Abstract
Purpose :
Self-efficacy, the belief in one’s own ability to accomplish goals, is a trait that could influence vision rehabilitation outcomes. The New General Self-Efficacy Scale (SS), is a validated and commonly used questionnaire of self-efficacy that has not yet been evaluated in visually impaired older adults. Here we evaluate psychometric properties of SS in a novel patient population.
Methods :
Visually-impaired adults (age≥55 years; n=122) enrolled in a randomized clinical trial to evaluate the impact of visual-assistive smartphone app use. Prior to intervention, they completed the SS and the Activity Inventory (AI), rating their ability to perform visually-mediated goals and tasks. Rasch analysis was conducted to estimate person measures and item measures based on pre-intervention SS and AI responses. Parameter standard errors and infit mean square statistics (infits) were also estimated from this analysis, and were evaluated to assess SS psychometric properties. Regressions were performed between SS, pre-intervention clinical exam data, and AI person measures.
Results :
Subjects (mean age=73, SD=10.8; mean Visual Acuity=0.73 logMAR, SD=0.38) underutilized rating categories corresponding to poorest self-efficacy (two poorest categories comprised 9% of all rating category selections). The distribution of person measures was shifted towards more positive values (mean=0.77logit, range -2.27,3.24) compared to the distribution of item measures (mean=0, range -0.90,0.64), with no item measures to target person measures >0.64 logit. Item measure standard error (mean=0.69, SD=0.15) and person measure standard error (mean=0.25, SD= 0.04) were greater for the most extreme item/person measures, as is expected by the Rasch model. SS person measure infits (mean=0.94, SD=0.69) approximated a chi-squared distribution (df=7), with infits underrepresented relative to what would be expected in the tails of the distribution. Item infit z-scores fell within ±1.5 SD from the expected value.
Weak correlations were observed between SS and AI (r=0.22) and contrast sensitivity (r=0.19).
Conclusions :
SS displays good psychometric properties, and may represent a unique trait that could influence success in integrating new visual assistive devices by persons with visual impairment. Future analysis should evaluate the impact of SS on patient outcomes following intervention, with consideration for the poor targeting observed in patients with high self-efficacy.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.