Abstract
Introduction
The instrumented Timed Up and Go test (iTUG) affords quantification of the sub-elements of the Timed Up and Go test to assess falls risk and physical performance. A miniature sensor applied to the back is able to capture accelerations and velocities from which the sub-elements of the iTUG can be quantified. This study is the first to compare iTUG performance between people with dementia (PWD) and their age matched caregivers. The aims of this study were to explore how age moderates the differences in performance on the instrumented Timed Up and Go test between PWD and their informal caregivers.
Methods
Eight-three community dwelling older PWD and their informal caregivers were recruited for this cross sectional, observational study. Participants were grouped by age; <70 years, 70-79 years and 80+ years old. Participants wore an inertial sensor while performing the iTUG in their home. The performance of the sub-elements sit to stand, walking and turning were captured through an algorithm converting accelerations and velocities into performance metrics such as duration and peak velocity. Performance for PWD were compared to caregivers for each age matched group and multiple regression models incorporating age, gender and presence or absence of dementia were computed.
Results
PWD took longer to turn in <70 year group, suggesting this may be an early indicator of functional decline in this age group. PWD took longer to complete the whole iTUG compared to caregivers in the 70-79 year old group. In the 80+ year old group PWD took longer to complete both walking phases, sit-to-stand and the full iTUG along with displaying slower turning velocity. Multiple regression models illustrated that gender failed to contribute significantly to the model, but age and presence of dementia explained around 30% of the variance of time to complete walking phases, total iTUG and turning velocity.
Conclusions
Differences were evident in performance of the iTUG between PWD and caregivers even after controlling for age. Age moderates the differences observed in performance.
Keywords: Inertial sensor, aging, dementia, balance, motor performance, timed up and go
Introduction
Falls in later life are globally recognised as a major public health issue.1 Among adults aged 65 and above, falls are the leading cause for emergency department presentation.2 Approximately 10% of falls among those aged 75 or above result in hip fracture,3 and only around a quarter of these patients return to their pre-fracture level of functioning within 90 days.4 Falls are often the reason for an older person to be admitted into long term residential care,3 and are associated with reduced social participation from a fear of falling and increased costs to health and social services.1,5,6
The risk of falls is higher among subgroups of the older population. The risk of falls increases with age3 and is different between men and women; women fall more often than men but men have more fatal falls.1 Another risk factor for falls is dementia; a degenerative neurological disease characterised by a chronic, global, and non-reversible loss of cognitive functioning.8 Estimates suggest that 46.8 million people had dementia in 2015 and that this figure will rise to 131.5 million worldwide by 2050.9 People with dementia (PWD) are more than twice as likely to fall and twice as likely to experience injurious falls compared to their cognitively intact peers.10,11 In addition, PWD are more likely to experience adverse health outcomes after injurious falls during their hospital stay and after discharge such as hospital readmission, institutionalisation, and mortality.12,13,14
The Timed Up and Go test (TUG) is one of the most frequently used tests to quantify physical function and falls risk for older adults.15,16 The TUG records time (seconds) to complete a continuous series of tasks (stand from a sitting position, walk 3 meters, turn and walk back finishing by returning to a sitting position). Previous research has proposed a threshold of 13.5 seconds or longer being associated with greater falls risk.16 The TUG is quick to administer but is criticized for providing a single value for a task involving multiple transitions, walking and turning all based on differing physiological constructs.17 In addition, there is evidence supporting a lack of prospective validity of the TUG to predict falls.18,19 A more detailed instrumented version of the TUG (iTUG), where the individual wears a body sensor on the low back during the test, generates metrics for each of the motor sub-elements of the TUG: peak acceleration and duration of the initial sit to stand; duration, regularity and symmetry of walking phases; as well as velocity and duration of the turning phases has been proposed.20 These sub-elements include the sit to stand phase, where peak accelerations and duration is quantified; the walking phases, where duration and metrics of regularity and symmetry are quantified and turning phases, where peak turning velocity and duration are quantified. These sub-phases are identified though identification of accelerations and velocities which physiologically correspond to those relative movements, captured by body worn sensors. The iTUG has been found to offer good repeated measures reliability and validity.20,21,22,23,24 The iTUG offers greater discriminatory ability for performance deficits than time to complete TUG25,26,27 and therefore may offer early insights into physical impairments. The iTUG sub-elements have been used to detect performance differences between people with mild cognitive impairment and age-matched peers,25 and to explore relationships between cognitive function, fear of falling and quality of life among PWD.28 However, iTUG performance has yet to be compared between PWD and age-matched peers. Therefore, the aims of this study were to explore how age moderates the differences in performance on the instrumented Timed Up and Go test between PWD and their informal caregivers. This will provide new insights into the iTUG performance deficits of dementia independent of age.
Methods
This cross section observational study used baseline data collected during the TACIT trial (NTC02864056), a randomized controlled trial to test the effects of Tai Chi on postural balance and falls in PWD and their informal caregivers.29 This study was approved by the West of Scotland Research Ethics Committee 4 (reference: 16/WS/0139 and the Health Research Authority (IRAS project ID: 209193).
