Interpreting oral health-related quality of life data.

The most common way of presenting data from studies using quality of life or patient-based outcome (PBO) measures is in terms of mean scores along with testing the statistical significance of differences in means. We argue that this is insufficient in and of itself and call for a more comprehensive and thoughtful approach to the reporting and interpretation of data. PBO scores (and their means for that matter) are intrinsically meaningless, and differences in means between groups mask important and potentially different patterns in response within groups. More importantly, they are difficult to interpret because of the absence of a meaningful benchmark. The minimally important difference (MID) provides that benchmark to assist interpretability. This commentary discusses different approaches (distribution-based and anchor-based) and specific methods for assessing the MID in both longitudinal and cross-sectional studies, and suggests minimum standards for reporting and interpreting PBO measures in an oral health context.