Objective To determine the reliability of patient reports of pain intensity during the application of physical tests.
Methods A single examiner required participants to use the numeric pain rating scale (NPRS) to report the intensity of pain provoked during physical testing of the hip. Standardized versions of 14 physical tests were used on 18 people with hip pain. Tests were repeated at 1 hour and 2-7 days later. Within- and between-session reliability of reports of pain intensity was calculated using intraclass correlation coefficients (ICCs) and Lin's concordance correlation coefficients (CCCs). Standard errors of measurement (SEMs) were calculated.
Results ICC/CCC values for within-session reports of pain intensity ranged from 0.34 (95% confidence interval [95% CI] -0.08, 0.66) to 0.88 (95% CI 0.73, 0.95). Patients demonstrated "substantial" or "almost perfect" reliability in reporting pain intensity with 9 of 14 tests and "moderate" reliability with 3 tests. Two tests were unreliable (ICC 0.35 and 0.34). ICC/CCC values for between-session reports of pain intensity ranged from -0.05 (95% CI -0.42, 0.34) to 0.84 (95% CI 0.44, 0.95). Patients demonstrated substantial or almost perfect reliability in reporting pain intensity with 11 of 14 tests. Two tests were unreliable (ICC 0.26 and -0.05). The average SEM value of 0.9 points on the NPRS was the same for both within- and between-session testing (range 0.6-1.6).
Conclusion Patient reports of the intensity of pain provoked by physical tests are sufficiently reliable to be clinically useful. However, the SEM of 0.9 points should be considered when precise calculations of changes in pain intensity are important.