Participants
Eighty-three persons with dementia and their informal caregiver were recruited from NHS databases, memory clinics, local charities and through self-referral from across the South of England. As this study used baseline data from the TACIT trial, inclusion criteria reflected recruitment for the TACIT trial. Caregivers needed to be living with the person with dementia or able to visit at least twice a week, able to participate in standing Tai Chi, and be able to commit to supporting the PWD in data collection and in Tai Chi weekly classes and home practice. Exclusion criteria included those caregivers with severe sensory impairment or lacked mental capacity to provide informed consent. PWD were included if they were aged 18 or more; lived at home; had a diagnosis of dementia (indicated on their medical record held by the national health service or general practitioner) and willing and able to complete standing Tai Chi (as part of the TACIT trial). PWD were excluded if they lived in a long term residential care facility; were in receipt of palliative care; scored <=9 on the Mini-Addenbrooke’s Cognitive Evaluation (M-ACE);30 had Lewy body dementia or dementia with Parkinson’s Disease; severe sensory impairment; were currently under the care of or had been referred to a falls clinic for assessment, currently attending a balance exercise program (e.g. Otago classes), or lacked mental capacity to provide informed consent. In addition participants were excluded if they had completed Tai Chi or similar exercise (yoga, Qi gong, or Pilates) once a week or more within six months of the commencement of the study.
The sample size was based on that used for the Tacit trial.29 This was to recruit a sample of 120, powered for a difference of 4 seconds in total time to complete TUG, with a standard deviation of 0.38, a correlation of 0.7 and a 2-sided 5% significance level and 90% power. While the recruited sample was below target at 83 PWD and their informal caregivers, we obtained smaller standard deviations than estimated for the TUG and the estimated smallest detectable change of a value of 4 was outside the 95% confidence interval (-2,17, 3.81) between the trial arms, suggesting that the testing on the TUG was adequately powered.
Instrumented Timed Up and Go test
Data were collected by a single investigator trained in the use of the iTUG during a visit at the PWD and caregiver's home. Each performed a standardized iTUG once the sensor was placed on their low back: rising for a chair, walking 3 meters to a mark on the floor, turning, then walking back to the chair, and returning to sitting. Participants were free to choose turning direction and a pragmatic approach to the particular chair available within the individuals’ home was used. Participants were encouraged to not use arm rests and were permitted to use their ‘usual’ walking aid, however only one person in the older PWD group used a cane during testing. Previous studies have demonstrated excellent reliability for total time to complete Timed-Up and Go and a range of minimal detectable change (MDC) values across a variety of clinical presentations, ranging from ICC=0.81, MDC=4.4s for persons with early Dementia 31 to ICC=0.97, MDC 1.1s for older adults with osteoarthiritis.32
A trunk mounted inertial measurement unit (Balance Sensor, THETAmetrix, Portsmouth, UK) was mounted over the middle of the individuals low back, reinforced with an elasticated strap. The Inertial Measurement Unit (IMU) incorporates a triaxial accelerometer and triaxial gyroscope providing linear accelerations and rotational velocities at 30Hz. Excellent reliability of the IMU has been previously reported.33 Data were exported to matlab for feature extraction. An automated algorithm was used to detect the sub-phases of the iTUG, as has been previously described.28 Data were filtered at 6Hz to remove high frequency noise. Temporal events and sub-phases were identified by local maxima/minima detection and zero crossing points from the respective acceleration and gyroscope traces. In addition, walking periods of the iTUG were used to compute measures of regularity (ACstep, ACstride) and symmetry (ratio) using autocorrelation methods.34 The variables investigated in this study are presented in table 1.
Table 1. Definition of the Variable Used in this Study.
| iTUG Variables | Definition |
|---|---|
| Standing Acc (m/s/s) | Peak acceleration of the most vertical axis of the accelerometer (meters per second per second) |
| Sit2stand time (s) | Duration of time taken to complete the sit to stand period (seconds) |
| Time Walk 1 (s) | Duration of time taken to complete the first walk period (seconds) |
| ACStepWalk1 | Regularity of steps in the first walk period as a correlation |
| ACStrideWalk1 | Regularity of strides in the first walk period as a correlation |
| RatioWalk1 | Symmetry of gait determined by step and stride ratio |
| Turning Vel1 (°/s) | Peak velocity of the first turning period (degrees per second) |
| Time for turn (s) | Duration of time taken to complete the first turn (seconds) |
| Time Walk 2 (s) | Duration of time taken to complete the second walk period (seconds) |
| ACStepWalk2 | Regularity of steps in the second walk period as a correlation |
| ACStrideWalk2 | Regularity of strides in the second walk period as a correlation |
| RatioWalk2 | Symmetry of gait determined by step and stride ratio |
| Turning Vel2 (°/s) | Peak velocity of the first turning period (degrees per second) |
| Total Time (s) | Duration of time taken to complete the iTUG (seconds) |
Statistical analysis
All data were assessed for normality using Shapiro-Wilk testing from which appropriate parametric and non-parametric statistics were followed. The two groups (PWD and Caregivers) differed significantly in age (p = 0.015), therefore two approaches were adopted to control for age. First, age brackets were defined for each group of young (<70 years (range age of youngest participant to 69.9 years)), middle (70-79 years), and older (80+ (range 80 to oldest participant)), from which independent t-tests or Mann Whitney-U tests were used to identify differences between PWD and Caregivers. These groupings were chosen to ensure similar amounts of the sample were present in each group (30%, 33% and 37% respectively). In addition multivariate regression models for each iTUG variable were explored with age, gender and diagnosis (PWD or Caregiver) as independent variables. To avoid type 1 error due to the multiple testing, we used Bonferroni corrections to reduce alpha from 0.05 to 0.004. Cohen’s-d effect sizes were calculated to determine the magnitude of difference between groups.
