• News New Blog Banner

  • New in the Literature: Clinical Balance Measures (Gait Posture. 2011;33:737-739)

    A change in single-leg-stance-time (SLST) performance in older adults should exceed 24.1 seconds in order to be considered real change, say authors of an article published in the April issue of Gait & Posture. SLST exhibits poor absolute reliability, the authors add, and appears unlikely to be sensitive to detecting change in performance in geriatric clinical settings and research studies.

    A measure of absolute reliability, the standard error of measurement (SEM), can be used to compute minimal detectable change (MDC), a clinically useful indicator of change in performance exceeding that attributable to measurement error. The purpose of this study was to quantify MDC for SLST in community-dwelling older adults.

    Twenty-five adults (60-89 years) performed repeated trials of SLST. Relative and absolute reliability for SLST were quantified using the intraclass correlation coefficient and SEM. The MDC was computed from the SEM.

    SEM was 8.7s and MDC at the 95% confidence level was 24.1 seconds for SLST. SLST exhibited large measurement error (40.8%) and high minimum change (113.1%) percent values.


    • I have a hard time accepting this conclusion. Does this mean to say that an older adult who goes from not being able to perform a single leg stance for 3 seconds initially but then improves to holding it for greater than 10 seconds has not made a real change?!? In my experience, patients usually report feeling significantly better balanced when they make such progress. I did not see an indication where the study focused on older adults who are at risk for falling, and I would suspect that a study of this sub-population may reveal different results.

      Posted by Paul Weiss on 6/3/2011 2:29 PM

    • So if the patient's SLST is 3 seconds and improves to 21 seconds, that's not a minimal detectable change? Please share.

      Posted by Gail Elliott on 6/3/2011 4:15 PM

    • No. According to the study the Minimal Detectable Change (MDC) is 24.1 seconds (at 95% confidence). The MDC was calculated, not measured, and represents an estimate of the true change score. The measurement used to derive the MDC had a high variation in the study. The Standard Error of the Measure (SEM) was 8.7sec. which means that most (~68%) of the SLS values were either 8.7sec. higher or lower than the average (mean). I think you can relate this to your clinical experience. Single Leg Stance times vary a lot even within a single session. Even some young people can't perform this test! They get a score of 0 sec. Despite the high variance and the apparent contradiction presented by the poor absolute reliability you may judiciously use this measure for clinical decision making given the following caveats: 1) There is no risk of harm to the patient 2) No alternative measure of superior quality exists The evidence base is fluid and incomplete. That is why humans are still necessary to make decisions in physical therapy. Tim Richardson, PT

      Posted by Charles Richardson -> =GR^EM on 6/3/2011 6:39 PM

    • I agree that these findings do not seem correct. An 8.7s SEM does not seem realistic here. That would impact all downstream calculations. I also agree with the above posting that what is true for a population of healthy / fit, community dwelling adults is far different from the figures for the home bound at-risk-of-falls population!

      Posted by Jeff Mauk PT on 6/3/2011 6:51 PM

    • I am sure that the lower the patient's initial SLST was initially, the less amount of time would indicate a significant change. A study should perhaps be done to measure significant change as a percentage. (Example: a 40% significance factor for someone who could do the SLST for 5 seconds would be 2 extra seconds ;However, 40% for someone who can do 100 seconds should be 140 seconds.) The fact that such a high number is given might be a tad misleading especially when many patients cannot perform the SLST for 5 seconds.

      Posted by Konrad Siemek on 6/3/2011 8:18 PM

    • I am a practicing P.T. for nearly 50 yrs. I have always included the SLST whenever I eval any patient considered a senior or who is referred with gait/posture/balance deficits. Anecdotally, I believe any patient who cannot SLS for at least 5 seconds is at severe risk of falling. Common sense + functional/measurable gain of 20% at first re-eval shows improvement to me that is significant and shows potential for more improvement in SLST.

      Posted by Herschel Budlow on 6/3/2011 9:41 PM

    • I have trouble believing this also. What was the testing protocol? It is true the variability of performance in the same person may be quite large at first (especially if they have not done that kind of thing in years), so you should give them a few practice trials, then get an official measure. It is usually a reasonable marker for progress if you test the same way and compare their best to their best. Was it physical therapists administering the test? You should see a more meaningful measure with a smaller standard of error if your testing protocol is reasonable.

      Posted by Adriene on 6/3/2011 11:54 PM

    • I am going to make a trip to the local university to read this article. I wish that the study had a larger sample size. The larger the sample size, the smaller the change will be needed to measure significance. Also, it would have been nice if the study performed an analysis of variance between different age groups separated by decades of life, since "normally functioning" 80 to 89 year-olds will present differently than 60-69 year-olds. Also, I wonder if they used functional outcome scales to compare balance measurements.

