The 6-minute walk test (6MWT) is commonly used in people with stroke undergoing rehabilitation.1–3 Although originally developed and validated as a submaximal oxygen consumption test for individuals with cardiac or pulmonary disease,4,5 the 6MWT is a valid6–11 and reliable12,13 measure of walking endurance and is highly recommended by the Academy of Neurologic Physical Therapy for use with people with stroke and other neurologic conditions across the continuum of care.14 More recently, the 6MWT has been used to predict community walking activity.15
An important psychometric property of any outcome measure is its sensitivity to change and responsiveness. Liang and colleagues16,17 define sensitivity to change as the ability of an instrument to measure change regardless of whether or not that change is important; it is the amount of change that exceeds measurement error and patient variability. Responsiveness is the ability of an instrument to measure important change. In particular, the minimal detectable change ([MDC] an index of sensitivity to change) and the minimal clinically important difference ([MCID] an index of responsiveness) are useful for clinicians and researchers when interpreting scores and/or change on an outcome measure. The MDC is an estimate of the measurement error and random fluctuation in the test score in patients who are stable.18,19
Although MDC is useful for interpreting change scores, it is not ideal, as it provides only the information that the change has exceeded measurement error and variability in patients who are stable. Conversely, the MCID is more useful clinically as it provides an index of important change. The MCID involves an anchor-based approach to estimating how much change in an outcome measure is clinically important and meaningful. The anchor is some external variable that is judged to be important.20 External anchors can be patients' perception of important change, clinicians' perception of important change, or an objective marker of important change (eg, discharge home).20 For example, Fulk and colleagues21 used patient and therapist's perception of important change measured with a Global Rating of Change Scale as an anchor to estimate clinically important change in the Arm Motor Ability Test. When estimating the MCID of gait speed, Tilson and colleagues22 used a 1-point improvement on the modified Rankin Scale (mRS) as the anchor of important improvement in disability.
Unfortunately, there is limited research on the sensitivity to change and responsiveness of the 6MWT in people with stroke. In people with chronic stroke, the MDC is estimated to be 29 m,12,23 while in people with stroke undergoing inpatient rehabilitation 30 days poststroke, the MDC is estimated to be 54 m.10 To the best of our knowledge, the MCID of the 6MWT has been reported for people with stroke in only 1 other study. Using data from a completed rehabilitation trial, Perera and colleagues24 estimated meaningful change in the 6MWT using 3 different methodologies. They used an anchor-based approach using decline on 2 items of the 36-Item Short Form Health Survey (walking 1 block and climbing a flight of stairs) as the anchors. Using a distribution-based approach, they calculated standard error of measurement, and they multiplied mean baseline 6MWT distance by a small (0.2) and medium effect size (0.5). Limitations in their findings are that the anchor-based approach used was in relation to decline in performance on the anchor and so should not be applied when trying to interpret improvement. The distribution-based methods Perera and colleagues24 used to estimate change in the 6MWT were based on patients whose condition was stable and are indices of sensitivity to change not responsiveness (ie, important change). However, the MCID of the 6MWT has been reported for other patient populations and has been estimated to be between 14.0 m and 156 m in people with chronic obstructive pulmonary disease, lung disease (lung disease), coronary artery disease, fibromyalgia, and older adults.25–28
The purpose of this research study was to estimate the MCID of the 6MWT in people with stroke undergoing outpatient rehabilitation 2 months poststroke using an anchor-based approach. Based on the MDC values reported in the literature, we hypothesized that the MCID would be greater than the reported MDC values.
This study was a secondary data analysis using data collected during the Locomotor Experience Post Stroke (LEAPS) rehabilitation trial.1,2 Inclusion criteria for the LEAPS trail were diagnosis of stroke, able to walk 3 m or more with at most maximal assistance from 1 person, gait speed less than 0.80 m/s, and paresis of the lower extremity affected by the stroke. Exclusion criteria were requiring assistance with activities of daily living prior to stroke, preexisting neurological health condition, and contraindications to exercise. Participants in the LEAPS trial provided informed consent and the appropriate institutional review boards approved the LEAPS trial. The authors' institutional review board ruled this secondary data analysis exempt.
