Abstract
Objective
To measure repeatability within and between sessions of the Timed Up and Go (TUG) test in geriatric dogs and investigate associations between TUG times and home activity as measured by accelerometry.
Methods
Geriatric dogs were recruited in April 2024 to perform 2 sessions of TUG testing 20 days apart. Each session consisted of 3 trials separated by 1-minute rest periods. Dogs wore collar-mounted accelerometers between sessions. Client-specific outcome measures, including Canine Brief Pain Inventory and Liverpool Osteoarthritis for Dogs, were compared between sessions to ensure functional stability. Repeatability was evaluated using intraclass correlation and Bland-Altman approaches.
Results
30 dogs were enrolled, of which 24 (median age, 13 years; median weight, 26 kg) met inclusion criteria for analyses. For all within- and between-sessions testing, coefficients of repeatability were < 1.63 seconds, intraclass correlation coefficients were > 0.92, and coefficients of variation were < 10%. Vigorous activity was negatively correlated with TUG times for both sessions (ρ = −0.5). No differences in client-specific outcome measures between sessions and no other correlations between activity measures and TUG times were detected.
Conclusions
The TUG test is repeatable within and between sessions in functionally stable geriatric dogs; however, differences of 2 seconds or less may represent normal variation. In-clinic TUG times correlate to vigorous activity at home, and both measures may be reflective of short-duration maximal capacity effort.
Clinical Relevance
This study further supports the TUG test as a reliable and valid measure of canine geriatric function.
Geriatric patients comprise an important cohort of both human and veterinary medicine. The Timed Up and Go (TUG) is a validated objective clinical functional test initially developed for geriatric people to help aid in prognosis and patient monitoring but has since been applied to numerous populations and conditions.1–3 Recently, our research demonstrated high inter-rater agreement and provided criterion validity of the canine TUG test in geriatric dogs.4 A reliable clinical test demands both strong inter-rater agreement and test-retest repeatability. However, to our knowledge, no studies exist investigating test-retest repeatability within a session or between sessions for the canine TUG.
Accelerometry objectively measures motion frequency, duration, and intensity in 3 dimensions, outputting useful data, such as step count, vector magnitude, and the intensity of activities performed.5–7 Consequently, the degree of activity a dog exhibits, as measured by accelerometry, may be employed as a clinical assessment tool for physical function or response to treatment.8,9 However, relationships between the TUG test and accelerometry (both of which provide different objective avenues to assess canine mobility—1 as a snapshot in clinic and the other at home over time) remain unexplored.
Therefore, we sought to investigate test repeatability by measuring both within-session (3 TUG trials separated by 1-minute rest periods) and short-term between-session (repeating the 3 TUG trials 20 days later) TUG times in geriatric dogs. Validated canine client-specific outcome measures (CSOMs) were used at both TUG sessions (A and B) to help ensure a functionally stable dog population for comparing between-session repeatability. We also sought to investigate associations between TUG times and accelerometry data with a previously validated canine accelerometer.5 We hypothesized that the TUG test would demonstrate strong within-session and between-session reliability. We further hypothesized that higher TUG times would correlate to lower activity at home as measured by accelerometry. Finally, we predicted that the CSOMs for Liverpool Osteoarthritis in Dogs (LOAD) and the Canine Brief Pain Inventory (CBPI) for pain sensitivity (psCBPI) and the CBPI pain interference (piCBPI) would remain unchanged between sessions.
Methods
Study population, recruitment, inclusion, and exclusion
Dogs were recruited from the Cornell College of Veterinary Medicine from a previous geriatric testing group.4 Recruitment efforts began 1 month before the scheduled testing dates, May 18 and June 8, 2024. Dogs were required to be tolerant of wearing an additional collar with the accelerometry monitor attached and be present for data collection. To help ensure consistency throughout the study, caregivers were given careful instructions not to change their dog’s environment, home lifestyle, diet, pain management, or exercise habits. Any unavoidable change in consistency was to be reported for review and potential exclusion. Veterinarian-diagnosed musculoskeletal or neurological diseases contributing to decreased mobility were recorded. Dogs that underwent surgery within the preceding 3 months or had any changes in pain medications, including therapeutic joint injections, within the week preceding or during the study period were excluded. Study approval was obtained by the Cornell University Veterinary Clinical Studies Committee for IACUC exemption, and informed client consent was collected prior to enrollment. All enrollees were provided a $100 incentive to participate for each TUG session, and no study results were shared with owners at any time. The demographic data collected included sex, age, body weight, body length, and body condition at the first TUG session.
