Mobility refers to the ability of an individual to move (independently or with aid) from one place to another.1 26% to 33% of stroke survivors will have mobility impairment.2,3 Mobility has been assessed by clinicians and researchers in terms of both functional mobility and living space mobility, and current research on mobility in stroke patients has focused on the former.4 Mobility is exceptionally important for reducing patients' physical disability and maintaining independent living,5 and it is also an important predictor of physical function, independence, and mortality,6 so an accurate assessment of mobility can help nurses identify potential functional deficits and risks, and provide a basis for care planning and resource allocation.7,8 Because of the dynamic and multidimensional nature of the concept of “mobility,”9 the assessment of mobility lacks a criterion standard and mostly uses self-reporting and performance measurement, where forward walking speed is a common parameter and indicator in performance measurement.9 However, some studies have shown that challenging dynamic assessments are more revealing of mobility deficits in patients with relatively good functional recovery.8 Backward walking test is one of the challenging items in the assessment of mobility deficits, and Taulbee et al8 argued that the backward walking speed is more sensitive to screen for mobility deficits. However, there is heterogeneity in the cutoff values for discriminating mobility deficits by walking speed in different populations,10 and it is believed that defining a cutoff value for mobility in a specific population is better to achieve the target management of the patient function.9 Therefore, it is unclear whether backward walking speed still has a high diagnostic value in the stroke patient population, and whether there is a difference in the identification of mobility deficits in stroke patients between backward maximum walking speed (BMWS) and backward comfortable walking speed (BCWS). Moreover, the determination of the respective cutoff values is an issue that needs further investigation.
In this study, a data-driven approach was used to derive the optimal cutoff values for BMWS and BCWS to distinguish mobility impairment in stroke patients, which provides a basis for the clinical formulation of targeted rehabilitation measures and also further expands the value of backward walking speed in clinical applications.
Stroke patients admitted to the hospital's neurology, neurosurgery, and neurorehabilitation departments from July 2021 to January 2022 were selected in chronological order of admission. Inclusion criteria are as follows: has been given a diagnosis of stroke by magnetic resonance imaging or computed tomographic scanning, being 18 years or older, physically able to get out of bed and walk independently for at least 10 meters, were able to follow simple instructions to complete tasks, and were able to give informed consent. The exclusion criteria are as follows: had any other neurological diagnoses that limit balance mobility (eg, Parkinson disease and Alzheimer disease), had any other severe musculoskeletal injury or cardiopulmonary diagnosis (acute myocardial infarction, unstable angina, etc), with severely impaired vision and hearing, or were unable to complete all 6 assessments during the test. This study was approved by the university medical ethics committee.
A general questionnaire was designed to collect information including age, sex, disease duration, stroke type, dominant side, and so forth.
Three-Meter Backward Walk Test
A 3-meter marker was made on a flat surface, with 1 meter reserved at each end. The patient stood at the starting point and was told to walk backward as fast as possible to the end point (until the end point was crossed by the side of the initial foot lift), and the exertion time was recorded.11 The total distance divided by the time consumed was used as the BMWS; the total distance divided by the time consumed by the patient walking backward to the end at the usual comfortable speed was recorded as the BCWS.
Timed Up and Go Test
The Timed Up and Go Test (TUGT) is a common evaluation method for gait balance and transfer ability.12 It consists of 4 functional tests of sitting-standing, walking, turning (around obstacles), and standing-sitting. The patient sat on a chair with a backrest of 46 cm in height at the starting point, walked forward at the fastest speed for a distance of 3 meters after the command “Start,” went around the obstacle, and then returned to the chair and sat on it until the hips touched the chair surface again. The consumption time was recorded as the test results.
Berg Balance Scale
The Berg Balance Scale (BBS) is a tool to assess an individual's static and dynamic balance. It consists of 14 items with a score of 0 to 56, with higher scores indicating better balance.
It includes 10 items on dressing, defecation, walking, and going upstairs and downstairs to evaluate the extent of the activities of daily living.13 The total score ranges from 0 to 100, and the higher score indicates better ability to perform activities of daily living.
