ADVANCES IN PAIN MEASUREMENT- ITEM RESPONSE THEORY (IRT) BASED METHODS AND THE PAIN IMPACT QUESTIONNAIRE (PIQ-6)
Author(s)
Janine Becker, PhD, Dipl-Psych, Scientist, Renee N. Saris-Baglama, PhD, Sientist, Jakob B. Bjorner, PhD, MD, Chief Science Officer, Mark Kosinski, MA, Scientist, John E. Ware, PhD, CEO, Chairman of the Board & Senior Scientist QualityMetric Incorporated, Lincoln, RI, USA
OBJECTIVES: To use IRT-based methods to develop a six-item questionnaire (PIQ-6) measuring pain severity and pain impact. IRT was used to: 1) identify the most informative items from a bank of 65 items, 2) evaluate construct validity, 3) norm the PIQ-6 to the US general population; 4) cross-calibrate it to the SF-36v2 BP scale; 5) develop a simple scoring approach based on response category weights matching the IRT score; and 6) build a computerized version to facilitate monitoring and management of pain. METHODS: Items analyses and selection were based on data from two web-based general population samples (n=782, n=829) and a chronic pain patients sample (n=306). Norming was achieved by recalibrating the IRT-based item parameter using 1998 US representative norm data of the SF36v2 (n=7069). Cross-calibration of the PIQ-6 and SF-36v2 was performed by estimating the expected SF-36v2 BP scale score using IRT-based item parameter. The PIQ-6 computerized version was programmed with Microsoft Visual Basic. RESULTS: Six items were identified that fitted an unidimensional IRT model, showed high measurement precision and did not show differential item functioning. Construct validity was supported by high correlations with other pain measures and strong discrimination of pain patients and the general population. The PIQ-6 is normed so that a score of 50 represents the US general population norm (sd=10). A cross-calibration table of the PIQ-6/SF-36v2 was created to facilitate communication between researchers using those different instruments. The simple hand scoring approach had satisfactory agreement with the IRT scores. The PIQ-6 was programmed as computerized stand-alone/PDA/online versions allowing for an easy administration, scoring and immediate feedback reporting including interpretation guidelines. CONCLUSIONS: Pain measurement can be improved by using IRT and computerized methods. Noteworthy advantages of the PIQ-6 are the enhancement of measurement precision and coverage as well as the facilitation of monitoring and managing pain.
Conference/Value in Health Info
2006-05, ISPOR 2006, Philadelphia, PA
Value in Health, Vol. 9, No.3 (May/June 2006)
Code
PPN5
Topic
Patient-Centered Research
Topic Subcategory
Patient-reported Outcomes & Quality of Life Outcomes
Disease
Systemic Disorders/Conditions