TUG testing procedures
Dogs underwent TUG testing in a quiet indoor environment at room temperature on a level, nonslip surface without distractions or owners present as previously described by McMullin et al4 (Supplementary Video S1). Collars and harnesses were removed, and a slip lead was placed around each dog’s neck. A muzzle (JorVet Nylon Plastic Basket Muzzle; Jorgensen Laboratories) was placed prior to testing dogs that appeared averse to handling, provided it did not result in observable behavior changes that would affect test results. To normalize for dog size, the body length (the distance from the tip of the nose to the base of the tail) of each dog was recorded, and the start and finish lines at 10 body lengths apart were marked against a wall. The handler positioned each dog in sternal recumbency, ensuring the front paws aligned with the starting line, and applied gentle pressure to the withers to maintain this position. A second person (caller) stood 5 yards beyond the finish and used verbal encouragement and hand signals to motivate the dogs to move at their quickest chosen pace, but treats or toys were not allowed as rewards. The test was run in a direction back toward the doors in which the dog entered, and the observer timing the TUG was positioned well outside the test path, looking directly across the finish line toward the wall. The time was recorded in seconds, to 2 decimal places, starting immediately upon finishing the callout “ready, set, go” and stopping when the dog’s nose crossed the finish line. Upon “go,” the handler and caller would encourage the dog to rise and begin moving at its chosen pace in a straight line toward the finish, using the wall as a guide on 1 side of its body and the handler on the other. Gentle tension on the lead was allowed to help direct the dog, but no additional forces that restricted or pulled were permitted. After completing the run, the dog was walked back to the starting line and allowed a 1-minute rest period, during which it could elect to stand, sit, or lie down. This procedure was repeated 2 additional times, for a total of 3 trials of the TUG test. Dogs that completed the first session returned for a second identical session 20 days later. Dogs were considered incapable of completing the test if during any trial they fell, dragged, exhibited behavior that obviously interfered with the result, or failed to finish in under 60 seconds. The same 5 handlers, experienced in conducting the TUG test at each position, participated in each TUG session and rotated through positions based on fatigue and availability.
Client-specific outcome measures
Similarly to the prior TUG study, owners completed web-based surveys regarding their dog’s functional and pain status over the 7 days prior to each TUG testing session using 2 previously validated questionnaires: LOAD and CBPI.10,11 The LOAD is a 13-item questionnaire scored on a 5-point Likert scale (0 to 4), with higher scores indicating greater mobility impairment. The CBPI questionnaire is divided into 3 sections, with 2 sections being examined for influence on comfort and function: (1) 4 questions were used to assess pain severity (psCBPI), and (2) 6 questions were used to evaluate pain interference (piCBPI). The psCBPI and piCBPI sections use an 11-point scale (0 to 10). The total scores for the psCBPI and piCBPI were divided by the number of questions (4 and 6, respectively) to produce a final mean score for each section. Higher scores for each component indicate greater pain severity or more pain interference with daily activities, respectively.
Accelerometry testing procedures
The accelerometry unit (Actigraph wGT3X-BT; Actigraph Holdings LLC) was mounted on its own separate collar (TrainPro replacement 3/4" dog collar strap band with double buckle loop; Grade-A Global LLC). The collar was appropriately and securely fitted to each dog with the size marked, positioned around the neck with the accelerometer centered ventrally, and shown to the client to help ensure consistency. Clients were instructed to remove this collar only during any anticipated water activity (grooming, underwater treadmill therapy, or swimming) and record the corresponding date, time, duration, and nature of activity for each removal. Data were continuously recorded between testing periods in dogs on 3 axes. Days where the collar was only partially worn, such as testing days, were excluded. The accelerometer sampled data at a rate of 30 times/s (30 Hz). The vertical component of the recorded activity was converted into numerical values using accelerometry software, generating activity counts for each 60-second epoch. Step counts and vector magnitude were employed as different methods of investigating cumulative movement. Activity intensities were categorized into 3 levels—sedentary, light to moderate, and vigorous—based on vector magnitude counts as previously defined.5 As such, activities like underwater treadmill therapy or swimming were classified as “light to moderate,” whereas any remaining lapse in time, like grooming appointments, were marked as “sedentary.” For instance, if a collar was recorded as removed for 45 minutes and an underwater treadmill session lasted 25 minutes, the remaining 20 minutes were marked as “sedentary.” Periods when the collars were reportedly removed were crossvalidated with nonwear periods using the algorithm developed by Choi et al.6
Statistics
To determine whether TUG was affected by age, body weight, or body condition score, the TUG times were plotted for each dog against these respective demographics, and then correlation was assessed using the Spearman ρ and Kendal τ. To ensure that dogs did not, as a group, worsen or improve function over the 2 sessions, TUG times were compared between sessions using a Wilcoxon signed rank test.
