Relationship of clinic based gait speed measurement to limitations in community based activities in older adults

Abstract

Objective

To examine the ability of clinic based assessments of gait speed to capture limitations in a broad range of home and community based activities.

Design

Cross-sectional study

Setting

Community based aging cohort study

Participants

655 community residing individuals (61% women) who were age 70 and older (mean 80.4 years).

Interventions

None

Main outcome measures

Limitations on three gait related activities of daily living (walking inside home and climbing up and down stairs) and six motor based but gait independent activities (bathing, dressing, getting up from a chair, toileting, shopping, and using public transportation).

Results

Gait speed was associated with presence of self-reported difficulty on all three home based activities that were directly gait related and in 5 out of the 6 motor based activities. Gait speed of ≤1 m/sec was associated with increased risk of limitations on at least one out the nine selected activities (odds ratio 3.21, 95% CI 2.24 to 4.58, p <0.001).

Conclusions

Gait speed measured in clinical settings has ecological validity as a clinical marker of functional status in older adults for use in clinical and research settings.

INTRODUCTION

Gait speed has been compared to a measure of vital signs in older adults; a screening measure that reflects integration of health, disease, fitness, and emotional state.¹ Gait speed shows high correlation with self reported walking ability, and is associated with health status,¹ frailty,² and falls³ in older adults. While home based assessments have ecological validity, it would not always be practical or feasible to always obtain gait assessments at home for clinical or research purposes. Moreover, differences in individual home environments and space limitations may reduce reliability of home based assessments and limit comparisons between individuals.

Documenting the ability of gait speed to capture limitations in a broad range of activities will support the ecological validity of this measure as an outcome measure for clinical and research purposes. Hence, we examined the association of gait speed with activities of daily living (ADL) that varied in their dependence on walking capacity in a cohort of nondisabled and nondemented older adults. We chose nine ADLs that are among those that older adults need to do successfully to live independently in the community, three that were gait based and six where walking was not the primary focus.

METHODS

Sample

We undertook a cross-sectional study in the Einstein Aging Study.⁴ Study design has been reported.⁴ Briefly, potential subjects (age 70 and over) identified from Bronx County population lists were contacted by letter and then by telephone. Subjects who gave verbal consent on the telephone were invited for in-person evaluation. Exclusion criteria included severe audiovisual loss, bed bound, and institutionalization. Additional exclusion criteria for this study included presence of dementia diagnosed by study clinicians or disability (unable to perform ADLs).^{3, 5} Study protocols were approved by the local institutional review board, and written informed consents were obtained from each subject.

Gait

Gait speed (m/s) was measured while subjects walked at their usual pace for two trials on a computerized walkway (180.0 × 35.5 × 0.25 inches) with embedded pressure sensors (GAITRite^a).³ White lines on the floor marked start and stop points, which included three feet from the walkway edge for initial acceleration and terminal deceleration. Based on footfalls recorded on the walkway, the software automatically computes gait speed as the mean of two trials during steady state walking. The GAITRite system is widely used and excellent reliability has been reported in our and other studies.³

Activity assessment

Activity questionnaires were administered by research assistants at the same visit as the gait assessment. We assessed subjects’ self-reported difficulty or limitations in doing nine home or community based ADLs using three validated functional assessment questionnaires.^5-7 The nine ADLs included three gait related activities (walking inside home, climbing up stairs and climbing down stairs) as well as six motor based ADLs that required some walking but could also be performed without walking (bathing, dressing, getting up from a chair, toileting, shopping and using public transportation). Bathing, dressing, walking inside home, getting up from a chair, and toileting were selected from a scale developed by Gill and colleagues for assessing function in community samples.⁵ We previously reported that self-reported difficulty in stair negotiation was correlated with activity limitations.⁷ Difficulty in climbing up or down stairs were ascertained from this scale.⁷ We assessed ability to shop and to use public transportation from the Lawton-Brody scale.⁶ There were not sufficient subjects with limitations in other activities such as managing finances from the Lawton-Brody scale.⁶

Clinical assessment

Clinical assistants used structured questionnaires to elicit history of medical illnesses, falls in the previous two months, fear of falling, medications, and depressive symptoms.⁸ Presence of self-reported depression, diabetes, heart failure, hypertension, angina, myocardial infarction, strokes, Parkinson’s disease, chronic obstructive lung disease, and arthritis were used to calculate a summary illness index.⁹ General cognitive status was assessed by the Blessed-Information-Memory concentration test (range 0 to 32, higher scores worse).¹⁰

Data analysis

We used logistic regression analysis to study the cross-sectional associations of gait speed with individual activities, adjusting for the effects of age, gender, years of education, illness index, falls, fear of falling, and Blessed test scores.

Other investigators^{11, 12} and a recent working group¹ recommended a cutscore of ≤1 m/sec on gait speed to define slow gait and to identify disability risk. We examined the association of this gait speed cutscore with community based ADL limitations using logistic regression. ADL limitations were categorized into four groups for this secondary analysis: 0 (reference), 1, 2 and three or more. Since all nine ADLs are essential for successful living in the community, we chose limitation in any one of these nine activities as the minimal threshold for abnormality.

