OBJECTIVES: To determine the predictive utility of verbal descriptors to distinguish between pain types following spinal cord injury (SCI). DESIGN: Cross-sectional. SETTING: USA. METHODS: Participants (n=29) completed the Short Form - McGill Pain Questionnaire (SF-MPQ) for each pain site reported. A total of 64 pain sites were reported with 80% of the sample reporting multiple pain sites. Each pain site was categorized using three different SCI pain classification schemes. The predictive utility of verbal descriptors to distinguish between pain types was examined statistically using (1) each word separately, (2) a combination of words (ie, the SF-MPQ total subscales, the number of words chosen on each scale), and (3) discriminant function analysis. RESULTS: There was a substantial overlap in the use of verbal descriptors across pain types. Few differences across pain types were found for endorsement of individual words, and differences across pain types were not found for any of the word combination scores. The majority of the verbal descriptors did not enter the step-wise discriminant functions for each SCI pain classification scheme, however, 'tingling' and 'aching' showed modest predictive utility for neuropathic and musculoskeletal pain, respectively. Correct classification was in the low range (ie, 39% to 82%, average=60%, with a 33% chance level). All three pain classification schemes showed the same general pattern of results. CONCLUSION: In general, verbal descriptors alone offered marginal utility with regard to identifying specific pain types following SCI. Future directions alone and implications are discussed.
OBJECTIVES: To determine the predictive utility of verbal descriptors to distinguish between pain types following spinal cord injury (SCI). DESIGN: Cross-sectional. SETTING: USA. METHODS:Participants (n=29) completed the Short Form - McGill Pain Questionnaire (SF-MPQ) for each pain site reported. A total of 64 pain sites were reported with 80% of the sample reporting multiple pain sites. Each pain site was categorized using three different SCI pain classification schemes. The predictive utility of verbal descriptors to distinguish between pain types was examined statistically using (1) each word separately, (2) a combination of words (ie, the SF-MPQ total subscales, the number of words chosen on each scale), and (3) discriminant function analysis. RESULTS: There was a substantial overlap in the use of verbal descriptors across pain types. Few differences across pain types were found for endorsement of individual words, and differences across pain types were not found for any of the word combination scores. The majority of the verbal descriptors did not enter the step-wise discriminant functions for each SCI pain classification scheme, however, 'tingling' and 'aching' showed modest predictive utility for neuropathic and musculoskeletal pain, respectively. Correct classification was in the low range (ie, 39% to 82%, average=60%, with a 33% chance level). All three pain classification schemes showed the same general pattern of results. CONCLUSION: In general, verbal descriptors alone offered marginal utility with regard to identifying specific pain types following SCI. Future directions alone and implications are discussed.
Authors: Thomas N Bryce; Cecilia Norrbrink Budh; Diana D Cardenas; Marcel Dijkers; Elizabeth R Felix; Nanna B Finnerup; Paul Kennedy; Thomas Lundeberg; J Scott Richards; Diana H Rintala; Philip Siddall; Eva Widerstrom-Noga Journal: J Spinal Cord Med Date: 2007 Impact factor: 1.985