• D Yarnitsky, E Sprecher, R Zaslansky, and J A Hemli.
• Department of Neurology and Clinical Neurophysiology, Rambam Medical Center, Haifa, Israel.
• Pain. 1996 Oct 1; 67 (2-3): 327-33.

AbstractExperimental heat pain transients were administered to 30 normal volunteers over four weekly sessions, measuring both heat pain (HP) threshold and suprathreshold magnitude estimation through VAS. Repeatability and bias for these two factors were evaluated. Heat pain thresholds measured through the method of limits were previously shown to have inter-session bias, presumably due to a practice effect. Existence of such a bias between first and second measurement sessions casts doubt on the usefulness of this parameter for pain assessment of individuals over time. In the present study, measurements of normal HP thresholds over four sessions showed that bias exists between the first and successive sessions, but not among sessions other than the first. It is concluded that (i) HP thresholds obtained from a single session are of limited value, and should be carefully interpreted. (ii) Long-term studies that use the HP threshold should take results from the second (or later) session as their baseline. The Visual Analog Scale (VAS) is considered the 'gold standard' for assessment of clinical and suprathreshold experimental pain, and changes in VAS score are regarded as significant evidence of individual response to treatment, placebo, or experimental manipulation. Although its overall group accuracy and precision have been examined for both clinical and experimental pain, and found adequate (Price 1988), the VAS has not been rigorously assessed for repeatability. Stimuli at three pain levels, 1.5, 3 and 4.5 degrees C above each individual's heat pain threshold as determined at each session, were given. Several models of analysis of the VAS were tested and repeatabilities (r) obtained from these analyses demonstrate poor precision for each of the tested analysis models. For example, inter-session repeatabilities for the three individual pain levels ranged from r = 3.8-4.7, effectively providing a confidence interval of 7.6-9.4 for any VAS reading on a 0- to 10-point scale. An examination of intra-session VAS provided somewhat better results. Thus, use of the VAS in similar experimental settings is called into question. The use of the VAS in clinical settings, where individual assessments are necessary, is also called into question, but remains to be specifically tested.

11 keywords...

hide…