes per day, and the median number of bouts per day was 176. Sample and mobility characteristics are detailed in tables 1 and 2, respectively.

Table 1.

Sample characteristics (n=49)

Variable	Mean ± standard deviation or frequency (%)	Median (IQR)
Age	59.1 ± 6.9	57.0 (10.0)
Male	32 (65.3)
Married/Common-law	23 (46.9)
Education*:
Some highschool	6 (12.5)
Highschool graduate	11 (22.9)
Some university	9 (18.8)
University graduate	21 (43.8)
Graduate degree	1 (2.1)
Employed/Volunteer	16 (32.7)
Income:
<14,999	9 (18.4)
15,000–29,999	11 (22.4)
30,000–44,999	5 (10.2)
45,000–59,999	6(12.2)
60,000–74,999	3 (6.1)
>75,000	8 (16.3)
Prefer not to answer	7 (14.3)
Diagnosis:
Spinal Cord Injury	26 (53.1)
Multiple Sclerosis	9 (18.4)
Stroke	1 (2.0)
Amputation	5 (10.2)
Other	8 (16.3)
FCI (0–18)	2.3 ± 2.1	2.0 (2.50)
ISEL (0–18)	14.0 ± 4.3	15.0 (4.50)
BI (0–20)	15.2 ± 2.3	15.0 (1.50)
WUSPI (0–150)	14.5 ± 21.6	3.3 (20.4)
Wheelchair-use (years)	24.5 ± 16.4	20.0 (31.0)
Self reported daily wheelchair-use (hrs)	12.4 ± 3.6	13.0 (4.8)
Received training to use a wheelchair	10 (20.4)
Requires assistance with wheelchair-use	6 (12.2)
SIT (≥2)	21 (42.9)

Note: FCI = Functional Comorbidity Index; SIT = Seating Identification Tool; ISEL = Interpersonal Support and Evaluation List; WUSPI = Wheelchair User Shoulder Pain Index; BI = Barthel Index;

^*n=48.

Table 2.

Descriptive statistics of the dependent and independent variables

	Mean ± standard deviation	Median (IQR)
LSA (0–90)*	43.8 ± 16.7	40.5 (26.5)
LLDI (0–100)	49.0 ± 6.1	49.5 (8.5)
Data-logger variables:
Distance travelled (daily meters)	2002.7 ± 1858.2	1429.7 (1510.7)
Occupancy time (daily minutes)	689.6 ± 237.4	739.3 (228.0)
Number of bouts	172.7 ± 78.9	175.6 (75.8)

LSA = Life-Space Assessment; LLDI = Late Life Disability Instrument;

^*the maximum total score is 120, but for people who use assistive devices it is 90.

Life-space mobility

The bivariable correlations between the wheeled-mobility variables and the LSA ranged from −0.25 to 0.35, as shown in table 3. There were statistically significant differences in the LSA scores by distance wheeled and occupancy time, with higher LSA scores reported by people who travel greater distances in their wheelchair, and occupy their wheelchair less throughout the day.

Table 3.

Pearson correlation matrix

	LSA	LLDI	Distance	Occupancy time
Distance (p-value)	0.35 (0.015)	0.29 (0.046)	-	-
Occupancy time (p-value)	−0.25 (0.079)	−0.27 (0.060)	0.37 (0.009)	-
Number of bouts (p-value)	0.18 (0.220)	0.25 (0.086)	0.50 (0.000)	0.45 (0.001)

Bold = statistically significant; LSA = Life-Space Assessment; LLDI = Late-Life Disability Instrument; Distance in meters/day; Occupancy time in minutes/day

After controlling for age and sex, the magnitude of the wheeled-mobility variables’ standardized regression coefficients ranged from −0.53 to 0.38, as shown in table 4. These variables independently accounted for 24% of the LSA variance, however, only occupancy time, and distance travelled were statistically significant predictors. No regression assumptions were violated.

Table 4.

Wheeled-mobility correlates of life-space in adult wheelchair-users

	Model 1				Model 2
	b	SE	β	95% CI	b	SE	β	95% CI
(Constant)	76.87	20.04		36.53, 117.20	51.59	18.82		13.64, 89.54
Age	−0.58	0.34	−0.24	−1.26, 0.10	−0.16	0.31	−0.06	−0.79, 0.48
Sex	−8.14	4.84	−0.24	−17.89, 1.60	−9.56	4.35	−0.28*	−18.32, −0.80
Distance travelled					0.003	0.001	0.38*	0.001, 0.01
Occupancy time					−0.04	0.01	−0.53*	−0.06, −0.02
Number of bouts					0.05	0.03	0.25	−0.02, 0.12
adj R2			0.07				0.31

b=unstandardized coefficients; SE=standard error; β=standardized coefficients; CI=confidence interval for b; adj=adjusted; Male and female sex were coded as −0.50 and 0.50, respectively; The units of measurement for age, distance travelled, occupancy time, and number of bouts are years, meters/day, minutes/day, and bout frequency/day, respectively; n=49.