RESULTS

Study population

Of the 709 subjects enrolled and assessed in our mobility study over a 46 month period, 54 were excluded due to dementia (26), disability (23), or missing gait assessments (5). Hence, 655 individuals were eligible. The average age was 80.4 years and years of education were 14 ± 3.4. There were 257 men (39 %) and 398 women (61 %). Mean gait speed was 0.94 ± 0.23 m/s. The mean Blessed test score was 1.9 ± 2.0.¹⁰ The mean score on the 15-item Geriatric depression scale was 2.1 ± 2.0.⁸ Fifteen percent reported a fall and 33% fear of falling. There was low illness burden (mean 1.3 illnesses per subject).

Gait speed

Table 1 shows that gait speed was associated with self-reported difficulty on all three gait related home based ADLs and in 5 out of the 6 motor based ADLs. These associations remained significant even after statistical corrections for multiple comparisons.¹³

Table 1.

Association of gait speed (per 1 cm/sec decrease) with limitations in home based activities in 655 seniors. Estimates derived from logistic regression analysis adjusted for age, gender, education, illness index, falls, fear of falling, and Blessed test scores

Activity limitations	N	Odds Ratio	95% CI	p-value
Gait related
Walking	214	1.04	1.02 - 1.05	<0.001
Climbing up stairs	301	1.03	1.02 - 1.04	<0.001
Climbing down stairs	147	1.04	1.03 - 1.06	<0.001
Motor based
Bathing	67	1.02	1.00 - 1.04	0.03
Dressing	70	1.03	1.01 - 1.04	0.002
Getting up from chair	159	1.03	1.08 - 1.04	0.02
Toileting	35	1.05	1.02 - 1.07	<0.001
Shopping	29	1.00	0.97 - 1.07	0.99
Using public transportation	36	1.05	1.03 - 1.07	<0.001

Of the 655 subjects, 423 (64.6%) reported difficulty on one or more ADLs from our list. Using the category of no difficulty in any ADL as the reference, gait speed of ≤1 m/s was associated with increased risk of limitations on any one out nine ADLs (odds ratio 2.97, 95% CI 1.72 to 5.13, p <0.001), on any two out of nine ADLs (odds ratio 2.80, 95% CI 1.37 to 5.74, p =0.005) and on any three or more out of the nine ADLs (odds ratio 7.01, 95% CI 3.40 to 14.44, p <0.001).

Figure 1 shows gait velocity levels for categories of activity limitations.

Figure 1.

Box plot showing gait speed values (cm/s) for categories of activity limitation. The line in the middle of the box represents the median value. The ends of the box represent the 25^th and 75^th quartile values.

DISCUSSION

In this large, well-characterized cohort of community residing nondisabled and nondemented older adults, clinic based gait speed measurement was associated with self-reported limitations in a range of activities that older adults commonly engage in home and community settings. The ADLs were not chosen to be an exhaustive list. Rather, we sought to study a range of home based activities that were both dependent as well as independent of gait. While it may be expected that gait speed best identifies activities that are gait related, it also captures limitations in a broader range of activities. Gait speed was associated with limitations on all three activities that were gait related and five out of six motor based activities. Gait speed is influenced by variables such as fitness, cognition, and mood. Hence, measuring gait speed might help capture limitations in those activities that dependent on the integrity of these underlying processes.

Limitations

Our examination was limited to gait speed. While we collect many other gait variables,⁹ we did not report their correlations as other measures (such as gait variability) are not yet recommended for routine use.^{1, 9} There is increasing interest in the role of laboratory based gait variables in predicting disability.¹⁴ Future studies should examine the relationship of these other measures with ADLs. While some of our motor based activities may involve some walking, these activities could be performed even if the subject cannot walk. For instance, older adults who are unable to walk due to medical conditions can still bathe or shop using assistive devices or with the help of caregivers. We excluded subjects with dementia to improve reliability of responses but acknowledge limitations of self-report. However, the observed associations might be strengthened by more precise activity estimates. The low illness score and low Blessed test scores suggest that our sample is healthier than the general population. Higher illness burden and cognitive impairment are associated with slow gait as well as activity limitations. Hence, stronger associations between gait speed and activity limitations may be expected in less healthier elderly samples. It was not our intention to validate a new cutscore for diagnosing disability in our secondary analysis. We used the literature based cutscore^{1,11, 12} to illustrate the relationship between in-house gait speed assessment and community based activities. The optimal gait speed cutscore using ROC analysis for identifying activity limitations in our study sample was 1.01 m/s, which was almost identical to the recommended cutscore.

Strengths include the large sample and standardized evaluation procedures. Gait speed was measured using an instrumented walkway but can also be calculated with a stopwatch in clinics.

Our findings show that gait speed measured in clinical settings has ecological validity as a clinical marker of functional status in older adults for use in clinical and research settings, and supports its use a simple and practical indicator of the effectiveness of treatment or rehabilitation programs.