Quality of Life Research

, Volume 22, Issue 10, pp 2769–2776

Development of a crosswalk for pain interference measured by the BPI and PROMIS pain interference short form


  • Robert L. Askew
    • Department of Rehabilitation MedicineUniversity of Washington
  • Jiseon Kim
    • Department of Rehabilitation MedicineUniversity of Washington
  • Hyewon Chung
    • Department of EducationChungnam National University
  • Karon F. Cook
    • Department of Medical Social SciencesNorthwestern University Feinberg School of Medicine
  • Kurt L. Johnson
    • Department of Rehabilitation MedicineUniversity of Washington
    • Department of Rehabilitation MedicineUniversity of Washington

DOI: 10.1007/s11136-013-0398-5

Cite this article as:
Askew, R.L., Kim, J., Chung, H. et al. Qual Life Res (2013) 22: 2769. doi:10.1007/s11136-013-0398-5



To help researchers in multiple sclerosis (MS) take advantage of the measurement properties of the PROMIS Pain Interference instrument while maintaining continuity with previous research, we developed and tested a crosswalk table to transform Brief Pain Inventory pain interference scale (BPI-PI) scores to PROMIS-PI short form (PROMIS-PI SF) scores.


The BPI-PI and the PROMIS-PI SF were administered in two studies that included persons with MS. One sample of 369 participants served as a developmental calibration sample, and a separate sample of 360 served as a validation sample. The crosswalk development included dimensionality assessment, item-level parameter estimation, and assessment of accuracy. BPI-PI and PROMIS-PI T scores were obtained from participants’ item responses, and using the crosswalk table, PROMIS-PI T scores were derived from responses to the BPI-PI items. Differences between observed and crosswalked T scores were compared in both samples.


For BPI-PI summary scores ranging from 0 to 10, corresponding T scores ranged from 38.6 to 81.2. The mean difference between observed and crosswalked T scores was 0.51 (SD = 3.9) in the calibration sample and −1.47 (SD = 4.2) in the validation sample. Approximately 80 % of crosswalked scores in the calibration sample were within four score points of the observed PROMIS-PI SF scores, and 70 % were within four points in the validation sample. In both samples, the largest differences were at lower levels of the pain interference continuum.


Crosswalked pain interference scores adequately approximated observed PROMIS-PI SF scores in both the calibration and validation samples. MS researchers and clinicians interested in adopting the PROMIS instruments can use this table to transform BPI-PI scores to enable comparisons with other studies and to maintain continuity with previous research.


CalibrationItem-response theoryBPIPROMISChronic painMultiple sclerosis

Copyright information

© Springer Science+Business Media Dordrecht 2013