Results
The frequency of gender and age range for the groups is presented in table 2. There was a propensity for caregivers to be female with a higher proportion in the younger age group compared to PWD. No significant differences were evident in the ages of the groups following categorization (table 3), however there was a wider age range for the young group of caregivers.
Table 2. The Distribution of Gender Across the Age Groups.
| Caregiver | PWD | |||
|---|---|---|---|---|
| <70 years old | Males | Females | Males | Females |
| Number | 4 | 31 | 10 | 4 |
| Age range (years) | 67.0 – 69.1 | 43.3 – 69.0 | 59.5 - 69.4 | 59.0 – 68.7 |
| 70-79 years old | Males | Females | Males | Females |
| Number | 6 | 19 | 18 | 12 |
| Age range (years) | 73.2 – 78.9 | 70.4 – 79.9 | 70.6 – 79.8 | 70.5 – 79.8 |
| 80+ years old | Males | Females | Males | Females |
| Number | 8 | 15 | 22 | 17 |
| Age range (years) | 82.2 – 88.0 | 80.0 – 96.0 | 80.1 – 97.5 | 80.0 – 90.5 |
Abbreviations: PWD; People with Dementia
Table 3. Comparison of iTUG Data for Under 70 year old Caregivers and PWD.
| <70 year old group | Caregivers (n = 35) | PWD (n = 14) | ||||
|---|---|---|---|---|---|---|
| Median | IQR | Median | IQR | P-value | Effect size | |
| Age (years) | 63.1 | 11.10 | 62.50 | 8.57 | 0.188 | 0.521 |
| Standing Accb (m/s/s) | -2.05 | 0.73 | -1.73 | 0.50 | 0.009 | 0.837 |
| Sit2stand time (s) | 1.62 | 0.54 | 1.90 | 0.40 | 0.209 | 0.378 |
| Time Walk 1 (s) | 2.60 | 0.67 | 3.02 | 2.19 | 0.068 | 0.769 |
| ACStepWalk1 | 0.81 | 0.29 | 0.75 | 0.46 | 0.790 | 0.074 |
| ACStrideWalk1 | 0.77 | 0.26 | 0.58 | 0.60 | 0.302 | 0.551 |
| RatioWalk1b | 1.04 | 0.27 | 1.26 | 1.13 | 0.030 | 0.975 |
| Turning Vel1 (°/s) | 2.32 | 0.57 | 2.11 | 1.00 | 0.198 | 0.447 |
| Time for turna (s) | 2.06 | 0.48 | 2.46 | 0.75 | 0.001 | 0.698 |
| Time Walk 2b (s) | 2.10 | 0.89 | 2.67 | 2.85 | 0.050 | 0.839 |
| ACStepWalk2 | 0.87 | 0.29 | 0.52 | 0.59 | 0.030 | 0.797 |
| ACStrideWalk2 | 0.75 | 0.44 | 0.66 | 0.72 | 0.954 | 0.122 |
| RatioWalk2 | 1.05 | 0.29 | 0.93 | 0.83 | 0.129 | 0.450 |
| Turning Vel2 (°/s) | 2.84 | 0.97 | 2.43 | 1.53 | 0.324 | 0.338 |
| Total Timeb (s) | 12.00 | 2.21 | 15.41 | 8.30 | 0.016 | 0.807 |
Abbreviations: PWD; People with Dementia, ACStepWalk; Step Regularity, ACStrideWalk; Stride regularity, RatioWalk; Gait Symmetry, Vel; Velocity
Denotes statistical significance, P < 0.004
Denotes large effect size, d > 0.8.
Age group categorization and statistical results for comparison between caregivers and PWD are presented in tables 3-5. The younger age group comparisons (table 3) demonstrated that, following Bonferroni correction, only time taken to complete the turn was significantly different between the groups. However large effect sizes (d>0.8), which in contrast to statistical testing are used to quantify the magnitude of the observed effect, were observed between the two groups for standing acceleration, time to complete walking phase 2, total TUG time and gait asymmetry during walk phase 1. The middle age group comparisons (table 4) demonstrated that PWD took significantly longer to complete the total iTUG. The older age group comparisons (table 5) demonstrated that PWD took significantly longer to complete sit-to-stand, both walking phases, total iTUG and turning velocity was lower. Large effect sizes were determined for time to complete both walking phases and the total iTUG in addition to turning velocity and step regularity during walk phase 2.
