Development and Validation of the Life for Low Vision Questionnaire (LIFE4LVQ) Using Rasch Analysis: A Questionnaire Evaluating Ability and Independence

Abstract

Low vision (LV) has a substantial impact on an individual's daily functionality and patient-reported outcome measures (PROMs) are increasingly incorporated into the evaluation of this problem. The objective of this study was to describe the design of the new "Life for Low Vision Questionnaire (LIFE4LVQ)", as a measure of daily functionality in LV and to explore its psychometric properties. A total of 294 participants completed the LIFE4LVQ and the data were subjected to Rasch analysis to determine the psychometric properties of the questionnaire, including response category ordering, item fit statistics, principal component analysis, precision, differential item functioning, and targeting. Test-retest reliability was evaluated with an interval of three weeks and intraclass correlation coefficients (ICC) were used. The correlation between the questionnaire score and Best Corrected Visual Acuity (BCVA) was examined using Spearman's correlation coefficient. Rasch analysis revealed that for most items the infit and outfit mean square fit values were close to 1, both for the whole scale and its subscales (ability and independence). The separation index for person measures was 5.18 with a reliability of 0.96, indicating good discriminant ability and adequate model fit. Five response categories were found for all items. The ICC was 0.96 (p < 0.001; 95% CI, 0.93-0.98), suggesting excellent repeatability of the measure. Poorer BCVA was significantly associated with worse scores (rho = 0.559, p < 0.001), indicating excellent convergent validity. The functional, 40-item LIFE4LVQ proved to be a reliable and valid tool that effectively measures the impact of LV on ability and independence.

Keywords: PROMs; Rasch analysis; ability; functionality; independence; low vision.

Figure 1

Category probability curves (CPCs) for the LIFE4LVQ. Each curve illustrates one response category (no difficulty/never = 1, little difficulty/little = 2, some difficulty/sometimes = 3, great difficulty/often = 4, and unable to do because of my vision/always = 5). The point where two adjacent curves overlap is the threshold. Thresholds in this case are adequately ordered.

Figure 2

Person–item map for the LIFE4LVQ. Participants (person ability) are presented on the left and the items (item difficulty) are on the right of the dashed line. Each “#” and “.” represent two participants and one participant, respectively. “Q” means the LIFE4LVQ item, and “M” indicates the mean measure (left, person ability; right, item difficulty). “S” shows 1 SD from the mean and “T” indicates 2 SDs, all expressed in logits. SD = standard deviation.