For the LEAPS trial, participants were randomized into 1 of 3 treatment arms: early locomotor training (2 months poststroke), home exercise intervention (2 months poststroke), or late locomotor training (6 months poststroke). For this study, we used data from the 2 groups (early locomotor training and home exercise intervention) that underwent the intervention at 2 months poststroke. The intervention sessions in all 3 groups were 90 minutes in length and provided 3 times a week for 12 to 16 weeks so that participants completed between 30 and 36 sessions. See the studies by Duncan et al1,2 for more details on the intervention dosage.
Outcome Measures and Anchors
At 2 months poststroke and before starting the intervention, and again at 6 months poststroke after completing the intervention, the participants performed a 6MWT using a standardized protocol.9 We used 2 different external anchors that were administered at 2 and 6 months poststroke upon completion of the intervention: the mRS29 and the Stroke Impact Scale (SIS).30,31 The mRS is a measure of disability poststroke and scores range from 0 (no symptoms) to 5 (severe disability) that are scored by a clinician. Participants who demonstrated improvement of 1 point or more on the mRS between 2 and 6 months poststroke after completion of the intervention were considered to have made important improvement in clinician-rated disability and those who did not were considered stable/not improved.
The SIS was designed from the viewpoint of, and with input from, people with stroke and their caregivers to assess health-related QOL in people with stroke.32,33 We used 1 item on the SIS that asks the person with stroke to rate his or her recovery: “On a scale of 0 to 100, with 100 representing full recovery and 0 representing no recovery, how much have you recovered from your stroke?” Participants who demonstrated an increase of 10% or more from 2 to 6 months after completion of the intervention were considered to have made important improvement in self-rated disability and those who did not were considered stable/not improved.34 These 2 anchors were chosen to provide an external criterion related to disability and to acquire 2 perspectives, 1 from the perspective of the clinician and 1 from the perspective of the participant.
Since there was no significant difference between the 2 intervention groups, we combined the data from both groups in these analyses to estimate the MCID of the 6MWT.1 Participants were dichotomized into 2 groups on the basis of their having achieved (improved) or not achieved (stable/not improved) important improvement in disability from 2 to 6 months poststroke after completion of the intervention for the 2 anchors as defined previously, an improvement of 1 or more on the mRS22 (improved) and an increase of 10% or more on the SIS recovery question (improved).34
To validate the anchors, independent t tests were performed comparing change in the 6MWT between the dichotomized groups of improved and stable/not improved and Spearman ρ correlation analyses were performed exploring the relationship between the anchors and change in 6MWT. There should be a statistically significant difference in change in 6MWT between the improved and stable/not improved groups and significant relationship between an anchor and the outcome of interest to support the validity of the anchors.
Receiver operating characteristic (ROC) curves were constructed by plotting sensitivity values (true-positive rate) on the y-axis and 1-specificity values (false-positive rate) on the x-axis for different changes in 6MWT for distinguishing important improvement from those participants without important improvement (as defined previously, improved vs stable/not improved). Separate ROC curves were constructed for the 2 different anchors (mRS clinician-rated disability and SIS participant-rated disability). The area under the curve (AUC) and 95% confidence intervals were obtained as a method for describing the ability of change in the 6MWT to distinguish participants who improved from those who did not improve on the 2 anchors. An AUC of 0.50 or less indicates that change in 6MWT has no ability beyond chance to distinguish between participants who improved and those who did not improve (as defined by the scores on the 2 anchors as defined previously), while a value of 1.0 would indicate perfect ability to distinguish between improved and not improved participants. The optimal MCIDs of the 6MWT were determined using Younden index (ie, the maximum sum of sensitivity and specificity), which is considered to be the best cutoff score for distinguishing improved from not improved participants for each of the different anchors. Sensitivity and specificity were calculated for each of the estimated MCIDs for the 2 anchors.