To examine the repeatability and reproducibility of TUG testing, a number of statistical tests were performed. Dogs that did not complete both sessions, had recorded times > 60 seconds for any test run, experienced behavioral interference with their TUG test, or dragged themselves or fell during testing were excluded.
To examine within-session repeatability, the intraclass correlation coefficient was calculated for each session, using the within-dog mean TUG times from each session employing the (3,k) model12; the coefficient of repeatability was determined for the within-dog mean TUG times from each session, as proposed by Bland and Altman,13 using a free online plotting tool14; and, finally, the within-dog coefficient of variation was compared using a Wilcoxon signed-rank test.
To examine between-session repeatability, mean TUG times were plotted for each session, creating a visual representation of their differences. The intraclass correlation coefficient was calculated using the within-dog mean TUG times from each session using the (3,k) model; determined the coefficient of repeatability for the within-dog mean TUG times from each session; finally, the within-session coefficient of variation was compared using a Wilcoxon signed rank test. Using within-session mean values would reduce the variability between sessions, but as the test usually requires multiple runs, it was determined that this was an acceptable approach.
To examine whether TUG times were correlated with accelerometry variables, the mean TUG times were plotted for each session against 5 within-dog variables measured by the accelerometer: median daily sedentary time, median daily light-to-moderate activity time, median daily vigorous activity time, median daily total number of steps, and median daily vector magnitude counts. The plots were visually examined for each session to look for a clear relationship. For any plot with an obvious relationship, the Spearman ρ and Kendal τ were calculated as a measure of correlation.
Finally, to compare the CSOMs between sessions, the differences in scores for each dog were plotted for each of the 3 measures. The scores were then formally compared between sessions using Wilcoxon signed rank tests.
Results
Population
Thirty geriatric dogs were recruited in order of willingness to participate and availability for TUG repeatability testing and collar-mounted accelerometer monitoring over a 3-week time span. Six dogs were excluded from TUG evaluation: 2 dogs were euthanized before the second session, 1 dog had values exceeding 60 seconds, 2 dogs had dragging hindquarters, and 1 dog was distracted with otitis externa and stopped midway to shake its head. Thus, TUG data from 24 dogs were analyzed, with a median age of 13.0 years (range, 9 to 16; quartiles, 11.0 and 15.0), a median weight of 26 kg (range, 5 to 49; quartiles, 19 and 33), and a median body condition score of 5/9 (range, 3 to 8; quartiles, 5 and 6); 9 were male, 15 were female, and 15 had a prior orthopedic or neurological diagnosis that could inhibit mobility. No association between TUG times and bodyweight (TUG A ρ = 0.08, τ = 0.05, P = .7; TUG B ρ = 0.009, τ = 0.003, P = .97) or between TUG times and body condition score (TUG A ρ = −0.03, τ = −0.02, P = .9; TUG B ρ = −0.22, τ = −0.17, P = .25) was observed. A moderate association between TUG times and age (TUG A ρ = 0.58, τ = 0.44, P = .002; TUG B ρ = 0.57, τ = 044, P = .003) was observed.
Client-specific outcome measure results
All CSOMs (LOAD, psCBPI, and piCBPI) for the 24 dogs were analyzed in TUG repeatability. In comparing the CSOMs between TUG A and B sessions, no evidence of a change in mobility or comfort was noted, and, therefore, no further dogs were dropped from the repeatability testing data set (Table 1).
Results of client-specific outcome measures (CSOMs; Canine Brief Pain Inventory pain severity [psCBPI], Canine Brief Pain Inventory pain interference [piCBPI], and Liverpool Osteoarthritis in Dogs [LOAD]) conducted during Timed Up and Go (TUG) repeatability testing sessions performed on May 18, 2024, and June 8, 2024.
| CSOM | Session | Median | Minimum | Maximum | 25th percentile | 75th percentile | P value |
|---|---|---|---|---|---|---|---|
| psCBPI | A | 2.0 | 0.0 | 7.5 | 0.3 | 3.3 | .17 |
| B | 1.4 | 0.0 | 7.5 | 0.2 | 3.3 | ||
| piCBPI | A | 1.2 | 0.0 | 9.7 | 0.0 | 3.5 | .99 |
| B | 1.4 | 0.0 | 9.2 | 0.2 | 3.6 | ||
| LOAD | A | 20 | 0 | 41 | 12 | 24 | .98 |
| B | 19 | 1 | 36 | 10 | 26 |
Values are reported as median, minimum, maximum, and the 25th and 75th percentiles.
