Interrater reliability of the International Association for the Study of Pain and Tunks' spinal cord injury pain classification schemes.

Abstract

To determine the interrater reliability of the International Association for the Study of Pain and Tunks' spinal cord injury pain classification schemes. A total of 64 pain sites reported by 29 individuals were classified using International Association for the Study of Pain and Tunks' classification schemes. Three raters independently categorized each pain site. In general, disagreement in pain classification between the three raters was found for about 50-70% of the pain sites. Disagreement between rater pairs (two raters at a time) was somewhat better, ranging from about 20% to 50%. The kappa statistic for interrater agreement was in the marginally acceptable range (i.e., 0.3 to 0.65). Although disagreement tended to be somewhat higher using the Tunks scheme, both classification schemes showed low interrater agreement. Consistent with our previous research using the Donovan spinal cord injury pain classification scheme, considerable variability between raters was demonstrated using the International Association for the Study of Pain and Tunks' spinal cord injury pain classification schemes.