Statistically and clinically important change of pain scores in patients with myogenous temporomandibular disorders

A within-patient change in pain score after treatment is statistically 'reliable' when it exceeds the smallest detectable difference (SDD). The aims of the present study were (i) to determine SDD for scoring pain behavior on a 0-5 point adjectival scale, and (ii) to explore the relationship between SDD, clinically important difference (CID) and effect size (ES) following treatment of known efficacy, and to compare these parameters of pain behavior with those of VAS-scores of pain intensity [van Grootel RJ, van der Bilt A, van der Glas HW. Long-term reliable change of pain scores in individual myogenous TMD patients. Eur J Pain 2007;11:635-43]. SDD was determined using duplicate scores on pain behavior from a pre-treatment diary that was completed by 118 patients with myogenous temporomandibular disorders (TMD). CID was determined as the mean change in score following treatment, and Cohen's ES as the ratio between mean change and SD of baseline values. The SDDs were 2-3 units (40-60% of the scale range) for test-retest intervals of 1-13 days. CID was 1.13 units (22.6%) and ES was 1.38. The normalized SDD and CID values and ES were similar for VAS-scores of pain intensity, i.e., 38-49% (SDD), 24.2% (CID) and 1.09 (ES). Because reliable change (change > SDD) exceeds CID, the responsiveness of scoring of pain variables is low for detecting CID. The finding of ES values that are larger than 0.5 (ES for patients with chronic degenerative diseases [Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life. The remarkable universality of half a standard deviation. Med Care 2003;41:582-92]) suggests that for myogenous TMD (chronic pain not caused by somatic disease and with a large chance on recovery following treatment), there are higher expectations of what constitutes important change.