Challenges in interpreting individual-level changes in health-related quality of life in patients with glioma using minimally important differences (MIDs) and a 4-point Likert scale.

Abstract

Purpose: Interpretation of changes on the individual level is often based on minimally important differences (MIDs) developed on the group level. We investigated the impact of applying different group-level MIDs (anchor-based and 10-point MIDs) to determine health-related quality of life (HRQoL) changes in glioma patients. We further explored directions and magnitudes of these changes and their relationship to response formats and types of scale.

Methods: We included 92 glioma patients at least 18 years old from a previously conducted randomized prospective study. We calculated changes in HRQoL (EORTC QLQ-C30 and QLQ-BN20) at individual levels over a two-week period and used anchor-based and 10-point MIDs to estimate if change is clinically meaningful; thereafter, we explored the direction and magnitude of changes.

Results: Between 8.8% and 66.3% of the patients had actual changes in estimated scales. While 16.3%-60.9% and 8.8%-59.8% of the patients changed to a clinically relevant extent using anchor-based and 10-point MIDs in any scale, respectively. Changes were mostly in the functional than symptom scales and mostly minor, i.e., changes between 'not at all' and 'a little' or 'a little' and 'quite a bit.'

Conclusion: 10-point compared to anchor-based MIDs underestimates clinically relevant changes. Therefore, the application of different MIDs to the same research question can lead to diverse result interpretations. As most changes were minor, it could be argued if these reflect actual relevant changes for an individual or that the current response scale lacks sufficient differentiating ability, warranting further research over the best method to evaluate individual-level changes.