^*p ≤ 0.05

Frequency of social participation

The bivariable correlations between the wheeled-mobility variables and the LLDI ranged from −0.27 to 0.29, as shown in table 3. There were statistically significant differences in the LLDI scores only by distance wheeled.

After controlling for age and sex, the magnitude of the wheeled-mobility variables’ standardized regression coefficients ranged from −0.54 to 0.34, as shown in table 5. Together, these variables independently accounted for 23% of the LLDI variance, with all variables being statistically significant predictors. Less wheelchair occupancy time, and higher number of bouts were predictive of higher LLDI scores.

Table 5.

Wheeled-mobility correlates of social participation frequency in adult wheelchair-users

	Model 1				Model 2
	b	SE	β	95% CI	b	SE	β	95% CI
(Constant)	61.54	7.52		46.22, 76.67	51.96	7.18		37.49, 66.43
Age	−0.21	0.13	−0.24	−0.47, 0.41	−0.05	0.12	−0.06	−0.30, 0.19
Sex	−0.42	1.81	−0.03	−4.08, 3.23	−1.22	1.66	−0.10	−4.56, 2.11
Distance travelled					0.001	0.000	0.29*	0.00, 0.002
Occupancy time					−0.01	0.004	−0.54*	−0.02, −0.01
Number of bouts					0.03	0.01	0.34*	0.00, 0.05
adj R2			0.02				0.25

b=unstandardized coefficients; SE=standard error; β=standardized coefficients; CI=confidence interval for b; adj=adjusted; Male and female sex was coded as −0.50 and 0.50 respectively; The units of measurement for age, distance travelled, occupancy time, and number of bouts are years, meters/day, minutes/day, and bout frequency/day, respectively; n=49.

^*p ≤ 0.05

Discussion

Participants in this research were community-dwelling adults with a variety of diagnoses, and many years of wheelchair-use experience. The sample’s mean life-space mobility is greater than reports from an older sample of individuals who use powered mobility devices,¹⁰ and less than that of younger individuals who use manual wheelchairs.³⁰ This sample’s life-space mobility may be considered low when considering the mean LSA score is below the midpoint of all possible scores for individuals who use assistive devices for mobility (i.e the maximum LSA score for individuals who use assistive devices is 90). It is therefore plausible that the life-space of the individuals in this study maybe largely limited to their homes and local neighbourhoods.

The primary objective of this study was to characterize the life-space mobility of adult manual wheelchair-users using multiple, objective measures of wheeled mobility. After controlling for age and sex, the objective measures of wheeled mobility (i.e. distance travelled, occupancy time, and number of bouts) explained 24% of the LSA variance. The number of bouts variable, however, did not contribute to the explanation of the LSA beyond the distance travelled and occupancy time variables. The results therefore provide partial evidence in support of our hypothesis that all of distance travelled, occupancy time, and number of bouts would each be independent and statistically significant predictors of life-space mobility.

Occupancy time was the strongest predictor of LSA. Interestingly, individuals occupying their wheelchair for fewer minutes throughout the day also reported having greater life-space mobility. A plausible explanation for this observation is that individuals reporting more distant life spaces (i.e. life space level 4 = to town, and life space level 5 = outside of town) would also likely transfer in and out of their wheelchair and use their automobiles or other forms of transportation to get to those life-spaces. In these cases, the individuals would have higher life-space mobility scores, along with low occupancy time because they are not using their wheelchair for longer-distance travel. Our results and other research showing the negative direction of the associations between age and both life-space mobility^16,31 and social and community participation^14,15 supports this reasoning. That is, younger individuals who report more social and community participation are likely to travel further distances by automobile than older individuals. The hypothesis that age (i.e. <50 years versus ≥50 years) moderates the association between occupancy time and life-space mobility is an area for future study. The statistically significant positive association between distance travelled and life-space mobility is not surprising. It is an indication that individuals with higher LSA scores also travel further distances in their wheelchair. When considered in combination with the results of occupancy time, distance travelled in the wheelchair may better reflect frequency of mobility within each life-space, whereas occupancy time may be an indicator of travel to distant life-space levels. Furthermore, Baker et al.⁸ has previously defined a change in the LSA by 10 points as clinically important. Using this criterion in combination with our regression results, wheelchair users would have to change their occupancy time by 250 minutes per day, daily distance by 3.3 kilometers, or a combination of occupancy time and distance travelled (e.g. 60 minutes of occupancy time and 2.6 kilometers) to experience a 10-point change in LSA. By providing this mobility-related context to the LSA, researchers will be able to better interpret the scores, and understand the nuances of life-space mobility among adult wheelchair-users.

