American Journal of Respiratory and Critical Care Medicine

From the Authors:

Chandra and colleagues have broadened our commentary on repeatability during clinical testing and interpretation of an individual's data. Our comments were based mainly on the analysis of real data (1) plus computational algorithms and simulation, founded on this real data, which is a recognized approach in many areas of healthcare. All patients had completed a practice test which was excluded from the analysis. Subsequently, there was no evidence of a “learning effect.” The Bland-Altman plots demonstrated no systematic error and met the statistical requirements for reporting repeatability (i.e., homoscedasticity). Given the random biological variability seen in all effort-dependent tests, assigning terms such as “learning” or “fatigue” is difficult when test repeatability is large.

We commented on the 6-minute-walk test (6MWT) as it illustrates the limitations of using the minimal clinically important difference to interpret changes in an individual patient. Although two 6MWTs are usually performed in clinical practice, other assessments such as FEV1 or peripheral muscle force usually include at least three attempts. The suggestion that a truer test would exhibit less random variation confuses individual random variability with responsiveness and precision. In the original study (1) the variability did not decrease with repeated 6MWTs. As with group data, the precision of estimating an individual's true value and variability increases with the number of valid tests performed.

We agree that carry-over effects, often referred to as a learning effect, on the 6MWT should be considered when interpreting the results. We also recognize the importance of controlling conditions to minimize the effect of fatigue (adequate rest) and learning (adequate practice). Nevertheless, the instability of this test, even after a practice walk, challenges its use as a single measure. In randomized controlled trials a systematic error can be accounted for. In an individual, this is more difficult and one may need to rely on more complicated strategies of assessment such as N of 1 trials (2).

The comments on regression to the mean, a consequence of using the maximum measure, are not relevant when the mean of repeated attempts is used. We presented additional arithmetic consequences that occur in real life for measures that follow the central limit theorem. Using the maximum can contribute to increasing systematic error and variability. The ATS has provided clearly defined criteria and limits for an acceptable test of spirometry. Defining limits for the 6MWT would be difficult given its relatively high variability. Even with well-defined tests such as spirometry, a recent communication (3) brought to our attention legitimate arguments for considering statistical approaches, requiring using the mean and variation of repeated measures, when assessing an individual's response to treatment.

The comments made by Chandra and colleagues encourage us to suggest that although highly variable measures such as the 6MWT, with poor repeatability and unquantifiable cofounding variables, provide useful information about groups of patients, these measures should be used in a more judicious way to assess outcomes in individual patients.

1. Goldstein RS, Gort EH, Stubbing D, Avendano MA, Guyatt GH. Randomised controlled trial of respiratory rehabilitation. Lancet 1994;344:13941397.
2. Guyatt G, Sackett D, Taylor DW, Chong J, Roberts R, Pugsley S. Determining optimal therapy–randomized trials in individual patients. N Engl J Med 1986;314:889892.
3. Hansen JE. Has my patient responded? Am J Respir Crit Care Med (In press)

Author Contributions: T.E.D. contributed to conception and drafted the letter; K.H., R.A.E., and R.S.G. contributed to revising the letter critically for important intellectual content.

Supported by the National Sanitarium Association-University of Toronto Chair in Respiratory Rehabilitation Research.

Author disclosures


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record