Because the amount of improvement on an outcome measure that is considered important may vary depending on the initial ability of the patient, in addition to estimating the MCID of the 6MWT using entire sample, we divided the sample into 2 groups on the basis of their initial walking function and estimated the MCID of the 6MWT for these 2 subgroups separately. For this subgroup analyses, we divided participants into 2 groups: participants with initial gait speed of less than 0.40 m/s and those with an initial gait speed of 0.40 m/s or more. These gait speed classifications were originally developed by Perry and colleagues35 and have been widely used to categorize functional walking ability. We used the same ROC analyses technique described previously for these 2 subgroups.
Data from 265 participants were used in these analyses. Mean age was 61.3 (12.8) years and the ratio of men to women was 57:43. For all participants, the 6MWT distance at 2 months poststroke was 125.2 (76.2) m, and mean change in 6MWT distance between 2 months and 6 months poststroke after the intervention was 78.9 (66.0) m. There was a significant difference in change in 6MWT distance between the improved and stable/not improved groups for the 2 anchors, P < 0.05. Two-month, 6-month, and change for the 6MWT data across the 2 anchor groups are presented in Table 1. There was a significant, minor relationship between mRS and change in 6MWT (r = 0.29, P < 0.05) and SIS recovery question (r = 0.29, P < 0.05).
For all participants, based on the mRS as the anchor of importance, the AUC for the ROC analysis was 0.66 (Figure 1). This resulted in an estimated MCID of the 6MWT of 71 m (Table 2). For participants whose initial gait speed was <0.40 m/s, based on the mRS as the anchor of importance, the AUC for the ROC analysis was 0.72 (Figure 1). This resulted in an estimated MCID of the 6MWT of 44 m. For participants whose initial gait speed was 0.40 m/s or more, based on the mRS as the anchor of importance, the AUC for the ROC analysis was 0.59. This resulted in an estimated MCID of the 6MWT of 71 m. For all participants, based on the SIS recovery question as the anchor of importance, the AUC for the ROC analysis was 0.59 (Figure 2). This resulted in an estimated MCID of the 6MWT of 65 m. For participants whose initial gait speed was less than 0.40 m/s, based on the SIS recovery question as the anchor of importance, the AUC for the ROC analysis was 0.62. This resulted in an estimated MCID of the 6MWT of 34 m (Table 2). For participants whose initial gait speed was 0.40 m/s or more, based on the SIS recovery question as the anchor of importance, the AUC for the ROC analysis was 0.56. This resulted in an estimated MCID of the 6MWT of 130 m.
Although the 6MWT is recommended for use in both research and clinical practice,14 we are not aware of previous studies that have estimated the threshold for clinically important change in relation to improvement for this commonly used outcome measure in people with stroke. The results of this study estimated that the MCID of the 6MWT in people undergoing physical therapy interventions between 2 and 6 months poststroke is 65 m or 71 m using anchors of clinician- or participant-reported disability. For individuals who are initially slow ambulators (gait speed <0.40 m/s), the MCID is 34 m or 44 m, and for those who are initially faster ambulators (gait speed ≥0.40 ms), the MCID is 130 m or 71 m.
Our estimated MCIDs of the 6MWT should be interpreted with caution; no threshold will ever be 100% accurate. Based on our findings, we are confident in recommending using an MCID of 44 m for people with stroke who walk slower than 0.40 m/s at the start of their physical therapy care at approximately 2 months poststroke. Our results were most accurate in this subgroup of participants. The AUC for this subgroup was 0.72 (sensitivity = 0.77, specificity = 0.60), based on the mRS as the anchor of importance. An AUC of at least 0.70 is considered adequate.36
The AUC in our analyses for all the participants and for those whose initial gait speed was ≥0.40 m/s indicates that the resulting estimated MCID values are not as accurate. The AUC for all participants was 0.59 or 0.56, depending on the anchor, are less than ideal. However, the 95% confidence intervals for the AUC for both anchors were not below 50%, indicating that these values were not found by chance. The AUC results (AUC 95% confidence interval encompassed 0.50) for the subgroup of participants whose initial gait speed was 0.40 m/s or more were even less accurate. Based on these findings, we would not recommend using the estimated MCID values with people whose gait speed is 0.40 m/s or more at the start of rehabilitation. Using the estimated MCID values with people with stroke who initially can walk ≥0.40 m/s may result in a misclassification of having experiencing important improvement in disability.