Results are included from 24 dogs with valid TUG testing trials from both sessions (3 TUG trials/session) and demonstrate no differences in CSOMs between sessions. Both psCBPI and piCBPI reflect a mean of 4 and 6 questions, respectively, from an 11-point scale (0 to 10), with higher values representing more pain and reduced function. The LOAD is a 13-item questionnaire additive scored on a 5-point scale (0 to 4), where higher scores indicate worse mobility.
TUG test repeatability
The cohort had a wide distribution of TUG times for each session, ranging mostly between 2 and 10 seconds, and TUG times did not differ between sessions (P = .28; Figure 1). Each session had excellent within-session intraclass correlation coefficients of 0.99 (95% CI, 0.98 to 1.00). The first TUG session had a coefficient of repeatability of 1.49 seconds (95% CI, 1.25 to 1.86); the second TUG session had a coefficient of repeatability of 1.02 seconds (95% CI, 0.85 to 1.28; Table 2). The second session had slightly smaller coefficients of variation than the first session (median, 7.4% vs 4.6%; P = .03). However, only 6 tests over either session (5 dogs; 12.5% of all tests) had coefficients of variation ≥ 11%, and only 3 were > 13%.
Scatter plot showing mean Timed Up and Go (TUG) results by dog for 2 sessions of repeatability testing. Each vertical circle pair corresponds to mean TUG time results (in seconds) per dog, with the black circle and white circle representing May 18, 2024 (session A), and June 8, 2024 (session B), respectively. Results are included for 24 dogs with valid testing trials (3 TUG trial repetitions in each session, with 1-minute rest periods between trials).
Citation: American Journal of Veterinary Research 86, 6; 10.2460/ajvr.25.02.0041
Coefficient of repeatability (CR; Bland-Altman analysis) and intraclass correlation coefficient (ICC) results from within-session testing (3 TUG trials separated by 1 minute of rest) and between-session testing (repeating the first session trials 3 weeks later). Timed Up and Go testing, session dates, and population are the same as described in Table 1.
| Repeatability | Session | Coefficient | 95% CI | |
|---|---|---|---|---|
| Between session | ICC (range, 0–1) | NA | 0.92 | 0.83–0.96 |
| CR (s)a | NA | 1.63 | 1.26–2.28 | |
| Within session | ICC (range, 0–1) | A | 0.99 | 0.98–1.00 |
| B | 0.99 | 0.98–1.00 | ||
| CR (s) | A | 1.49 | 1.25–1.86 | |
| B | 1.02 | 0.85–1.28 | ||
There were 23 dogs used for this analysis as 1 dog was removed as an outlier.
Intersession repeatability was also excellent, with a between-session intraclass correlation coefficient of 0.92 (95% CI, 0.82 to 0.96) and a coefficient of repeatability of 1.63 seconds (95% CI, 1.26 to 2.28; Table 2). One outlier was excluded from the between-session coefficient of repeatability Bland-Altman analysis (a dog that had the longest TUG times and a difference between sessions of 3.4 seconds; Supplementary Figures S1 and S2). The absolute differences between the mean TUG A and mean TUG B times were examined—of the remaining 23 dogs, 22 had absolute differences < 1.5 seconds, whereas 1 dog had an absolute difference of 2.1 seconds.
Associations between TUG times and accelerometry measures
Accelerometry data are presented for the same 24 dogs with valid TUG data (Table 3). Visual examination of plots of accelerometry variables against TUG times failed to reveal any obvious relationships for all accelerometry variables other than vigorous activity (Figure 2), which was negatively correlated with TUG times for both sessions (ρ = −0.5, τ = −0.35; P = .015; Table 4).
Accelerometry results, including data from 24 dogs measured over the 3-week testing period (as described in Table 1).
| Daily activity | Median | Minimum | Maximum | 25th percentile | 75th percentile |
|---|---|---|---|---|---|
| Sedentary | 1,282 | 1,183 | 1,343 | 1,260 | 1,315 |
| Light-moderate | 148 | 94 | 248 | 119 | 163 |
| Vigorous | 9 | 3 | 25 | 6 | 11 |
| Steps | 3,884 | 2,311 | 5,989 | 3,106 | 4,454 |
| Vector magnitude | 539,548 | 317,580 | 750,439 | 422,062 | 636,613 |
Values are reported as median, minimum, maximum, and the 25th and 75th percentiles.
