Comparison of pain behaviors in multiple sclerosis, back pain, and arthritis.

TitleComparison of pain behaviors in multiple sclerosis, back pain, and arthritis.
Publication TypeJournal Article
2012
AuthorsCook KF, Bamer AM, Amtmann D, Jensen MP, Johnson KL, Callahan L, Kim J, Keefe FJ, Revicki D, Roddey TS
JournalQuality of life research
Volume20
IssueSuppl 1
Pagination6
22298117
Humans, Quality of Life, Research

Aims: To compare pain behaviors in three samples: multiple sclerosis (MS), back pain, and arthritis using pain behavior frequency counts (video-taped), and self- and significant other (SO)-responses to candidate items for a new pain behavior measure. Methods: A sample of patient/SO pairs (N=620 pairs) completed measures of pain, function, disability, and other pain correlates. In addition, a sample of 30 individuals with back pain, 26 with arthritis, and 30 with MS were videotaped for 10 minutes while sitting, standing, walking, and lying down. Videotapes were coded to obtain pain behavior frequency counts by category (guarding, sighing, bracing, rubbing, and grimacing) and total behavior counts. Results: Mean item responses (1 to 5 response scale) in MS, arthritis, and back pain were, respectively, 2.7 (SD=0.56), 2.7 (SD=0.71), and 3.0 (SD=0.69) Spearman correlation coefficients between patient and SO pain responses were 0.55 (MS), 0.60 (arthritis), and 0.67 (back pain). Mean item score differences between self and SO item scores (1-5 response scale) were highest for persons with arthritis (0.08 higher) and lowest for persons with MS (0.02 higher). Self-reported pain behaviors and pain behavior frequency counts (videotapes) were moderately correlated and varied by item. Items with highest correlations were items about using a cane or other support (0.62), asking for help when walking (0.53), and the item, “You could hear it in my voice” (0.50). Classes of behaviors most correlated with self-report pain behaviors varied by diagnosis. In the back pain sample, guarding behavior counts had the strongest correlation with self-report (0.50). In the arthritis sample, the highest correlations were between self-reported pain behaviors and guarding (0.47) and total behavior counts (0.53). In the sample with MS, the highest values were for counts of rubbing (0.49) and total behavior counts (0.64). Conclusions: Pain behaviors vary somewhat by diagnosis but there also are substantial similarities. Significant others reported higher levels of pain behaviors than were self-reported, but mean differences were less than 1 response category on a 1-5 response scale. The correlations among self-report, SO-report, and frequency counts based on videotaped observations support the validity of candidate items for a new pain behavior measure.

http://www.ncbi.nlm.nih.gov/pubmed/22298117