Calibration and initial validation of a general self-efficacy item bank and short form for the NIH PROMIS®

Abstract

AimsGeneral self-efficacy is associated with adaptive coping and positive health outcomes. The Patient-Reported Outcomes Measurement Information System (PROMIS®) has developed self-efficacy item banks for managing chronic conditions, but lacks a general self-efficacy measure. We sought to refine and validate an item-response theory (IRT)-based measure of general self-efficacy for PROMIS®.MethodsTen items were modified from the NIH Toolbox® Self-Efficacy Item Bank by creating “confidence” response options, and administered to a general population sample (n\u2009=\u20091000) with the Toolbox® Self-Efficacy Item Bank, Life Orientation Test-Revised (LOT-R), and Generalized Expectancy for Success Scale (GESS). We split the sample in half for exploratory factor analysis (EFA) followed by confirmatory factor analysis (CFA). IRT analyses included evaluation of differential item functioning (DIF).ResultsParticipants had a mean age of 47.8\xa0years and 50.3% were male. EFA showed all items loaded onto one dominant factor and CFA yielded a good fitting model for a general self-efficacy bank with confidence response options (CFI\u2009=\u20090.987, TLI\u2009=\u20090.984, RMSEA\u2009=\u20090.090). Items showed no evidence of DIF by gender, age, education, or race. Internal consistency reliability was α\u2009=\u2009.94 and .88 for a new 10-item general self-efficacy bank and 4-item short form, respectively. The new bank was correlated with the LOT-R (r\u2009=\u2009.58), the GESS (r\u2009=\u2009.55), and the Toolbox® Self-Efficacy Item Bank (r\u2009=\u2009.87).ConclusionsThe PROMIS® General Self-Efficacy measure demonstrated sufficient unidimensionality and displayed good internal consistency reliability, model fit, and convergent validity. Further psychometric testing of the PROMIS® General Self-Efficacy Item Bank and Short Form can evaluate its utility in people with chronic health conditions.