The median number of daily minutes dogs spent engaging in sedentary, light-to-moderate, or vigorous activity are presented. Additionally, median daily step counts and median daily vector magnitude counts are presented. Sedentary indicates the median daily minutes of sedentary activity. Light-moderate indicates the median daily minutes of light-to-moderate activity. Vigorous indicates the median daily minutes of vigorous activity. Steps indicates the median total steps per day. Vector magnitude indicates the median vector magnitude counts per day.
Scatter plots showing a curvilinear relationship between median daily minutes of vigorous activity and TUG time in seconds for the first session (A) and second session (B). The dotted line represents the line of best fit, highlighting a negative correlation between TUG times and vigorous activity for both sessions (ρ = −0.5, τ = −0.35; P = .015). Sessions and population as described in Figure 1.
Citation: American Journal of Veterinary Research 86, 6; 10.2460/ajvr.25.02.0041
Results for Spearman ρ and Kendal τ correlations between accelerometry parameters (as described in Table 3) correlated to TUG times from sessions A and B with data included from 24 dogs, highlighting significant correlations regarding daily time of vigorous activity and TUG time. Timed Up and Go testing, session dates, and population are the same as described in Table 1.
| TUG A | P value | TUG B | P value | |
|---|---|---|---|---|
| Sedentary vs TUG | ||||
| Spearman ρ | 0.009 | .950 | –0.020 | .930 |
| Kendal τ | 0.003 | .980 | –0.030 | .840 |
| Light-moderate vs TUG | ||||
| Spearman ρ | 0.070 | .740 | 0.100 | .650 |
| Kendal τ | 0.060 | .670 | 0.080 | .580 |
| Vigorous vs TUG | ||||
| Spearman ρ | –0.490 | .015 | –0.510 | .015 |
| Kendal τ | –0.350 | .015 | –0.340 | .015 |
| Steps vs TUG | ||||
| Spearman ρ | 0.150 | .480 | 0.200 | .340 |
| Kendal τ | 0.090 | .530 | 0.150 | .290 |
| Vector magnitude vs TUG | ||||
| Spearman ρ | –0.100 | .650 | –0.090 | .670 |
| Kendal τ | –0.080 | .600 | –0.060 | .690 |
Bold indicates significance.
Discussion
Our study demonstrates that both within-test and between-test repeatability were strong across multiple statistical measures, particularly the intraclass correlation coefficient, supporting our first hypothesis (Table 2). The within-test and between-test coefficients of repeatability suggest that changes of 2 seconds or less may simply represent normal variation for an individual with static physical function and TUG times of 10 seconds or less.
Our sample of 24 geriatric dogs received no reported changes in treatment, environment, home lifestyle, diet, pain management, or exercise habits during the study. As predicted, their mobility issues remained static over the study period as evidenced by the lack of change in the CSOMs, psCBPI, piCBPI, and LOAD (Table 1). Previous investigators have correlated changes in these outcome measures with TUG testing; therefore, the lack of change helped assure that our study cohort remained stable for repeatability analysis. Interestingly, multiple dogs changed their scores by 1.5 for psCBPI, 2.5 for piCBPI, and 4 for LOAD, not necessarily in any particular direction, potentially reflecting random variability of these measures. Previous investigators have suggested that decreases of less than or equal to 1 unit for psCBPI or 2 units for piCBPI indicate a response to therapy (and, presumably, an increase > 1 unit indicates worsening) in dogs with osteoarthritis.10 Similarly, a minimally clinically important difference for LOAD has been suggested to be 4 units for orthopedic disease and osteoarthritis.15,16 Our findings in a small sample of dogs (of which 15 had confirmed mobility issues) that did not undergo any treatment during the study suggests that changes of this magnitude are potentially too small and within the range of random variation. However, our findings must not be overinterpreted as our study was focused on function in a heterogenous group of geriatrics and not a homogeneous group of dogs suffering from a specific mobility disease.
