The brief pain inventory-Interference Subscale has acceptable reliability but questionable validity in acute back and neck pain populations

Abstract

BACKGROUND: The Brief Pain Inventory-Interference Subscale (BPI-IS) is a subscale of the BPI assessment tool developed to rapidly assess the impact of a person's pain on their function. It is uncertain whether it has one or two factors, and whether it has acceptable clinimetric properties in a mixed spinal pain (back and/or neck) population. OBJECTIVES: To determine the clinimetric properties of the BPI-IS in a population with mixed spinal pain. METHODS: We completed a clinimetric evaluation with a test-retest design, factor analysis, and hypothesis testing. We used data collected for a randomised clinical trial including a population presenting to primary care or emergency departments with acute spinal pain (back and/or neck). RESULTS: Confirmatory factor analysis better supported the two-factor model of the BPI-IS (physical interference factor and affective interference factor) as compared to the one-factor model. Both one and two-factor models had acceptable reliability (high internal consistency and no evidence of floor or ceiling effects). Both models failed to reach our a-priori thresholds for acceptable construct (cross sectional) validity, and responsiveness (longitudinal validity) in either back or neck pain populations. CONCLUSION: The BPI-IS has two factors and both have acceptable reliability, but tests for validity did not reach our a priori thresholds for acceptability (construct validity and responsiveness). The BPI-IS may not be suitable to measure the impact of pain on function in back and neck pain populations.