Ten-Meter Walk Test
It is a common tool for assessing walking speed in patients with neurological disease.14 A 10-meter marking was made on flat ground, with 5 meters reserved at each end, and patients walk forward at the fastest speed. The total distance divided by the time consumed was used as the forward maximum walking speed (FMWS). All the tools applied in this study have good reliability and validity and have been widely used in stroke patients.14–17
Power analysis and Sample Size 15.0 software was used to calculate the sample size. The expected area under the receiver operating characteristic (ROC) curve analysis would be 0.7, the α value would be .05, and intended power of 0.8 indicated that 48 participants would be sufficient.7 Liston and Brouwer18 argued that the sample size for cluster analysis can be calculated according to 5*2k (where k is the number of independent variables); it was estimated that a total sample size of 40 would have been adequate. A total sample size of 98 participants was included in this study.
In addition to the evaluation of general demographic data, all included stroke patients underwent 6 functional tests, including the 3MBW (which contains 2 data items: BMWS and BCWS), the TUGT, the BBS, the Barthel Index (BI), and the 10-Meter Walk Test. The previously mentioned tests collected data on BMWS, BCWS, transferability, balance, activities of daily living, and FMWS, respectively. All walking tests were performed 3 times, with the first time as a pretest to improve patient adaptation, and the average of the last 2 tests was recorded as the final test result. The study nurses were trained and qualified by physiotherapists to administer the previously mentioned tests.
SPSS statistics 26.0 software and MedCalc19.5.6 were used for statistical analysis, and Kolmogorov-Smirnov was used for normality testing. The quantitative data followed a normal distribution and were presented as mean (SD), and independent-samples t tests were performed to compare differences between groups; the median (interquartile range) was used if the data were not normally distributed, and the Mann-Whitney U test was used for comparisons between groups.
The K-means clustering algorithm was used to analyze the data in the study. We selected K = 2 according to the purpose of the study, with Euclidean distance as the distance metric to generate the central points of each cluster. The clustering result of FMWS, TUGT, and BBS divided the stroke patients into 2 clusters: the high–mobility function group and the low–mobility function group. Then, BI score was used to evaluate the effectiveness of classification. Receiver operating characteristic curves were generated to evaluate the diagnostic value of BMWS and BCWS for discriminating the patients with mobility deficits based on the clustering result. The DeLong method was used to compare the difference between the 2 ROC curves and determine the optimal cutoff values of the 2 indicators. P values < .05 were considered as statistically significant.
A total of 98 stroke patients, aged 23 to 84 years (67 men and 31 women; 89 with ischemic stroke, 9 with hemorrhagic stroke, and 19 with a history of stroke), were included in this study (Table 1). The clustering results distinguished 76 patients as the high–mobility function group and 22 as the low–mobility function group. Further comparison of the characteristics of the 2 categories of patients showed that there were significant differences in the results of all 6 functional tests (P < .05) except for age and duration of disease (Table 1).
TABLE 1 -
Comparison of General Geographic and Function Tests Between the 2 Groups
|Variables, Median (IQR)
||All Stroke Patients (98)
||Low-Mobility Group (22)
||High-Mobility Group (76)
||Comparison Between Groups (t/Z, P)
|Disease duration, d
|FMWS, mean (SD), m/s
|BMWS, mean (SD), m/s
|BCWS, mean (SD), m/s
Abbreviations: BBS, Berg Balance Scale; BCWS, backward comfortable walking speed; BI, Barthel Index; BMWS, backward maximum walking speed; FMWS, forward maximum walking speed; IQR, interquartile range; TUGT, Timed Up and Go Test.
aMann-Whitney U test.
The ROC curves showed that the area under the curve of both indicators was greater than 0.9 (see Figure, Supplemental Digital Content, available at https://links.lww.com/JNN/A438). The optimal cutoff value of BMWS was 0.3 m/s (sensitivity, 0.93; specificity, 0.86), and the optimal cutoff value of BCWS was 0.27 m/s (sensitivity, 0.8; specificity, 0.91; Table 2). Z test for diagnostic efficacy between the 2 indicators showed no statistical significance (P > .05; Table 2).