These findings add further support to the clinical usefulness of the 6MWT. The estimated MCID of 44 m is greater than the reported variability of the 6MWT in stable patients who are more than 6 months poststroke (MDC = 29 m) and slightly less than the reported variability in the 6MWT in people 30 days poststroke (MDC = 54 m).10,12,23 The estimated MCID of the 6MWT determined here is in the range of the meaningful change and standard error of measurement values reported by Perera and colleagues.24 They reported a standard error of measurement of 22 m, which is an index of sensitivity to change, so it is appropriate that our MCID finding of 44 m is larger. It is not appropriate to compare our MCID findings with their meaningful change values (21 m and 54 m) as theirs are based on a decline in function while we examined an improvement in disability; they also do not provide information on the accuracy of their estimated MCID values and so we cannot compare accuracy and sensitivity/specificity findings. These differences among the sensitivity to change and responsiveness indices of the 6MWT from the different studies are likely due to differences in time poststroke of the participants and in the analysis techniques by which the MDC and MCID values were determined.
Clinicians and researchers can use the estimated MCID to assist when making clinical decisions regarding patient change and differences between groups. For example, if a patient's initial 6MWT distance was 120 m at initial examination and at discharge it improved to 150 m (a change less than our estimated MCID), using our results the interpretation of this change would indicate that the patient did not experience a clinically important improvement in the 6MWT in relation to overall disability. However, if the patient's 6MWT at discharge was 175 m (a change greater than our estimated MCID), the interpretation of this change would indicate that the patient experienced a clinically important improvement in the 6MWT in relation to overall disability.
Researchers can use the estimated MCIDs to interpret differences between groups in intervention studies. In addition to comparing the difference between groups using statistical tests such as a t test, researchers can compare the proportion of subjects who exceed the MCID between groups using the number needed to treat.37 The number needed to treat is the number of patients who need to be treated with the experimental intervention before you can be sure that 1 patient improved, who would not have improved with the comparison intervention. The number needed to treat likely has greater clinical utility when interpreting results compared with statistical differences between groups.
These MCIDs should ideally be used only to interpret change in people with stroke in the subacute stage of recovery who initially walk slower than 0.40 m/s and whose initial 6MWT is within the range of those in this study (80 m ± 50 m). The different MCID values and accuracies based on initial walking function found in this study illustrate the importance of applying MCID values with the appropriate patient and within the correct context. The MCID values are likely to vary depending on the baseline scores and other patient characteristics.38,39 In addition, when using these MCID values to interpret change in the 6MWT, it should be done within the context of the external criterion used. In this case, overall disability as measured by the mRS. Change in the 6MWT that exceeds the MCID values would indicate a possible improvement in overall disability. The selection of these 2 anchors may have played a role in the relatively low accuracy of the estimated MCID values for the participants as a whole and in those who initially had less walking disability. The construct measured by the 6MWT (walking endurance) is not strongly related to overall disability.
These are the first reported MCIDs for improvement in the 6MWT; further research is needed with other anchors of importance, at different stages of recovery poststroke, and with different levels of 6MWT ability. A Global Rating of Change Scale that is more directly linked to the construct measured by the 6MWT (walking endurance) would be another anchor to assess in future research. However, the anchors used in the study have strong validity to support their use. The mRS is a measure of global disability with strong convergent, construct validity. It is strongly related to physical function, ability to perform ADLs, and other measures of disability.40–43 However, the fact that the mRS is primarily used in acute stroke trials and may not be as sensitive to change in the subacute stage of recovery may have influenced the estimated MCID values. The SIS is a valid measure of health-related quality of life and disability that was developed with input from people with stroke.30–33 The specific cutoff scores on these 2 anchors that were used to dichotomize participants into improved and stable/not improved groups have been used in other studies that have estimated the MCID of different outcome measures.22,34
Our results provide an initial estimate of the MCID of the 6MWT in people with stroke. Patients 2 months poststroke who initially walk slower than 0.40 m/s who exhibit an improvement in the 6MWT during outpatient rehabilitation of 44 m or greater are likely to experience clinically meaningful improvement in disability. Clinicians and researchers can use this estimated MCID of the 6MWT to interpret change and set goals. However, we would not recommend the use of the estimated MCID values for people poststroke whose initial gait speed is 0.40 m/s or more due to the low accuracy of these indices.