Table 5. Comparison of iTUG Data for 80+ year old Caregivers and PWD.
| 80+ years old group | Caregivers (n = 23) | PWD (n = 39) | ||||
|---|---|---|---|---|---|---|
| Median | IQR | Median | IQR | P-value | Effect size | |
| Age (years) | 83.20 | 4.53 | 84.30 | 4.40 | 0.221 | 0.160 |
| Standing Acc (m/s/s) | -1.52 | 0.97 | -1.57 | 0.62 | 0.889 | 0.029 |
| Sit2stand timeab (s) | 1.68 | 0.81 | 2.22 | 0.70 | 0.001 | 0.757 |
| Time Walk 1 ab (s) | 3.17 | 2.01 | 4.44 | 2.48 | 0.001 | 0.847 |
| ACStepWalk1 | 0.63 | 0.44 | 0.58 | 0.41 | 0.565 | 0.157 |
| ACStrideWalk1 | 0.64 | 0.44 | 0.71 | 0.48 | 0.780 | 0.075 |
| RatioWalk1 | 1.00 | 0.71 | 0.95 | 0.24 | 0.675 | 0.095 |
| Turning Vel1 ab (°/s) | 2.18 | 0.61 | 1.71 | 0.47 | 0.001 | 1.010 |
| Time for turn (s) | 2.41 | 0.78 | 2.79 | 0.73 | 0.014 | 0.648 |
| Time Walk 2 ab (s) | 2.44 | 1.07 | 4.16 | 3.94 | <0.001 | 1.122 |
| ACStepWalk2 b | 0.73 | 0.36 | 0.74 | 0.50 | 0.060 | 0.803 |
| ACStrideWalk2 | 0.83 | 0.24 | 0.52 | 0.59 | 0.419 | 0.238 |
| RatioWalk2 | 1.05 | 0.27 | 0.96 | 0.35 | 0.087 | 0.075 |
| Turning Vel2 ab (°/s) | 2.54 | 0.73 | 1.71 | 0.57 | <0.001 | 1.081 |
| Total Time ab (s) | 13.65 | 4.94 | 18.72 | 7.84 | <0.001 | 1.352 |
Abbreviations: PWD; People with Dementia, ACStepWalk; Step Regularity, ACStrideWalk; Stride regularity, RatioWalk; Gait Symmetry, Vel; Velocity
Denotes statistical significance, P < 0.004,
Denotes large effect size, d > 0.8.
Table 4. Comparison of iTUG Data for 70-79 year old Caregivers and PWD.
| 70-79 years old group | Caregivers (n = 25) | PWD (n = 31) | ||||
|---|---|---|---|---|---|---|
| Median | IQR | Median | IQR | P-value | Effect size | |
| Age (years) | 74.15 | 4.50 | 75.60 | 4.70 | 0.116 | 0.197 |
| Standing Acc (m/s/s) | -1.75 | 1.29 | -1.63 | 1.01 | 0.423 | 0.076 |
| Sit2stand time (s) | 1.81 | 0.70 | 2.06 | 0.92 | 0.098 | 0.573 |
| Time Walk 1 (s) | 3.17 | 1.33 | 3.81 | 2.25 | 0.053 | 0.551 |
| ACStepWalk1 | 0.70 | 0.39 | 0.62 | 0.45 | 0.176 | 0.431 |
| ACStrideWalk1 | 0.72 | 0.37 | 0.61 | 0.42 | 0.247 | 0.313 |
| RatioWalk1 | 1.03 | 0.37 | 1.04 | 0.63 | 0.487 | 0.271 |
| Turning Vel1 (°/s) | 2.25 | 0.93 | 1.91 | 0.95 | 0.233 | 0.270 |
| Time for turn (s) | 2.19 | 0.68 | 2.48 | 0.79 | 0.011 | 0.775 |
| Time Walk 2 (s) | 2.65 | 1.41 | 3.75 | 2.03 | 0.021 | 0.667 |
| ACStepWalk2 | 0.72 | 0.32 | 0.61 | 0.44 | 0.498 | 0.269 |
| ACStrideWalk2 | 0.64 | 0.52 | 0.66 | 0.28 | 0.545 | 0.109 |
| RatioWalk2 | 0.96 | 0.47 | 1.03 | 0.50 | 0.957 | 0.150 |
| Turning Vel2 (°/s) | 2.35 | 0.84 | 2.20 | 0.92 | 0.144 | 0.293 |
| Total Timea (s) | 13.62 | 3.98 | 16.53 | 5.84 | 0.002 | 0.741 |
Abbreviations: PWD; People with Dementia, ACStepWalk; Step Regularity, ACStrideWalk; Stride regularity, RatioWalk; Gait Symmetry, Vel; Velocity
Denotes statistical significance, P < 0.004
Denotes large effect size, d > 0.8.
Multivariate regression models were sequentially built to determine if the inclusion of age, gender, and presence/absence of dementia (independent variables) might predict each specific iTUG variable (dependent variable). This sequential process enabled the understanding of the impact of adding each independent variable to the model. Details of the model and contribution are displayed in table 6. Gender made no difference to the predictive capacity of the regression model, but adding dementia diagnosis improved the predictive capacity. The model explained 26% - 33% of the variance of time to complete the walking phases and total time to complete iTUG. Regarding turning, the model explained 21% - 28% of turning velocity and 15% of the variance of time to turn. The higher percentages of variances explained for each significant variable were from the model with dementia diagnosis added.
