Assessment of the psychometrics of a PROMIS item bank: self-efficacy for managing daily activities
Gepubliceerd in: Quality of Life Research | Uitgave 9/2016Log in om toegang te krijgen
The aim of this study is to investigate the psychometrics of the Patient-Reported Outcomes Measurement Information System self-efficacy for managing daily activities item bank.
The item pool was field tested on a sample of 1087 participants via internet (n = 250) and in-clinic (n = 837) surveys. All participants reported having at least one chronic health condition. The 35 item pool was investigated for dimensionality (confirmatory factor analyses, CFA and exploratory factor analysis, EFA), item-total correlations, local independence, precision, and differential item functioning (DIF) across gender, race, ethnicity, age groups, data collection modes, and neurological chronic conditions (McFadden Pseudo R 2 less than 10 %).
The item pool met two of the four CFA fit criteria (CFI = 0.952 and SRMR = 0.07). EFA analysis found a dominant first factor (eigenvalue = 24.34) and the ratio of first to second eigenvalue was 12.4. The item pool demonstrated good item-total correlations (0.59–0.85) and acceptable internal consistency (Cronbach’s alpha = 0.97). The item pool maintained its precision (reliability over 0.90) across a wide range of theta (3.70), and there was no significant DIF.
The findings indicated the item pool has sound psychometric properties and the test items are eligible for development of computerized adaptive testing and short forms.