1. Duncan PW, Sullivan KJ, Behrman AL, et al Body-weight-supported treadmill rehabilitation after stroke. NEJM. 2011;364:2026–2036.
2. Duncan PW, Sullivan KJ, Behrman AL, et al Protocol for the Locomotor Experience Applied Post-stroke (LEAPS) trial: a randomized controlled trial. BMC Neurol. 2007;7:39.
3. Kluding PM, Dunning K, O'Dell MW, et al Foot drop stimulation versus ankle foot orthosis after stroke: 30-week outcomes. Stroke. 2013;44:1660–1669.
4. Guyatt GH, Sullivan MJ, Thompson PJ, et al The 6-minute walk: a new measure of exercise capacity in patients with chronic heart failure. Can Med Assoc J. 1985;132:919–923.
5. Lipkin DP, Scriven AJ, Crake T, Poole-Wilson PA. Six minute walking test for assessing exercise capacity in chronic heart failure. Br Med J (Clin Res Ed). 1986;292:653–655.
6. Brogardh C, Flansbjer UB, Lexell J. Self-reported walking ability in persons with chronic stroke and the relationship with gait
performance tests. PM R. 2012;4:734–738.
7. Eng JJ, Chu KS, Dawson AS, Kim CM, Hepburn KE. Functional walk tests in individuals with stroke: relation to perceived exertion and myocardial exertion. Stroke. 2002;33:756–761.
8. Kelly JO, Kilbreath SL, Davis GM, Zeman B, Raymond J. Cardiorespiratory fitness and walking ability in subacute stroke patients. Arch Phys Med Rehabil. 2003;84:1780–1785.
9. Pohl PS, Duncan PW, Perera S, et al Influence of stroke-related impairments on performance in 6-minute walk test. J Rehabil Res Dev. 2002;39:439–444.
10. Fulk GD, Echternach JL, Nof L, O'Sullivan S. Clinometric properties of the six-minute walk test in individuals undergoing rehabilitation poststroke. Physiother Theory Pract. 2008;24:195–204.
11. Fulk GD, Reynolds C, Mondal S, Deutsch JE. Predicting home and community walking activity in people with stroke. Arch Phys Med Rehabil. 2010;91:1582–1586.
12. Eng JJ, Dawson AS, Chu KS. Submaximal exercise in persons with stroke: test-retest reliability and concurrent validity with maximal oxygen consumption. Arch Phys Med Rehabil. 2004;85:113–118.
13. Flansbjer UB, Holmback AM, Downham D, Patten C, Lexell J. Reliability of gait
performance tests in men and women with hemiparesis after stroke. J Rehabil Med. 2005;37:75–82.
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;42:170–212.
15. Fulk GD, He Y, Boyne P, Dunning K. Predicting home and community walking activity poststroke. Stroke. 2017;48:406–411.
16. Liang MH. Longitudinal construct validity: establishment of clinical meaning in patient evaluative instruments. Med Care. 2000;38:II84–II90.
17. Liang MH, Lew RA, Stucki G, Fortin PR, Daltroy L. Measuring clinically important changes with patient-oriented questionnaires. Med Care. 2002;40:II45–II51.
18. Stratford PW, Riddle DL. When minimal detectable change exceeds a diagnostic test-based threshold change value for an outcome measure: resolving the conflict. Phys Ther. 2012;92:1338–1347.
19. Stratford PW. Getting more from the literature: estimating the standard error of measurement from reliability studies. Physiother Can. 2004;56:27–30.
20. Haley SM, Fragala-Pinkham MA. Interpreting change scores of tests and measures used in physical therapy. Phys Ther. 2006;86:735–743.
21. Fulk G, Martin R, Page SJ. Clinically important difference of the Arm Motor Ability Test in stroke survivors. Neurorehabil Neural Repair. 2017;31:272–279.
