Reliability and validity of verbal descriptor scales of painfulness

Abstract

Previous studies have provided information about the reliability and validity of verbal descriptor scales of sensory intensity and unpleasantness and have shown that these two dimensions can be differentially affected by pharmacological manipulations. Since the relation between these dimensions and the general term ‘pain’ is not known, two experiments developed a verbal descriptor scale of painfulness and compared the sensitivity of this scale to pharmacological manipulations used previously with scales of sensory intensity and unpleasantness. In exp. I, 20 subjects used cross-modality matching to both handgrip force and tone duration to quantify the amount of pain implied by verbal descriptor phrases such as ‘slightly painful’, ‘somewhat painful’ and ‘very painful’. Ratio scales of relative magnitude for each individual were highly correlated within subjects (mean r = 0.92) and between a scale from each individual and a combined scale from others in the group (mean r = 0.93). These correlations indicate agreement between individual scales; an individual's scale values were predicted equally well by that individual or by a group of similar persons. In exp. II, 4 groups of 10 subjects rated the magnitude of painful tooth pulp sensations by choosing pain descriptors from randomized lists. Seven electrical stimuli spaced between individually determined pain threshold and tolerance values were delivered in random sequence 6 times before and after double-blind intravenous infusions of placebo, 0.11 mg/kg diazepam, 0.66 μg/kg fentanyl or a combination of the diazepam and fentanyl doses. Mean responses were reduced significantly after all active drugs but not after placebo. These results suggest that the term pain does not represent a simple combination of sensory intensity and/or unpleasantness and shows that the sensitivity to an inert placebo, an active placebo, and an analgesic can vary with the type of pain assessment procedure.