Table 6. Multivariate Regression Models, Incorporating Age, Gender and Dementia Diagnosis (yes/no).
| Age | Age+ Gender | Age+ Gender+ Diagnosis | ||||
|---|---|---|---|---|---|---|
| Adj R2 | p | Adj R2 | p | Adj R2 | p | |
| Standing Acc | 0.014 | 0.070 | 0.010 | 0.172 | 0.030 | 0.053 |
| Time S2Sa | 0.076 | <0.001 | 0.070 | 0.002 | 0.127 | <0.001 |
| Time Walk1a | 0.111 | <0.001 | 0.134 | <0.001 | 0.278 | <0.001 |
| ACStepWalk1 | 0.021 | 0.036 | 0.016 | 0.100 | 0.017 | 0.128 |
| ACStrideWalk1 | 0.003 | 0.484 | 0.007 | 0.650 | 0.008 | 0.245 |
| RatioWalk1 | 0.001 | 0.298 | 0.004 | 0.492 | 0.022 | 0.089 |
| Turn Vel1a | 0.112 | <0.001 | 0.127 | <0.001 | 0.210 | <0.001 |
| Time for turna | 0.066 | 0.001 | 0.066 | 0.002 | 0.154 | <0.001 |
| Time Walk2a | 0.083 | <0.001 | 0.082 | 0.001 | 0.260 | <0.001 |
| ACStepWalk2a | 0.010 | 0.105 | 0.006 | 0.229 | 0.067 | 0.003 |
| ACStrideWalk2 | 0.002 | 0.411 | 0.008 | 0.711 | 0.010 | 0.717 |
| RatioWalk2 | 0.009 | 0.122 | 0.013 | 0.131 | 0.021 | 0.096 |
| Turn Vel2a | 0.234 | <0.001 | 0.230 | <0.001 | 0.278 | <0.001 |
| Total Time iTUGa | 0.102 | <0.001 | 0.099 | <0.001 | 0.316 | <0.001 |
Abbreviations: ACStepWalk; Step Regularity, ACStrideWalk; Stride regularity, RatioWalk; Gait Symmetry, Vel; Velocity, Adj; Adjusted
Denotes statistical significance, P < 0.004
Discussion
The aims of this study were to explore how age moderates the differences in performance on the instrumented Timed Up and Go test between PWD and their informal caregivers. The novel sensor technology and derived algorithms were capable of quantifying the sub-phases of iTUG and demonstrated that age moderates the differences in iTUG performance observed between PWD and caregivers. As all testing was completed in the individuals homes this offers a significant potential for quantification of performance in clinical practice.
Differences between PWD and caregivers in the youngest age group (<70 years of age) were demonstrated for time for turn, and may offer early indications of deterioration in function. PWD took around 20% longer (0.4 seconds) to complete the turn than caregivers of a similar age. Turning has been identified as a complex task requiring a coordinated sequence of axial rotations of multiple body parts,35 all of which may require longer processing in PWD. In addition, large effect sizes were identified for standing acceleration, gait symmetry, time to walk phase 2 and total time to complete iTUG. These did not meet the stringent criteria for Bonferroni correction, however the magnitude of actual difference was similar (around 20%). Total time to complete was quite variable in PWD (coefficient of variation (calculated by dividing the mean by the standard deviation) = 54%), suggesting great variability in performance of the whole iTUG across the group. In addition the observed difference between the groups was 3.4s, slightly below that identified as the MDC31. Standing acceleration was much less variable, while still demonstrating a large effect size between groups, suggesting this could be used as a key performance indicator in younger PWD. Sit-to-stand acceleration may represent early deficits in power from the lower limbs.17
The middle aged group (70-79 years of age) demonstrated total time to complete iTUG was significantly different with the greatest statistical confidence and around a 20% real difference. This finding is in line with numerous studies identifying deficits in total time to complete iTUG in frail older adults36,37 and in fallers,38 and now in those with dementia. This demonstrates that total time identifies performance difference even when controlled for age. Despite this the effect size was only moderate and the magnitude of difference between groups was below that of the MDC identified previously.31 Total time to complete TUG has been strongly correlated to time to complete walking phases.28 Therefore, it is highly likely that walking speed is a significant contributor to overall iTUG time. In addition turning time demonstrated the largest effect size (0.775) suggesting this slowed speed was sustained from the younger age group, despite the fact that the caregivers in this age group took slightly longer to turn than the younger caregivers.
The divergence in metrics is much clearer in the older group (80+ years of age). Walking durations, turning velocity and sit to stand time were prolonged suggesting a strong down gearing of movement velocity with actual differences of around 25%. This suggests a loss of around a quarter of these higher functions. This is corroborated by effect size analysis where large effect sizes were determined for the above variables, except sit to stand time. There seems to be little difference in these values for the caregivers as they age, but a sharper drop in performance in PWD noted between the 70 and 80+ year old group. It is possible that this is due to a progression in dementia impacting on performance, illustrating that as the disease progresses the performance declines resulting in the divergence observed after 80 years of age. However, despite our previous study demonstrating that walk time and TUG time were correlated with dementia disease severity (as measured using the M-ACE), the strength of the relationship was weak at 0.25-0.28 and non-existent for other sub-phases of the iTUG.28 This throws into question the mechanism behind such marked deficits observed in the older age group. The current findings illustrate that it is not simply age, and our previous findings illustrate it is not simply disease severity that reduces performance on the iTUG and its subtasks. This is corroborated further by our multivariate models that showed an increased variance explained by adding dementia diagnosis to age as the predictor variables. This suggests that a more complex multifactorial explanation is required. It is possible that fear of falling is important as this correlated with iTUG in PWD, explaining up to 20% of total time to complete iTUG.28 However perhaps deconditioning plays an important role also,39,40,41 where activity down-regulation results in a reduction in physical capacity.