Our population captured a wide range (2 to 10 seconds) of TUG times among participants (Figure 1). Therefore, our data provide repeatability results for various levels of canine mobility, especially when considering a previously proposed 4-second separation between highly functional and less functional dogs.4 Future studies could focus on measuring TUG repeatability within geriatric dogs by degree of function (high, moderate, and low) or by specific disease. Theoretically, a cohort of dogs with a specific disease may affect function similarly and lead to greater test reliability; however, geriatric populations often suffer from multiple comorbidities affecting function, confounding assessment. Furthermore, the diagnosis of a disease often fails to translate to functional change.1,17,18 In human medicine, within-session repeat testing can result in TUG time increases during the session (in patients that easily fatigue) or TUG time decreases (in patients that learn to perform more efficiently).3,19–21 Also, geriatric people with longer TUG times tend to have more variability in their performances between test sessions, similar to our observation of 1 dog that had the longest TUG times,3,22 which we considered an outlier and removed from between-session coefficient of repeatability analysis (Supplementary Figures S1 and S2). Given this singular observation, TUG time repeatability might be even larger (worse) for dogs with very long TUG times (eg, > 12 seconds) and, therefore, might require an even greater change in CSOMs to demonstrate true improvement or worsening. Furthermore, since many geriatric patients suffer from chronic progressive disease affecting mobility, longer term TUG times would be expected to increase, but this hypothesis has yet to be explored. Greater age, unlike body weight and body condition, was positively correlated with longer TUG times, like our previous study.4 These findings are unsurprising given that our patient population is a subset of the prior study. The compounding impacts of various comorbidities and cellular senescence due to aging could worsen TUG results; however, age itself should not impact repeatability for an individual. Lastly, both testing environment and procedural factors can influence TUG repeatability in human medicine23; therefore, further research is ideal to determine if environment, personnel, and equipment (such as leash type) can influence the canine TUG.
We observed that TUG times correlated moderately with time spent partaking in vigorous activity but no other activity variables, partially supporting our second hypothesis (Figure 2; Table 4). In people, moderate-to-vigorous activity was most strongly associated with better functional testing outcomes (including TUG), whereas sedentary behaviors have been associated with lower outcomes.24,25 In dogs, less intense physical activity has been previously documented with greater age.26 Light-to-moderate activity might be further reduced by a pet’s lifestyle and environment.18 We were unsurprised by the lack of association between sedentary behavior and TUG times because the environment and home lifestyle of a pet can limit movement. Specifically, many dogs spend much of the day at home sedentary while the family is out of the house regardless of the pet’s fitness capacity.27 In our study, dogs were sedentary a median of more than 21 hours daily (nearly 90% of their time) and engaged in vigorous activity for less than 1% of the day (Table 3). Such a high proportion of sedentary behavior will naturally limit the impact of accelerometry on capturing a dog’s full mobility profile. On the other hand, there remains the opportunity to achieve short bursts of high-intensity activity during nonsedentary time that could explain our findings. For example, a relatively more mobile dog could still choose to participate or be offered small bouts of vigorous exercise (like sprinting or play). In other words, both the in-clinic TUG test and the at home accelerometer may overlap in their ability to measure short bursts of maximal function in older dogs.
Our study has some limitations. The population of participants were derived from a prior study4 that acknowledged that the recruited dogs may not be representative of typical geriatric populations given their association with the Cornell University College of Veterinary Medicine and not being overtly overweight or obese. We did not repeat weighing and body condition scoring dogs at the second TUG session because significant change impacting the study seemed unlikely over such a short period. The total enrollee number and order in this study were based on client willingness to participate and not power calculations. Ideally, the population would have been recruited to satisfy a broad spectrum of geriatric function. Regardless, functional diversity within our population was reflected by a wide and relatively evenly distributed spectrum of TUG times and approximately 60% of the dogs having had a prior diagnosis for an orthopedic and/or neurological condition inhibiting mobility. In the future, a larger population aimed at capturing more dogs, including those at slower TUG times, would allow better exploration of the potential influence of varying function upon TUG time repeatability. Comparative medical examinations were not conducted after each TUG session to better ensure that the dog populations were functionally stable. There was concern that orthopedic examination might exacerbate lameness or pain in some of our dogs, affecting home activity, CSOM scoring, or even repeat TUG testing. Regardless, the lack of difference in multiple CSOMs and the short between-session duration helped to minimize this limitation. Excluding dogs that received therapeutic joint injections or changes in pain medication within a week of TUG testing was based on the prior manuscript that provided criterion validity to the TUG test. However, a longer period of exclusion from the time of any pain management changes, particularly joint injections, may be more ideal given that some injectates have different durations of effect and can also cause transient soreness.28 Having stable CSOMs throughout the study period helped mitigate this limitation; however, we also retrospectively investigated our population and found that only 1 dog received a therapeutic joint injection procedure within 5 months of testing (injected 1 month prior to the study), making this unlikely to influence study outcomes. An ideal period of rest between TUG tests remains unknown; however, within this study, a 1-minute rest period still resulted in excellent test repeatability. Some dogs participated in consistent weekly underwater treadmill therapy. Such exercise was classified as light to moderate based on prior validation5 and, therefore, would not have influenced TUG time correlations with vigorous activity. Individual dog behavior affecting such a trial cannot be completely eliminated as a possibility. Despite being experienced in all aspects of the TUG test, handler differences might still influence the test.