      Posted by Rex Fujiwara on 6/4/2011 2:55 AM

    • It interesting that many of the previous posts, start out with a dismissal of the results based on clinical observations. While clinical expertise and experience is certainly included with in the evidence hierachy it is a lower level of evidence. Perhaps an example of how resistant experienced clinicians can be to a change in practice when newer evidence challenges our practice patterns, assessments, or intervention selection. The reading of this study challanged my beliefs regarding the utility of single leg stance as assessment tool. While I agree that one should be cautious in applying these results to all populations, this study does seem to questions the utility of single leg stance as assessment tool. It is probably safe to say that single leg stance is common assessment that used by many in daily practice as if it is a gold standard of balance assessment. Although single leg stance MAY provide insight into the nature of the balance impairment Perhaps it is time to question this assessment as standard element of the evaluation or at least its purpose? In this study, the study population consisted of generally healthy community dwelling individuals with roughly a mean single leg stance time of around 20 seconds (with faily large variation). So, how much single leg stance time is really required to function in daily life? Some degree of brief single leg time probably required for walking, stair climbing, standing and putting on pants -- So why not directly assess these items, rather then use what appears to be a fairly unreliable assessment to infer improvement in functionality and balance improvement? MDC value aside, what is perhaps most significant in this study is the poor reliability demonstated of the single leg stance assessment. Given that a number of balance and functional assessments exists of higher quality it would seem that from a clinical standpoint other assessments would be better choice when assessing balance, establishing fall risk, or demonstrating improvement in physical performance and function. Any thoughts? Brad Tracy, PT

      Posted by Brad Tracy, PT, DPT on 6/4/2011 12:29 PM

    • My home care experience, (30 yrs.) and published clinical research, (The Spring Scale Test SST: A Valid and Reliable.., JGPT Dec. 2009) supports the findings of Dr Golberg and the comments of Brad Tracy PT. In our study, four balance related performance measures were examined; (TUGT, Gait Speed, SLST and Tandem Stand Time (TST) on a population of 58 active independent community dwelling adults aged 65 to 90+, 29 non fallers, and 29 with at least 1 fall in prior 2 years. The SLST sensitivity was 48.3 % , specificity 89.7% at ROC determined cut point of 6.5 seconds. The ability to perform the SLST is indeed highly variable in this age group thus limiting the usefullness of this test. Many individuals 80+ years old without balance impairment and without a prior fall cannot perform the SLST. Related to SLST, Tai Chi has come under recent scrutiny as well (Logghe IH, Arch Gerontol Geriatr 2010, and Prev Med 2010. Low S, Arch Gerontol Geriatr, 2009). Both static and unchallenging dynamic tests are of limited use in determining reactive stepping responses which are essential for perturbation responses ans preventing a fall. I agree with Brad's conclusions that physical therapists must explore alternative fall risk tests. Louis DePasquale PT,MA

      Posted by Louis DePasquale PT on 6/5/2011 4:42 PM

    • I agree with Louis and Brad....using the SLST should be a PART of a more dynamic assessment. SLS is part of the Berg and Romberg progression but should never be used as a sole test. More dynamic testing gives a fuller picture of what the patient can actually do. Also, many of the perceived practices of therapists need to be challenged in order to challenge the aging adult much more than is typically done.

      Posted by Jill Heitzman, PT, DPT, GCS, CWS, CEEAA on 6/6/2011 12:20 PM

    • I agree with Brad and Louis that the single limb stance test does not seem to be the best tool for measuring change in our older patients, especially considering its poor reliability and lack of direct functional relevance compared to other more thorough measures. However, I think that more research needs to be done before we dismiss the measure completely. I want to present 2 caveats that people should know about when interpreting articles on minimal detectable change (MDC) scores that were calculated based on standard error of measurements and reliability coefficients. First, MDCs are not the only way to determine change scores. The calculation of minimal clinical important difference (MCID, MID) based on patient perceived improvement is likely to be a better indicator of clinically relevant change (Turner et al, 2010. Journal of Clinical Epidemiology). The second caveat is that baseline functional levels do impact MCID change values, so measures that are good for one patient, may not be appropriate for other patients with different diagnoses or different levels of function (Wang et al. 2011 PTJ). Many of the comments above refer to much lower level patients than the subjects in this recent study. I think it would be interesting to compare MDC and MCID levels for single limb stance for very impaired older populations versus healthy, independent, community dwelling elders. Miriam Rafferty, PT, DPT, NCS

      Posted by Miriam Rafferty -> ?MW[BI on 6/6/2011 12:56 PM

    • I, too, have difficulty using SLS as a major measure of balance. It is only one of many. It also only looks at static balance. The touching of a 2 to 4 inch step alternating feet gives me a better sense of a person's real balance as well as requiring that he/she stand SLS for at least 3 seconds during that motion. Remember that 3 seconds is a minimum for a person to safely step, for example, over a door sill, such as going in or out of a sliding patio door. But the dynamic of that motion versus static standing SLS is better tested with Berg task #12. For my home dwelling largely older population, 10 seconds is a huge success...20 seconds would probably never be achieved during current allowable therapy times.

      Posted by Sami Jo Magoffin, MPH, PT on 6/7/2011 10:38 AM

    Leave a comment
    Name *
    Email *