The findings from this study have a number of important clinical implications. Firstly the results demonstrate that the sub-phases of the iTUG are able to detect differences in PWD from their age-matched caregivers, thus separating out those changes due to age versus those due to Dementia. Such deficits are different depending on the age bracket investigated with most divergence evident over 80 years of age. Deficits identified in the under 70 year old bracket were quite pronounced and may offer early clinical targets for intervention to minimize functional decline. These deficits were only visualized with the addition of the instrumented Timed Up and Go test, such as standing acceleration, suggestive of a decline in lower limb power. As individuals age, other sub-phases may offer clues for functional deficits demonstrating the importance of an assessment which offers a detailed breakdown of the iTUG. This probably reflects the underlying complexity pertaining to the iTUG with its differing physiological constructs underpinning its differing sub-phases. The ability to assess these complex tasks not only differentiates between performances of PWD but is able to evaluate early changes in function offering highly specific clinical rehabilitation targets.
Despite being commercially available, devices and algorithms for quantifying iTUG are not commonplace in clinical practice. In the absence of such methods, the findings of this study can still guide clinicians in their approach to assessment and management of PWD. Understanding that specific elements, such as sit to stand, may be the first clues to deterioration of function in PWD under 70 years old. Many assessment strategies exist to quantify performance of sit to stand and the findings of the current study encourages clinicians to integrate such assessment for PWD under 70 years old. Assessment of individuals in the 70-79 demonstrate the original total time to complete TUG is able to detect differences and so should remain as an important variable for assessing function in this age group. Therefore clinicians should be mindful that age moderates the performance of the iTUG differently for caregivers and PWD.
There are several limitations with this study. A cross-sectional design was used therefore no inferences about causation can be made. The age group categorizations resulted in unequal group sizes. The data were collected in individual’s homes therefore a pragmatic approach was taken towards chair height and a standard 3m Timed Up and Go was adopted which can affect the ability to use autocorrelation analysis methods. Future research could aim to determine if the identified performance deficits in iTUG sub-phase are modified with rehabilitation in PWD such as to prevent functional decline.
Conclusion
This study demonstrated significant differences in performance of specific elements of the iTUG in PWD compared to caregivers matched for age. These include time for turn in the <70 year olds, total iTUG time in the 70-79 year olds and sit to stand time, walk time, total time to complete iTUG and turning velocity for the >80 year olds.
Acknowledgements
The authors acknowledge colleagues from the TACIT Trial, from which the data for this paper was drawn. In particular, the authors acknowledge Chris Hayward, Wendy Ingram, and Jeanette Sanders (Peninsula Clinical Trials Unit, University of Plymouth), Peter Thomas, Sarah Thomas, and Helen Allen (Bournemouth University Clinical Research Unit, Bournemouth University), Michael Vassallo (Centre of Postgraduate Medical Research and Education, Bournemouth University), James Raftery (Faculty of Medicine, University of Southampton), and Iram Bibi and Yolanda Barrado-Martín (Department of Psychology and Ageing & Dementia Research Centre, Bournemouth University). The authors acknowledge Southern Health NHS Foundation Trust for sponsorship of the trial.
The authors acknowledge the assistance of Mr Andrew Watt in testing the iTUG algorithm and advice received from Dr Shanti Shanker in regard to cognitive testing and our public and patient involvement group on our approach to recruitment and data collection. The authors thank the Alzheimer’s Society for their assistance with publicizing the study, and the support of the National Institute for Health Research Clinical Research Network (NIHR CRN). The authors thank the General Practice surgeries in Wessex that assisted with recruitment, and the three main recruitment sites: Memory Assessment Research Centre, Southern Health NHS Foundation Trust (Principal Investigator: Brady McFarlane), Memory Assessment Service, Dorset HealthCare University NHS Foundation Trust (Principal Investigator: Kathy Sheret and then Claire Bradbury), and Research and Improvement Team and Older People’s Mental Health Service, Solent NHS Trust (Principal Investigator: Sharon Simpson). We also thank the Trial Steering Committee for their expert input (Independent Chair: Frances Healey, NHS Improvement).
Funding
SRN was funded by a National Institute for Health Research (NIHR) Career Development Fellowship Award for this project. The data for this study was drawn from independent research funded by the NIHR. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.
Footnotes
Declaration of Interests: JMW has consulted with THETAmetrix, the company from which the sensor was purchased. SRN declares no conflicts of interest.
References
- 1.World Health Organization. WHO global report on falls prevention in older age. [Accessed November 5, 2019];2007 https://www.who.int/ageing/publications/Falls_prevention7March.pdf.
