DISCUSSION
There was a strong relationship between cognition and performance measures of gait and balance/mobility col- lected by physical therapists, regardless of living situation.
Table 3. Associations* between Dynamic Gait Index (DGI) (n = 3,877) and Performance Oriented Measurement Assessment (POMA) (n = 6,154) measures and likelihood of Outcome and Assessment Information Set (OASIS) (M0560) cognitive functional impairment relative to independence: odds ratios with 95 percent confidence intervals (CIs).
*Adjusted for age, sex, and modified Clinical Test of Sensory Integration and Balance.
1094
JRRD, Volume 50, Number 8, 2013
Gait performance as measured by the DGI and mobility as measured by the POMA were independently associated with requiring cognitive prompting by worse scores on the gait measures. Assistance or dependence with cognitive tasks was also associated with the inability to perform any portion of the mCTSIB. This is the first study to our knowledge to find a relationship between cognition and standard measures of gait performance in older adults undergoing physical therapy services in the home. The difference in DGI scores between the cognitively intact and the considerable assistance groups was 2.5 points. These people in the home all had very impaired gait; all mean scores for all four groups were below a mean of 11.4 out of 24 points. Other measures of gait performance have been shown to be associated with cognitive impairment. Five and seven percent, respectively, of variance in gait speed was predicted by standard measures of cognitive function (the Trail Making-Part B and the Stroop Interference tests) in community-living older adults with mild cogni- tive decline [25]. Although the relationship between cog- nition and gait performance is statistically significant in our study, the clinical importance of our findings is diffi- cult to interpret as there is little to no data in the literature that describes these relationships in the home healthcare practice setting. Our findings show a 2 percent prediction in variance in DGI score due to cognitive function and a 7 percent reduction in likelihood of cognitive independence for every 1-point change in DGI. The OASIS measure of cognitive function is a 4-level scale, unlike many other more precise measures of cognitive function. There was a difference of 2.5 DGI points between the group of people rated as cognitively independent and those who required cognitive assistance after adjustment for age, sex, and standing balance. A 3-point change is considered a clini- cally meaningful change in persons living in the commu- nity who have higher functional performance levels [26]. Further investigation is required to explain the clinical significance of the association between cognition and gait performance in people receiving home-care services. Patients typically referred to home-care services have difficulty walking and also may have cognitive dysfunc- tion. If patients walked well, they would be seen in an outpatient setting for care. The minimal clinically impor- tant difference (MCID) has not been previously identi- fied in the home-care population, so it is impossible to determine whether the MCID has been reached, but the
difference of 2.5 may be clinically important in patients with limited gait function. Therapist choice of gait measure appeared to be affected by the patients’ cognition, with patients who were more cognitively aware asked to perform the DGI. As part of the DGI, four items have three or more com- mands that must be remembered in order to perform the gait task correctly. The physical therapists appear to have selected people with higher cognitive functioning to per- form the DGI. A minimal detectable change at the 95 percent CI for the POMA and the DGI has been reported as a 3- and a 2-point change, respectively [27]. The mean difference between the POMA scores of those who were indepen- dent and those who required considerable assistance was 1.9 (minimal detectable change is 3). The differences on the POMA did not exceed the minimal detectable change. For the DGI, the mean difference in the independent group versus the considerable assistance group was 2.6 (minimal detectable change score is 2). Because the DGI requires more complex, multistep commands, it is no sur- prise that those who required cognitive prompting would have lower scores on the DGI than the POMA. Although we did not examine change over time, it appears that the DGI might be more responsive to variations in cognitive status than the POMA. Walking (gait) is important for maintaining func- tional independence. While reliant on the functioning of peripheral organs such as the musculoskeletal system, vestibular apparatus, and cardiovascular fitness, normal gait performance also depends on cortical input. Walking is the most common form of physical activity among older adults [28], is increasingly being examined as a means of health promotion and disease prevention, and has been suggested as being protective for cognitive health [9]. Although there is evidence that walking may be protective of cognitive status, there are findings that suggest that walking speed can also be used as a determi- nant of cognitive function [29]. It has been demonstrated in community-based studies in older adults that cognitive decline and dementia are strongly associated with higher risk of developing subsequent slow gait speed, especially in advanced states [30–31], and lower frequency of walk- ing in the neighborhood [9]; also, decreases in cognition and gait speed have been found to be parallel over time [32–34]. Gait performance as assessed by the DGI and the POMA suggests that our homebound sample had gait and
1095
WHITNEY et al. Cognition and gait performance in older adults
balance impairments. As expected, average DGI and POMA scores indicated that the subjects were at risk for falling because of subject selection, as they were referred to the home-care program based on their increased risk for falls at onset of their episode of care [16–17]. Gait speed may be an early marker of subsequent cognitive impairment [34]. Changes in coordination, finger tapping speed, and time to walk 30 ft may be attributable to path- ological processes associated with cognitive decline [35]. Our study also supports these findings that walking per- formance and cognition are related in a large sample of people receiving care in their homes. Several mechanisms might explain the relationship between slow walking speed and subsequent decline in cognitive function. One possible mechanism is that walk- ing speed could be a marker of brain lesions (white mat- ter disease) that are associated with future cognitive decline [32]. Progression of these small white matter lesions could produce abnormalities in walking speed early in the process of decline, and only later might impairments in cognition be detected [32]. Periventricu- lar white matter hyper intensities (high signal) on mag- netic resonance imaging (MRI) scans and increased ventricular volume, even in apparently healthy older adults, have also been associated with slowing of walking speed [36]. These central nervous system changes con- tribute to important markers of aging [35]. More recently, changes on MRI scans in conjunction with presence of apolipoprotein E epsilon 4 predicted future cognitive decline and poor physical function in 444 older adults over a 10 yr period. Radiological findings in conjunction with apolipoprotein E epsilon 4 may single out a group at higher risk for dementia [37]. Cognitive decline and poor physical function are risk factors for disability in old age and were related in our sample. Cognitive impairment has been considered to be a negative prognostic factor for successful rehabilitation and functional recovery [38]. Lower scores on the MMSE have also been associated with poorer physical performance on tests of gait speed, chair stand, and max- imal inspiratory pressure and predicted 2 yr declines in physical performance on a summary measure of gait speed, chair stand, and standing balance tests [33]. Suc- cessful execution of a physical task demands cognitive processes regardless of the nature of the task. Research- ers using performance-based measures of physical func- tioning should pay particular attention to the cognitive capacities of their subjects because these might influence
their assessment [34]. Cognition and physical perfor- mance in persons undergoing home-care interventions should be further studied with longitudinal follow-up in order to improve current knowledge on the effect of cog- nitive decline on gait/balance performance.