Graded response model fit, measurement invariance and (comparative) precision of the Dutch-Flemish PROMIS® Upper Extremity V2.0 item bank in patients with upper extremity disorders

Abstract

Background: The Dutch-Flemish PROMIS® Upper Extremity (DF-PROMIS-UE) V2.0 item bank was recently developed using Item Response Theory (IRT). Unknown for this bank are: (1) if it is legitimate to calculate IRT-based scores for short forms and Computerized Adaptive Tests (CATs), which requires that the items meet the assumptions of and fit the IRT-model (Graded Response Model [GRM]);(2) if it is legitimate to compare (sub) groups of patients using this measure, which requires measurement invariance; and (3) the precision of the estimated patients' scores for patients with different levels of functioning and compared to legacy measures. Aims were to evaluate (1) the assumptions of and fit to the GRM, (2) measurement invariance and (3) (comparative) precision of the DF-PROMIS-UE v2.0. Methods: Cross-sectional data were collected in Dutch patients with upper extremity disorders. Assessed were IRT-assumptions (unidimensionality [bi-factor analysis], local independence [residual correlations], monotonicity [coefficient H]), GRM item fit, measurement invariance (absence of Differential Item Functioning [DIF] due to age, gender, center, duration, and location of complaints) and precision (standard error of IRT-based scores across levels of functioning). To study measurement invariance for language [Dutch vs. English], additional US data were used. Legacy instruments were the Disability of the Arm, Shoulder and Hand (DASH), the QuickDASH and the Michigan Hand Questionnaire (MHQ).

Development and validation of a short version of the Sarcopenia Quality of Life questionnaire: the SF-SarQoL

“…Monotonicity was evaluated with Mokken scaling carried out with the R package "Mokken" (version 3.0.2), using the scalability coefficient H for each item and the questionnaire in its entirety. The assumption of monotonicity was confirmed if the item scalability coefficients were ≥ 0.3 and the scalability coefficient H i for the entire questionnaire was ≥ 0.5 [36].…”

“…In other words, the item responses should be entirely a function of the underlying trait, and not (partly) dependent on a second factor [34]. To determine this, we looked at the residual correlation matrix from the previously described single-factor CFA, and considered a value of 0.2 above the average residual correlation as the cut-off for local independence [36]. (3) Lastly, the concept of monotonicity was examined.…”

Psychometric assessment of scales for measuring loneliness and social isolation: an analysis of the household, income and labour dynamics in Australia (HILDA) survey

“…We ran a MIRT graded response model using flexMIRT software [ 53 ] which is used for polytomous responses such as Likert scales [ 54 ], to examine the extent to which the items differentiate among individuals across different degrees of the underlying traits. The slope parameter ( a ) refers to the item’s discriminative ability, with higher values indicating a stronger association with the construct [ 55 ], and the intercept ( d ) refers to the log-odds of responding above a given category when the latest trait θ = 0. Items with a slope parameter ( a ) between 1.35 and 1.70 were considered to have high discrimination and above 1.70 considered to have very high discrimination [ 56 ].…”

Multidimensional Item Response Theory to Assess Psychological Properties of GHQ-12 in School Children Parents

“…To evaluate the item fit, the generalized Orlando and Thissen's S-X 2 index for polytomous data was used [35], comparing the observed and expected response frequencies under the estimated MIRT model. Eventually items with S-X 2 p-value<0.001 were considered poorly fitted [36].…”