• Eva Skovlund, Michael Bretthauer, Tom Grotmol, Inger Kristin Larsen, and Geir Hoff.
• Section of Medical Statistics, University of Oslo, Oslo, Norway. eva.skovlund@basalmed.uio.no
• Clin J Pain. 2005 Jul 1; 21 (4): 292-6.

ObjectiveThe purpose of this study was to compare the sensitivity of two commonly used pain-rating scales, the Visual Analog Scale and the 4-point verbal rating scale. Both are considered reliable and valid, but previous studies regarding sensitivity of rating scales have lead to different conclusions, and there is no firm agreement as to the best scale to choose.MethodsThe sensitivity of the Visual Analog Scale and the 4-point verbal rating scale was compared by stochastic simulation. In the simulation model, we used 168 pairs of pain ratings on the Visual Analog Scale and the 4-point verbal rating scale from individuals undergoing a lower gastrointestinal endoscopy, maintaining the true relation between ratings from the same individual. We created empirical distributions mimicking 2 independent groups of pain ratings. Random samples from the 2 groups were compared by the Wilcoxon-Mann-Whitney U test in 10,000 repetitions of a computer algorithm. By increasing the proportion of individuals with a high level of pain in one group, we increased the true difference between pain ratings and estimated a statistical power function.ResultsIn the present pain model with pain ratings from healthy individuals undergoing endoscopy, the Visual Analog Scale is consistently more sensitive than the four-point verbal rating scale.DiscussionBecause each individual provided one Visual Analog Scale and one 4-point verbal rating scale rating for the same pain experience, the ability of the two scales to detect differences between groups of pain ratings could be compared. The use of a simulation model enabled estimation of a power function and reduced the probability of basing the conclusion on a chance finding.

26 keywords...

hide…