For maintaining adequate psychometric properties when reducing the number of items from an instrument, item level psychometrics is crucial. Strategies such as low item correlation or factor loadings, using classical test theory, have traditionally been advocated. The purpose of this study is to describe the development of a new short form assessing the impact of low back pain on physical activity. Rasch measurement model has been applied to the International Classification of Functioning, Disability and Health Activity Measure (ICF-AM). One hundred and one individuals with low back pain aged 19-89 years (mean age: 48.1±17.3) who live in the community were participated in the study. Twenty-seven items of lifting/carrying construct of the ICF-AM were analyzed. Ten items were selected from the construct to create a short form. Item elimination criteria include: 1) high or low mean square (out of the range: .6-1.4 for the fit statistics), 2) similar item calibrations to adjacent items, 3) person separation value, and item-person map for potential gap in person ability continuum. All 10 items of the short form fit to the Rasch model except one item (i.e., carrying toddler on back). Despite its high infit and outfit statistics (1.90/2.17), the item had to be reinstated due to potential gaps at the upper extreme of person ability level. The short form had a slightly better spread of person ability continuum compared to the entire set of item. The created short form separated individuals with low back pain into nearly 4 groups, while the entire set of items separated the individuals into 6 groups. The findings prompted multidimensional models for better explanation of the lifting/carrying domain. The item level psychometrics based on the Rasch model can be useful in developing short forms with rationally retained items.

Differential item functioning (DIF) based on Rasch model can be used to examine whether the items function similarly across different groups and identify items that appear to be too easy or difficult after controlling for the ability levels of the compared groups. The Oswestry low back pain disability (Oswestry) has traditionally been proved as an effective instrument measuring disability resulting from low back pain (LBP). In this study, DIF method was used to explore whether items on the Oswestry perform similarly across two different groups (participants with LBP and no LBP). A series of Rasch analyses on the 10 items of the Oswestry were performed using Winsteps(R) software. Forty-two participants with back pain were recruited from 3 rehabilitation hospitals in Gainesville, Florida. Another 42 participants with no LBP were recruited from several public places in the rehabilitation hospitals. Based on the DIF analysis across the two groups, several items were found to have an uniform DIF. Participants with no LBP had more difficulty on lifting and personal care items and participants with LBP had more difficulty on sleeping and social life items. For non-LBP group, a high ceiling effects (83% of participants with non-LBP) was detected, which was not be able to be effectively measured with the Oswestry items. Although 4 items of the Oswestry function differently across the two groups, all items of the Oswestry were well targeted the LBP group.