Reproducibility of pain measurement and pain perception

The reproducibility of both the conscious experience of pain and the reproducibility of psychophysical assessments of pain remain critical, yet poorly characterized factors in pain research and treatment. To assess the reproducibility of both the pain experience and two methods of pain assessment, 15 subjects evaluated experimental heat pain during four weekly sessions. In each session, both brief (5s) and prolonged (90s) heat stimuli were utilized to determine effects of stimulus duration on reproducibility. Multiple presentations of the brief heat stimuli in each session were used to evaluate effects of response averaging. Both visual analog scales (VAS) and randomized verbal descriptor scales (VDS) were employed to better distinguish variations in the pain experience from variations in pain scale usage. Subjects also rated the intensity of visual stimuli in order to provide an independent assessment of the session-to-session variation in the use of both types of scales. Within-subjects analyses revealed that ratings of visual stimuli exhibited significantly less session-to-session variation than ratings of heat pain. Thus, pain perceptions were more variable than perceptions of visual stimuli after controlling for session-to-session variations in scale usage. Comparisons between scales indicated that intensity ratings acquired with the VAS had significantly smaller session-to-session variation than those acquired with the VDS, although VDS ratings were spread across a larger range of the scale. For both scales, analyses of the effects of stimulus averaging and stimulus duration revealed that averaging multiple assessments of the same stimulus substantially reduces session-to-session variation and that multiple assessments of brief stimuli produce responses which are more reproducible than a single presentation of a prolonged stimulus. However, the VAS was significantly more sensitive to small differences in perceived pain intensity and pain unpleasantness, and did not exhibit some of the order effects present with the VDS. Taken together, these results indicate that the reproducibility of psychophysical ratings of pain can be maximized: (1) by averaging responses to multiple, brief stimuli; (2) by providing subjects with a training period distinct from the study period; and (3) by ensuring that interpretation of scale parameters remains constant over time. Thus, although the experiences of both experimental and clinical pain are highly variable, pain assessment procedures can be structured to minimize session-to-session variability.