Background In the assessment of chronic pain patients it is important to measure pain behavior and pain interference. The National Institutes of Health's Patient-Reported Outcomes Measurement Information System (PROMIS) initiative developed item banks for measuring pain behavior and pain interference. These PROMIS item banks were translated into Dutch-Flemish language according to the FACIT methodology1.
Objectives The aim of current study was to validate the Dutch-Flemish translation of the PROMIS pain behavior item bank (DF-PROMIS-PB) and the Dutch-Flemish PROMIS pain interference item bank (DF-PROMIS-PI) in patients with chronic pain.
Methods A paper-and-pencil or web-based survey, including the full DF-PROMIS-PB (39 items, 6-point Likert scale) and DF-PROMIS-PI (41 items, 5-point Likert scale), was completed by 1046 chronic pain patients referred to an outpatient secondary care center for rheumatology and rehabilitation in the Netherlands. One-dimensionality was evaluated by one-factor confirmatory factor analysis. With the future strategy to develop computer adaptive tests (CAT), item response theory (IRT) models were used to evaluate the item characteristics of the two item banks. A graded item response model (GRM) was fitted and Differential Item Functioning (DIF) was evaluated for e.g. language (Dutch vs. English), by ordinal regression models. Furthermore, construct validity was studied.
Results DF-PROMIS-PB and DF-PROMIS-PI demonstrated good fit to a one-dimensional model (CFI=0.963; 0.988 resp. and TLI=0.961; 0.987 resp). The first factor accounted for 42% (DF-PROMIS-PB) and 66% (DF-PROMIS-PI) of the questionnaire variance. The results showed acceptable testinformation (SE<0.3) for theta between -1.9 and 3.6 for DF-PROMIS-PB and between -3.3 and 2.8 for DF-PROMIS-PI. 23 out of 741 (3%) DF-PROMIS-PB item pairs and 62 out of 820 (7.6%) DF-PROMIS-PI item pairs were marked as possibly locally dependent. The items demonstrated no DIF with respect to age, gender, and survey version. DIF with respect to language was present for 6 DF-PROMIS-PB items and 2 DF-PROMIS-PI items. However, the impact of DIF on the total item scores was minimal.
Conclusions The DF-PROMIS-PB and the DF-PROMIS-PI fit a GRM and demonstrate good coverage across the range of the pain behavior and pain interference domain. Nearly all Dutch item parameters match the American item parameters and likely Dutch-specific item calibrations are not needed. The DF-PROMIS-PB and DF-PROMIS-PI can be used to develop a CAT.
Terwee CB, Roorda LD, de Vet HCW, Dekker J, Westhovens R, van Leeuwen J, Cella D, Correia H, Arnold B, Perez B, Boers M. Dutch-Flemish translation of 17 item banks from the Patient Reported Outcomes Measurement Information System (PROMIS). Quality of Life Research in press.
Disclosure of Interest : None declared