22. Tilson JK, Sullivan KJ, Cen SY, et al Meaningful gait
speed improvement during the first 60 days poststroke: minimal clinically important difference. Phys Ther. 2010;90:196–208.
23. Liu J, Drutz C, Kumar R, et al Use of the six-minute walk test poststroke: is there a practice effect? Archives Phys Med Rehabil. 2008;89:1686–1692.
24. Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc. 2006;54:743–749.
25. Bohannon RW, Crouch R. Minimal clinically important difference for change in 6-minute walk test distance of adults with pathology: a systematic review. J Eval Clin Pract. 2017;23:377–381.
26. du Bois RM, Weycker D, Albera C, et al Six-minute-walk test in idiopathic pulmonary fibrosis: test validation and minimal clinically important difference. Am J Respir Crit Care Med. 2011;183:1231–1237.
27. Kaleth AS, Slaven JE, Ang DC. Determining the minimal clinically important difference for 6-minute walk distance in fibromyalgia. Am J Phys Med Rehabil. 2016;95:738–745.
28. Kwok BC, Pua YH, Mamun K, Wong WP. The minimal clinically important difference of six-minute walk in Asian older adults. BMC Geriatr. 2013;13:23.
29. Banks JL, Marotta CA. Outcomes validity and reliability of the modified Rankin Scale: implications for stroke clinical trials: a literature review and synthesis. Stroke. 2007;38:1091–1096.
30. Duncan PW, Wallace D, Lai SM, Johnson D, Embretson S, Laster LJ. The Stroke Impact Scale version 2.0. Evaluation of reliability, validity, and sensitivity to change. Stroke. 1999;30:2131–2140.
31. Lai SM, Perera S, Duncan PW, Bode R. Physical and social functioning after stroke: comparison of the Stroke Impact Scale and Short Form-36. Stroke. 2003;34:488–493.
32. Lai SM, Studenski S, Duncan PW, Perera S. Persisting consequences of stroke measured by the Stroke Impact Scale. Stroke. 2002;33:1840–1844.
33. Duncan PW, Bode RK, Min Lai S, Perera S. Rasch analysis of a new stroke-specific outcome scale: the Stroke Impact Scale. Arch Phys Med Rehabil. 2003;84:950–963.
34. Lin KC, Fu T, Wu CY, et al Minimal detectable change and clinically important difference of the Stroke Impact Scale in stroke patients. Neurorehabil Neural Repair. 2010;24:486–492.
35. Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of walking handicap in the stroke population. Stroke. 1995;26:982–989.
36. Terwee CB, Bot SD, de Boer MR, et al Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol. 2007;60:34–42.
37. Dalton GW, Keating JL. Number needed to treat: a statistic relevant for physical therapists. Phys Ther. 2000;80:1214–1219.
38. Beaton DE, Boers M, Wells GA. Many faces of the minimal clinically important difference (MCID): a literature review and directions for future research. Curr Opin Rheumatol. 2002;14:109–114.
39. Guyatt GH, Osoba D, Wu AW, Wyrwich KW, Norman GR. Methods to explain the clinical significance of health status measures. Mayo Clin Proc. 2002;77:371–383.
40. Cup EH, Scholte op Reimer WJ, Thijssen MC, van Kuyk-Minis MA. Reliability and validity of the Canadian Occupational Performance Measure in stroke patients. Clin Rehabil. 2003;17:402–409.
41. Kwon S, Hartzema AG, Duncan PW, Min-Lai S. Disability measures in stroke: relationship among the Barthel Index, the Functional Independence Measure, and the Modified Rankin Scale. Stroke. 2004;35:918–923.
42. Weimar C, Kurth T, Kraywinkel K, et al Assessment of functioning and disability after ischemic stroke. Stroke. 2002;33:2053–2059.
43. Tilley BC, Marler J, Geller NL, et al Use of a global test for multiple outcomes in stroke trials with application to the National Institute of Neurological Disorders and Stroke t-PA Stroke Trial. Stroke. 1996;27:2136–2142.