- 2.Samaras N, Chevalley T, Samaras D, Gold G. Older patients in the emergency department: A review. Ann Emerg Med. 2010;56(3):261–269. doi: 10.1016/j.annemergmed.2010.04.015. [DOI] [PubMed] [Google Scholar]
- 3.Rubenstein LZ. Falls in older people: Epidemiology, risk factors, and strategies for prevention. Age Ageing. 2006;35(S2):ii37–ii41. doi: 10.1093/ageing/afl084. [DOI] [PubMed] [Google Scholar]
- 4.Laxton CE, Freeman CJ, Todd CJ, et al. Morbidity at 3 months after hip fracture: Data from the East Anglian audit. Health Trends. 1997;29(2):55–60. 1997. [Google Scholar]
- 5.Finnes TE, Meyer HE, Falch JA, Medhus AW, Wentzel-Larsen T, Lofthus CM. Secular reduction of excess mortality in hip fracture patients >85 years. BMC Geriatr. 2013;13:e25. doi: 10.1186/1471-2318-13-25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kerr J, Marshall S, Godbole S, et al. The relationship between outdoor activity and health in older adults using GPS. Int J Environ Res Public Health. 2012;9(12):4615–4625. doi: 10.3390/ijerph9124615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Peel NM. Epidemiology of falls in older age. Can J Aging. 2011;30(1):7–19. doi: 10.1017/S071498081000070X. [DOI] [PubMed] [Google Scholar]
- 8.Butler R, Radhakrishnan R. Dementia: Systematic review 1001. [Accessed November 5, 2019];BMJ Clinical Evidence. 2012 2012, doi: clinicalevidence.bmj.com/x/systematic-review/1001/overview.html. [PMC free article] [PubMed]
- 9.Prince M, Wimo A, Guerchet M, Ali G-C, Wu Y-T, Prina M. World Alzheimer report 2015. The global impact of dementia: An analysis of pravalence, incidence, cost and trends. [Accessed Novemeber 5, 2019]; https://www.alz.co.uk/research/world-report-2015.
- 10.Taylor M, Delbaere K, Close J, Lord S. Managing falls in older people with cognitive impairment. Ageing Health. 2012;8(6):573–588. [Google Scholar]
- 11.Winter H, Watt K, Peel NM. Falls prevention interventions for community-dwelling older persons with cognitive impairment: A systematic review. Int Psychogeriatr. 2013;25(2):215–227. doi: 10.1017/S1041610212001573. [DOI] [PubMed] [Google Scholar]
- 12.Dramé M, Fierobe F, Lang P, et al. Predictors of institution admission in the year following acute hospitalisation of elderly people. J Nutr Health Aging. 2011;15(5):399–403. doi: 10.1007/s12603-011-0004-x. [DOI] [PubMed] [Google Scholar]
- 13.Secretariat MA. Caregiver- and patient-directed interventions for dementia: An evidence-based analysis. Ont Health Technol Assess Ser. 2008;8(4):1–98. [PMC free article] [PubMed] [Google Scholar]
- 14.Watkin L, Blanchard M, Tookman A, Sampson E. Prospective cohort study of adverse events in older people admitted to the acute general hospital: Risk factors and the impact of dementia. Int J Geriatr Psychiatry. 2012;27(1):76–82. doi: 10.1002/gps.2693. 2012. [DOI] [PubMed] [Google Scholar]
- 15.Podsiadlo D, Richardon S. The Timed Up and Go: A test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39:142–148. doi: 10.1111/j.1532-5415.1991.tb01616.x. [DOI] [PubMed] [Google Scholar]
- 16.Shumway-Cook A, Brauer S, Wollacott M. Predicting the probability for falls in community-dwelling older adults using the Timed Up & Go test. Phys Ther. 2000;80(9):896–903. [PubMed] [Google Scholar]
- 17.Barry E, Galvin R, Keogh C, Horgan F, Fahey T. Is the Timed Up and Go test a useful predictor of falls in community-dwelling older adults: a systematic review and meta-analysis. BMC Geriatr. 2014;14:14. doi: 10.1186/1471-2318-14-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Shoene D, Wu SM, Mikolaizak AS, Menant JC. Discriminative ability and predictive validity of the Timed Up and Go test in identifying older people who fall: systematic review and meta-analysis. J Am Geriatr Soc. 2013;61(2):202–208. doi: 10.1111/jgs.12106. [DOI] [PubMed] [Google Scholar]
- 19.Beauchet O, Fantino B, Allali G, Muir SW, Montero-Odasso M, Annweiler C. Timed Up and Go test and the risk of falls in older adults: a systematic review. J Nutr Health Aging. 2011;15(10):933–938. doi: 10.1007/s12603-011-0062-0. [DOI] [PubMed] [Google Scholar]
- 20.Salarian A, Horak FB, Zampieri C, Carlson-Kuhta P, Nutt JG, Aminian K. iTUG, a sensitive and reliable measure of mobility. IEEE Trans Neural Syst Rehabil Eng. 2010;18(3):303–310. doi: 10.1109/TNSRE.2010.2047606. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Mellone S, Tacconi C, Chiari L. Validity of a smartphone-based instrumented Timed Up and Go. Gait Posture. 2012;36(1):163–165. doi: 10.1016/j.gaitpost.2012.02.006. [DOI] [PubMed] [Google Scholar]
- 22.Van Lummel RC, Walgaard S, Hobert MA, et al. Intra-rater, interrater and test retest reliability of an instrumented Timed Up and Go (iTUG) test in patients with Parkinson’s Disease. PLoS One. 2016;11(3) doi: 10.1371/journal.pone.0151881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wuest S, Masse F, Aminian K, Gonzenbach R, de Bruin ED. Reliability and validity of the inertial sensor-based Timed Up and Go test in individuals affected by stroke. J Rehabil Res Dev. 2016;53(5):599–610. doi: 10.1682/JRRD.2015.04.0065. [DOI] [PubMed] [Google Scholar]