In conclusion, TUG demonstrated strong repeatability within a session and between sessions separated by 20 days in a stable geriatric dog population. Such findings further support the TUG as a reliable clinical instrument for geriatric function; however, more research is warranted to determine how repeatability is affected by different canine populations, environment, handling, and equipment. Over the short term, changes in TUG time of less than 2 seconds may represent natural variability in dogs scoring under 10-second times. Lastly, increased time of vigorous activity at home was associated with improved TUG times in clinic, revealing some overlap in these different objective measures of mobility and lending more criterion validity to the canine TUG test.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
None reported.
Disclosures
Dr. Rishniw is a member of the JAVMA Scientific Review Board, but was not involved in the editorial evaluation of or decision to accept this article for publication.
No AI-assisted technologies were used in the composition of this manuscript.
Funding
Funding was provided by Nestlé Purina PetCare in support of Cornell University’s Canine Healthy Aging and Mobility Program.
ORCID
M. Rishniw https://orcid.org/0000-0002-0477-1780
C. W. Frye https://orcid.org/0000-0002-5531-1012
References
- 1.↑
Podsiadlo D, Richardson S. The timed “Up & Go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39(2):142–148. doi:10.1111/J.1532-5415.1991.TB01616.X
- 2.
Freter SH, Fruchter N. Relationship between Timed “Up and Go” and gait time in an elderly orthopaedic rehabilitation population. Clin Rehabil. 2000;14(1):96–101. doi:10.1191/026921500675545616
- 3.↑
Christopher A, Kraft E, Olenick H, Kiesling R, Doty A. The reliability and validity of the Timed Up and Go as a clinical tool in individuals with and without disabilities across a lifespan: a systematic review. Disabil Rehabil. 2021;43(13):1799–1813. doi:10.1080/09638288.2019.1682066
- 4.↑
McMullin KM, Carney PC, Juran IG, et al. Timed Up and Go demonstrates strong interrater agreement and criterion validity as a functional test in geriatric dogs. Am J Vet Res. 2024;85(8):ajvr.24.03.0062. doi:10.2460/AJVR.24.03.0062
- 5.↑
Yam PS, Penpraze V, Young D, et al. Validity, practical utility and reliability of Actigraph accelerometry for the measurement of habitual physical activity in dogs. J Small Anim Pract. 2011;52(2):86–91. doi:10.1111/J.1748-5827.2010.01025.X
- 6.↑
Choi L, Liu Z, Matthews CE, Buchowski MS. Validation of accelerometer wear and nonwear time classification algorithm. Med Sci Sports Exerc. 2011;43(2):357–364. doi:10.1249/MSS.0B013E3181ED61A3
- 7.↑
Preston T, Baltzer W, Trost S. Accelerometer validity and placement for detection of changes in physical activity in dogs under controlled conditions on a treadmill. Res Vet Sci. 2012;93(1):412–416. doi:10.1016/J.RVSC.2011.08.005
- 8.↑
Brown DC, Boston RC, Farrar JT. Use of an activity monitor to detect response to treatment in dogs with osteoarthritis. J Am Vet Med Assoc. 2010;237(1):66–70. doi:10.2460/JAVMA.237.1.66
- 9.↑
Rialland P, Bichot S, Moreau M, et al. Clinical validity of outcome pain measures in naturally occurring canine osteoarthritis. BMC Vet Res. 2012;8:162. doi:10.1186/1746-6148-8-162/TABLES/5
- 10.↑
Brown DC, Bell M, Rhodes L. Power of treatment success definitions when the Canine Brief Pain Inventory is used to evaluate carprofen treatment for the control of pain and inflammation in dogs with osteoarthritis. Am J Vet Res. 2013;74(12):1467–1473. doi:10.2460/AJVR.74.12.1467
- 11.↑
Walton MB, Cowderoy E, Lascelles D, Innes JF. Evaluation of construct and criterion validity for the “Liverpool Osteoarthritis in Dogs” (LOAD) clinical metrology instrument and comparison to two other instruments. PLoS One. 2013;8(3):e58125. doi:10.1371/JOURNAL.PONE.0058125