TABLE 2 -
Diagnostic Accuracy Measures for Cutoffs of BMWS and BCWS for Mobility Deficit
|Variables and Cutoff Value, m/s
||AUC and 95% CI
|BMWS at 0.3
|BCWS at 0.27
Abbreviations: AUC, area under the curve; BCWS, backward comfortable walking speed; BMWS, backward maximum walking speed; CI, confidence interval; +LR, positive likelihood ratio; −LR, negative likelihood ratio; YI, Youden's index.
aNull hypothesis: true area = 0.5.
bComparison of AUC of BMWS and BCWS.
Our results confirm the high diagnostic value of backward walking speed for motor impairment in stroke patients and also suggest the optimal cutoff values for BMWS and BCWS to identify mobility impairment in stroke patients. K-means clustering analysis is the most common clustering algorithm in the data-driven field.7 Forward maximum walking speed, TUGT, and BBS were used to assess the walking ability, transferability, and balance ability of stroke patients, respectively, all of which are indicators of physical status related to mobility. In this study, K-means cluster analysis was used to cluster stroke patients into high– and low–mobility function groups based on these 3 categories of data, which better fit the distribution characteristics of the sample itself. The BI is a commonly used scale for assessing activities of daily living,19 with 40% of the scores related to assessing mobility, and other studies have shown that there are many similar properties between the concepts of activity of daily living and mobility,1 so the BI score can also reflect patients' mobility function. Table 1 shows that the BI scores were statistically different between the 2 groups (P < .05), which could verify the discriminatory validity of this clustering model.
The area under the ROC curve > 0.9 is considered to represent high diagnostic performance.20 The areas under the curve of both BMWS and BCWS in this study were >0.9, confirming the high diagnostic value of both for identifying stroke patients with mobility deficits. The likelihood ratio is an indicator that reflects both sensitivity and specificity but is more stable than them. The negative likelihood ratio for BMWS was <0.1, indicating a significantly higher likelihood of excluding mobility deficits when BMWS > 0.3 m/s; the positive likelihood ratio for BCWS was 8.83, indicating a higher likelihood of having mobility deficits for patients with BCWS < 0.27 m/s. For mobility assessment, high sensitivity is more clinically meaningful than specificity because it helps nurses to identify stroke patients with mobility deficits in time to take measures to prevent complications such as falls and injuries. However, the difference between the sensitivity and specificity of the 2 indicators in this study was not significant. In addition, we conducted Z tests to compare the diagnostic efficacy of the 2 indicators, and we found that there was no difference in the diagnostic value of BMWS compared with BCWS (P > .05). However, Clark et al21 concluded that maximum walking speed is more advantageous than comfortable walking speed when assessing age-related decline in mobility function due to neuromuscular injury. Our results are inconsistent with the previously mentioned findings, which may be because the participants in Clark et al's study were well-functioning healthy elderly people and the monitoring index was forward walking speed, whose maximum walking speed and comfortable walking speed differed significantly, and therefore the maximum walking speed reflected more differences in physiological function.22 However, the mean values of BMWS and BCWS in this study were 0.54 and 0.39 m/s, respectively, and the difference between them was only 0.17 m/s. Furthermore, the BMWS and BCWS values were the same in 5 patients during the trial, demonstrating that the additional neuromuscular and perceptual demands during backward walking can lead to the inability of patients with mobility deficits to accelerate. So the difference between BMWS and BCWS in this study was not significant, leading to no difference in the diagnostic value of the 2 indicators.
The optimal cutoff values for differentiating mobility deficits in stroke patients were 0.3 and 0.27 m/s for BMWS and BCWS in this study, respectively. Taulbee et al8 also confirmed the effectiveness of BCWS as an outcome indicator to screen the elderly population with mobility deficits in the community, but its cutoff value (0.73 m/s) was higher than the results in this study (0.27 m/s), probably because the patients in this study were in the poststroke recovery period and the poorer functioning of the patients led to a lower judgment threshold as well. Because few studies have attempted to compare the diagnostic capabilities of BMWS and BCWS in mobility function, it is not possible to compare them with the results of this study. In summary, both BMWS (<0.3 m/s) and BCWS (<0.27 m/s) are effective in identifying patients with poststroke mobility deficits. However, some studies have suggested that there is an increased risk of falls when walking in the backward direction23; therefore, it is recommended that the BCWS be selected for testing in clinical practice to ensure patient safety.
