Evaluating the Statistical Significance of Health-Related Quality-Of-Life Change in Individual Patients

Abstract

Assessing individual change is feasible and potentially useful in clinical practice. This article provides an overview of the evaluation of statistically significant change in health-related quality of life (HRQOL) for individual patients. We review the standard error of measurement, standard error of prediction, and reliable change indices using a sample of 54 patients receiving care at the UCLA Center for East-West Medicine. The largest amount of change necessary for statistical significance was found for the reliable change index and the smallest change was needed for the standard error of measurement. The amount of change required for statistical significance was intermediate for the standard error of prediction. The median kappa for classifying change (declined, stayed the same, improved) by different indices was .82, indicating a high level of agreement. Future research is needed to determine if one index is most appropriate for evaluating the significance of individual change.

The minimal clinically important difference in the Gastrointestinal Quality-of-Life Index after cholecystectomy

“…The RCI indicates whether change in an individual patient is beyond the measurement error of an instrument. The advantage of RCI is that it tests directly for significant individual change [12]. As expected, clinically significant changes in scores for the symptoms, emotions, physical functions, and social functions subscales were larger for the individual patients (9.23-10.67) than for the overall group (6.42-7.64 for the ''somewhat better'' group), and the RCI decreased as subscale test-retest reliability increased.…”

“…The reliable change index (RCI) was used to evaluate the magnitude of change in the scores that was statistically reliable and not due to random measurement error. The RCI is a function of the standard deviation and the reliability coefficient of an instrument [12]. In this study, an RCI exceeding 1.96 indicated that the postsurgery score was unaffected by random measurement error and that the change was reliable (p \ 0.05).…”

Lumbar Computerized Adaptive Test and Modified Oswestry Low Back Pain Disability Questionnaire: Relative Validity and Important Change

“…Anchor-based methods use a reference standard that is clinically relevant as an anchor for determining change that is important to the patient, such as the patient global rating of change (GROC). As suggested by Hays et al 43 and de Vet et al, 15 anchor-based methods estimate whether group change is big enough to be regarded as clinically important, and the concept of "minimal importance" is explicitly defined and incorporated in these anchor-based methods. Equality of the areas under the curves (AUCs) for any pair of ROC analyses was assessed using an algorithm suggested by DeLong et al 11 for dependent data.…”

Evaluation of a Multi-disciplinary Back Pain Rehabilitation Programme—Individual and Group Perspectives

“…A number of methods are available to assess statistically significant individual change [30] and clinical significance/meaningful change [31]. Ferguson et al [19] describe the use of Reliable Change Index with SF-36 data and the approach they used has been adopted within this study, but using relevant UK norm data [32,33].…”

“…Hays et al [30] describe the use of the standard error of measurement and the standard error of prediction alongside RCI. The paper emphasises the point that whichever measure is used, changes required to be significant at an individual level are much greater than those required at a group level.…”