- 12.↑
Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med. 2016;15(2):155–163. Published correction appears in J Chiropr Med. 2017;16(4):346. doi:10.1016/J.JCM.2016.02.012
- 13.↑
Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res. 1999;8(2):135–160. doi:10.1177/096228029900800204
- 14.↑
Goedhart J, Rishniw M. BA-plotteR – a web tool for generating Bland-Altman plots and constructing limits of agreement. Res Vet Sci. 2021;137:281–286. doi:10.1016/J.RVSC.2021.05.017
- 15.↑
Innes JF, Morton MA, Lascelles BDX. Minimal clinically-important differences for the ‘Liverpool Osteoarthritis in Dogs’ (LOAD) and the ‘Canine Orthopedic Index’ (COI) client-reported outcomes measures. PLoS One. 2023;18(2):e0280912. doi:10.1371/JOURNAL.PONE.0280912
- 16.↑
Alves JC, Innes JF. Minimal clinically-important differences for the “Liverpool Osteoarthritis in Dogs” (LOAD) and the “Canine Orthopedic Index” (COI) in dogs with osteoarthritis. PLoS One. 2023;18(9):e0291881. doi:10.1371/JOURNAL.PONE.0291881
- 17.↑
Gordon WJ, Conzemius MG, Riedesel E, et al. The relationship between limb function and radiographic osteoarthrosis in dogs with stifle osteoarthrosis. Vet Surg. 2003;32(5):451–454. doi:10.1053/JVET.2003.50051
- 18.↑
Frye C, Carr BJ, Lenfest M, Miller A. Canine geriatric rehabilitation: considerations and strategies for assessment, functional scoring, and follow up. Front Vet Sci. 2022;9:842458. doi:10.3389/FVETS.2022.842458
- 19.↑
Bloch ML, Jønsson LR, Kristensen MT. Introducing a third Timed Up & Go test trial improves performances of hospitalized and community-dwelling older individuals. J Geriatr Phys Ther. 2016;40(3):121–126. doi:10.1519/JPT.0000000000000080
- 20.
Faleide AGH, Bogen BE, Magnussen LH. Intra-session test-retest reliability of the Timed “Up & Go” test when performed by patients with hip fractures. Eur J Physiother. 2015;17(2):89–97. doi:10.3109/21679169.2015.1043579
- 21.↑
Viteckova S, Krupicka R, Dusek P, et al. The repeatability of the instrumented Timed Up & Go test: the performance of older adults and Parkinson’s disease patients under different conditions. Biocybern Biomed Eng. 2020;40(1):363–377. doi:10.1016/J.BBE.2019.12.001
- 22.↑
Nordin E, Rosendahl E, Lundin-Olsson L. Timed “Up & Go” test: reliability in older people dependent in activities of daily living—focus on cognitive state. Phys Ther. 2006;86(5):646–655. doi:10.1093/PTJ/86.5.646
- 23.↑
Lee SP, Dufek J, Hickman R, Schuerman S. Influence of procedural factors on the reliability and performance of the Timed Up-and-Go test in older adults. Int J Gerontol. 2016;10(1):37–42. doi:10.1016/J.IJGE.2015.10.003
- 24.↑
Cruz D, Freire YA, Rosa-Souza FJ, et al. Accelerometer-measured volumes of moderate-to-vigorous physical activity and functional fitness in older adults. J Aging Phys Act. Published online November 11, 2024. doi:10.1123/JAPA.2023-0384
- 25.↑
Semanik PA, Lee J, Song J, et al. Accelerometer-monitored sedentary behavior and observed physical function loss. Am J Public Health. 2015;105(3):560–566. doi:10.2105/AJPH.2014.302270.
- 26.↑
Lee H, Collins D, Creevy KE, Promislow DEL. Age and physical activity levels in companion dogs: results from the Dog Aging Project. J Gerontol A Biol Sci Med Sci. 2022;77(10):1986–1993. doi:10.1093/GERONA/GLAC099
- 27.↑
Scaglia E, Cannas S, Minero M, Frank D, Bassi A, Palestrini C. Video analysis of adult dogs when left home alone. J Vet Behav. 2013;8(6):412–417. doi:10.1016/J.JVEB.2013.04.065
- 28.↑
Miller AV, Carney PC, Markmann A, Frye CW. Retrospective analysis describes safety of therapeutic joint injections in dogs. J Am Vet Med Assoc. 2023;261(3):397–402. doi:10.2460/JAVMA.22.11.0483