This study has limitations. Individual mobility is determined by a combination of environmental, cognitive, psychological, and physical functioning factors.1 Forward maximal walking speed, BBS, and so forth only respond to motor and balance ability, which may not be sufficient to fully reflect the patient's level of mobility function, and other socioecological parameters can be added subsequently for measurement. In addition, patients were not allowed to use assistive devices in the test, so a proportion of patients with poor functional levels have been excluded from this study, and the included patients are not representative of all poststroke patients.
Because rehabilitation measures to improve mobility are considered the most effective strategy for stroke patients24 and mobility is associated with depression,25 caregiver burden,26 and adverse outcomes, backward walking speed can guide nurses in developing rehabilitation plans, assessing rehabilitation outcomes, and developing risk intervention strategies in community rehabilitation. The 3MBW is currently a common tool for testing backward walking speed. Because of its ease of operation and the low professionalism required of the testers, the tool has clinical application and promotion value. Assessment is considered as one of the emerging trend areas for neuroscience nursing research in the next 3 to 5 years,27 and we suggest that future studies increase cross-sectional comparisons between multiple measurement tools to obtain the most appropriate patient assessment tools for use in stroke.
The results of this study showed that the relevant measures of backward walking speed (including BMWS and BCWS) can contribute to the early identification of mobility deficits in stroke patients and can provide a basis for poststroke rehabilitation treatment and care. There was no difference in diagnostic efficacy between the two, but considering that backward walking increases the risk of falls, we recommend that BCWS be selected as a screening tool in clinical practice.
We acknowledge the support from the Second Affiliated Hospital of Wenzhou Medical University. We are grateful to the clinical staff who participated in this study.
1. Moulton E, Wilson R, Deluzio K. Movement and mobility: a concept analysis. ANS Adv Nurs Sci
. 2019;42(4):E11–E23. doi:10.1097/ANS.0000000000000247
2. Crichton SL, Bray BD, McKevitt C, Rudd AG, Wolfe CD. Patient outcomes up to 15 years after stroke
: survival, disability, quality of life, cognition and mental health. J Neurol Neurosurg Psychiatry
3. Obembe AO, Simpson LA, Sakakibara BM, Eng JJ. Healthcare utilization after stroke
in Canada—a population based study. BMC Health Serv Res
. 2019;19(1):192. doi:10.1186/s12913-019-4020-6
4. Rössler R, Bridenbaugh SA, Engelter ST, et al. Recovery of mobility function and life-space mobility after ischemic stroke
: the MOBITEC-Stroke
study protocol. BMC Neurol
. 2020;20(1):348. doi:10.1186/s12883-020-01920-z
5. Lininger MR, Warren M, Knecht J, Verheijde J, Tyler B, Tompkins J. Clinical instruments for bedside functional assessment: convergent validity among the AM-PAC '6-Clicks' and BMAT. J Clin Nurs
. 2021;30:2048–2056. doi:10.1111/jocn.15761
6. Buvarp D, Rafsten L, Sunnerhagen KS. Predicting longitudinal progression in functional mobility after stroke
: a prospective cohort study. Stroke
. 2020;51(7):2179–2187. doi:10.1161/STROKEAHA.120.029913
7. Kwong PWH, Ng SSM. Cutoff score of the lower-extremity motor subscale of Fugl-Meyer assessment in chronic stroke
survivors: a cross-sectional study. Arch Phys Med Rehabil
. 2019;100(9):1782–1787. doi:10.1016/j.apmr.2019.01.027