In the secondary investigation, social participation frequency was regressed on the objective wheeled-mobility variables. The results indicate that occupancy time and number of bouts were statistically significant predictors of social participation, measured using the LLDI. Similar to the life-space mobility model, occupancy time had a negative association with social participation, and was also the strongest predictor. This may imply that social participation activities are often performed while individuals are not in the wheelchair. For example, when meeting with friends in public places, such as restaurants, or at organized social events, such as at community or recreation centres, individuals may transfer from their wheelchair to a location that is more convenient for socializing (e.g. at a chair at a table). It is plausible that this may speak to the issue of accessibility of public spaces for wheelchair-users. Evidence shows that wheelchair-users who report more frequent participation, also report experiencing more perceived environmental barriers.¹² This corroborates our speculation that individuals who have higher levels of social participation only do so at an inconvenience to them due to accessibility issues.

The finding of a significant association between the number of bouts and social participation variables is a novel contribution to the wheelchair literature. This observation may indicate that those more active in their social participation also use their wheelchair to transition more often between the various activities. While this seems obvious, we speculate that it raises an important safety concern when considering that these individuals are perhaps transferring in/out of their wheelchair more often than individuals reporting less participation and more occupancy time in their wheelchair. Research shows that transferring in/out of the wheelchair is a risk factor for both serious injury and death. In fact, between 1986 and 1990, 17% of the estimated 36,000 wheelchair-related accidents warranting a visit to an emergency department in the United States were falls during transfers.³² In 2003, of the 102,300 wheelchair-related injuries in the United States, more than 65% of the injuries were due to tips and falls of which many were during transfers.³³ Further research examining the associations between social participation with component measures of wheelchair mobility such as, occupancy time, distance traveled, and number of bouts is warranted.

Limitations

This study has several limitations. First, the results are limited to individuals with similar characteristics to those in this study’s sample. We used a single MMSE cutscore to determine exclusion, however, alternative interpretations of the MMSE scores may vary with age, sex, and education³⁴ that we did not consider. Moreover, the results of the statistical analyses using the composite LSA score may be difficult to interpret³⁵ given the score is formed by multiplying three separate variables (frequency, level, and independence). However, the composite score has been validated previous studies. Next, the self-report nature of the questionnaire data from a volunteer sample may be influenced by selection and measurement bias and/or social desirability. As a result, the volunteer sample may not accurately represent the population as a whole. As well, due to the use of a cross-sectional study design causality cannot be established. In addition, objectively monitoring wheelchair mobility also introduces limitations related to altered mobility patterns as well as instrumentation failure and data loss. For example, in our study the number of days complete data was captured slightly differs for each participant. Finally, we may have excluded important mobility variables (e.g. bout speed, wheeling time) or interaction effects (e.g. to test whether the associations differ by age, occupancy time or other variables) that that may contribute to the explanation of life-space mobility and participation. However, given that the number of independent variables that we could model was limited due to our sample size, we believe we included the most important mobility variables that provide a broad description of wheeled mobility.

Conclusion

Wheelchair occupancy time and distance travelled are statistically significant predictors of life-space mobility. Whereas lower occupancy time is an indication of travel to more distant life-spaces, distance travelled is likely a better reflection of mobility within each life-space. Occupancy time, distance travelled, and number of bouts have important associations with social participation. That lower occupancy time and greater number of bouts are associated with more frequent participation raises accessibility and safety issues for manual wheelchair-users.

List of abbreviations

BI

Barthel Index

FCI

Functional Comorbidity Index

ISEL

Interpersonal Support and Evaluation List

LLDI

Late Life Disability Instrument

LSA

Life-Space Assessment

MMSE

Mini-Mental State Examination

SIT

Seating Identification Tool

WUSPI

Wheelchair User Shoulder Pain Index