- 24.Caronni A, Sterpi I, Antoniotti P, et al. Criterion validity of the instrumented Timed Up and Go test: A partial least square regression study. Gait Posture. 2018;61:287–293. doi: 10.1016/j.gaitpost.2018.01.015. [DOI] [PubMed] [Google Scholar]
- 25.Mirelman A, Weiss A, Buchman AS, Bennett DA, Giladi N, Hausdorff JM. Association between performance of Timed Up and Go subtasks and mild cognitive impairment: further insights into the links between cognitive and motor function. J Am Geriatr Soc. 2014;62(4):673–678. doi: 10.1111/jgs.12734. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Galan-Mercant A, Cuesta-Vargas AI. Clinical frailty syndrome assessment using inertial sensors embedded in smartphones. Physiol Meas. 2015;36(9):1929–1942. doi: 10.1088/0967-3334/36/9/1929. [DOI] [PubMed] [Google Scholar]
- 27.Palmerini L, Mellone S, Avanzolini G, Valzania F, Chiari L. Quantification of motor impairment in Parkinson's diseased using an instrumented Timed Up and Go test. IEEE Trans Neural Syst Rehabil Eng. 2013;21(4):664–673. doi: 10.1109/TNSRE.2012.2236577. [DOI] [PubMed] [Google Scholar]
- 28.Williams JM, Nyman SR. Association between the instrumented Timed Up and Go test and cognitive function, fear of falling and quality of life in community dwelling people with dementia. J Frailty Sarco Falls. 2018;3(4):185–193. doi: 10.22540/JFSF-03-185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nyman S, Hayward C, Ingram W, et al. A randomised controlled trial comparing the effectiveness of Tai Chi alongside usual care with usual care alone on the postural balance of community-dwelling people with dementia: Protocol for The TACIT Trial (TAi ChI for people with dementia) BMC Geriatr. 2018;18:e263. doi: 10.1186/s12877-018-0935-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Hsieh S, McGrory S, Leslie F, Dawson K, Ahmed S, Butler C. The Mini-Addenbrooke’s Cognitive Examination: A new assessment tool for dementia. Dement Geriatr Cogn Disord. 2015;39:1–11. doi: 10.1159/000366040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Blackwood J. Reliability, validity and minimal detectable change in the Timed Up and Go and five times sit to stand tests in older adults with early cognitive loss. J Physiother Rehabili. 2017;1:1. [Google Scholar]
- 32.Alghadir A, Anwer S, Brismee J-M. The reliability and minimal detectable change of Timed Up and Go test in individuals with grade 1-3 knee osteoarthritis. BMC Musculoskelet Disord. 2015;16:174. doi: 10.1186/s12891-015-0637-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Williams JM, Dorey C, Clark S, Clark C. The within-day and between-day reliability of using sacral accelerations to quantify balance performance. Phys Ther Sport. 2016;17:45–50. doi: 10.1016/j.ptsp.2015.04.002. [DOI] [PubMed] [Google Scholar]
- 34.Moe-Nilssen R, Helbostad JL. Estimation of gait characteristics by trunk accelerometry. J Biomech. 2004;37(1):121–126. doi: 10.1016/s0021-9290(03)00233-1. [DOI] [PubMed] [Google Scholar]
- 35.Buchman AS, Boyle PA, Leurgans SE, Barnes LL, Bennet DA. Cognitive function is associated with the development of mobility impairment in community-dwelling elders. Am J Geriatr Psychiatry. 2011;19(6):571–580. doi: 10.1097/JGP.0b013e3181ef7a2e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Ansai JH, Farche ACS, Rossi PG, de Andrade LP, Nakagawa TH, Takahashi ACM. Performance of different Timed Up and Go subtasks in frailty syndrome. J Geriatr Phys Ther. 2017;42(4):287–293. doi: 10.1519/JPT.0000000000000162. [DOI] [PubMed] [Google Scholar]
- 37.Eagles D, Perry JJ, Sirois MJ, et al. Timed Up and Go predict functional decline in older patients presenting to the emergency department following minor ttrauma. Age Ageing. 2017;46(2):214–218. doi: 10.1093/ageing/afw184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ansai JH, Andrade LP, Nakagawa TH, Rebelatto JR. Performances on the Timed Up and Go test and subtaks between fallers and non-fallers in older adults with cognitive impairment. Arq Neuropsiquiatr. 2018;76(6):381–386. doi: 10.1590/0004-282X20180055. [DOI] [PubMed] [Google Scholar]
- 39.Suzuki M, Ohyama N, Yamada K, Kanamori M. The relationship between fear of falling, activities of daily living and quality of life among elderly individuals. Nurs Health Sci. 2002;4(4):155–61. doi: 10.1046/j.1442-2018.2002.00123.x. 70. [DOI] [PubMed] [Google Scholar]
- 40.Wijlhuizen GJ, de Jong R, Hopman-Rock M. Older persons afraid of falling reduce physical activity to prevent outdoor falls. Prev Med. 2007;44(3):260–264. doi: 10.1016/j.ypmed.2006.11.003. 71. [DOI] [PubMed] [Google Scholar]
- 41.Choi K, Ko Y. Characteristics Associated With Fear of Falling and Activity Restriction in South Korean Older Adults. J Aging Health. 2015;27(6):1066–83. doi: 10.1177/0898264315573519. [DOI] [PubMed] [Google Scholar]