8. Taulbee L, Yada T, Graham L, et al. Use of backward walking speed to screen dynamic balance and mobility deficits in older adults living independently in the community. J Geriatr Phys Ther
. 2021;44(4):189–197. doi:10.1519/JPT.0000000000000290
9. Chung J, Demiris G, Thompson HJ. Instruments to assess mobility limitation in community-dwelling older adults: a systematic review. J Aging Phys Act
. 2015;23(2):298–313. doi:10.1123/japa.2013-0181
10. Miller ME, Magaziner J, Marsh AP, et al. Gait speed and mobility disability: revisiting meaningful levels in diverse clinical populations. J Am Geriatr Soc
. 2018;66(5):954–961. doi:10.1111/jgs.15331
11. Carter V, Jain T, James J, Cornwall M, Aldrich A, de Heer HD. The 3-m backwards walk and retrospective falls: diagnostic accuracy of a novel clinical measure. J Geriatr Phys Ther
. 2019;42(4):249–255. doi:10.1519/JPT.0000000000000149
12. Son H, Park C. Effect of turning direction on Timed Up and Go test results in stroke
patients. Eur J Phys Rehabil Med
. 2019;55(1):35–39. doi:10.23736/s1973-9087.18.05202-4
13. Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J
14. Moore JL, Potter K, Blankshain K, Kaplan SL, O'Dwyer LC, Sullivan JE. A core set of outcome measures for adults with neurologic conditions undergoing rehabilitation: a clinical practice guideline. J Neurol Phys Ther
. 2018;42(3):174–220. doi:10.1097/npt.0000000000000229
15. Abit Kocaman A, Aydoğan Arslan S, Uğurlu K, Katırcı Kırmacı Zİ, Keskin ED. Validity and reliability of the 3-meter backward walk test in individuals with stroke
. J Stroke Cerebrovasc Dis
. 2021;30(1):105462. doi:10.1016/j.jstrokecerebrovasdis.2020.105462
16. 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
17. Hsueh I, Lee M, Hsieh C. Psychometric characteristics of the Barthel activities of daily living index in stroke
patients. J Formos Med Assoc
18. Liston RA, Brouwer BJ. Reliability and validity of measures obtained from stroke
patients using the Balance Master. Arch Phys Med Rehabil
. 1996;77(5):425–430. doi:10.1016/s0003-9993(96)90028-3
19. Balkaya M, Cho S. Optimizing functional outcome endpoints for stroke
recovery studies. J Cereb Blood Flow Metab
. 2019;39(12):2323–2342. doi:10.1177/0271678X19875212
20. Swets JA. Measuring the accuracy of diagnostic systems. Science
. 1988;240(4857):1285–1293. doi:10.1126/science.3287615
21. Clark DJ, Manini TM, Fielding RA, Patten C. Neuromuscular determinants of maximum walking speed in well-functioning older adults. Exp Gerontol
. 2013;48(3):358–363. doi:10.1016/j.exger.2013.01.010
22. Wu T, Zhao Y. Associations between functional fitness and walking speed in older adults. Geriatr Nurs
. 2021;42(2):540–543. doi:10.1016/j.gerinurse.2020.10.003
23. Hawkins KA, Balasubramanian CK, Vistamehr A, et al. Assessment of backward walking unmasks mobility impairments in post-stroke
community ambulators. Top Stroke Rehabil
. 2019;26(5):382–388. doi:10.1080/10749357.2019.1609182
24. Sahely A, Giles D, Sintler C, Soundy A, Rosewilliam S. Self-management interventions to improve mobility after stroke
: an integrative review. Disabil Rehabil
. 2022;31:1–18. doi:10.1080/09638288.2022.2028019
25. Bhimani R, Chappuis D, Mathiason MA, Anderson LC. Spasticity, pain, and fatigue: are they associated with functional outcomes in people with stroke
. 2022;47(2):60–71. doi:10.1097/RNJ.0000000000000357
26. Shamoun C, Harris MA. Impact of stroke
survivors' physical disability on the presence of caregiver burden: a systematic review. J Neurosci Nurs
. 2022;54(4):146–150. doi:10.1097/JNN.0000000000000651
27. Hinkle JL, Alexander S, Avanecean D, et al. Research priority setting: the current landscape of neuroscience nursing research. J Neurosci Nurs
. 2022;54(2):55–60. doi:10.1097/JNN.0000000000000638