Validation of the brain injury associated visual impairment - impact questionnaire (BIVI-IQ)

Abstract

Purpose: The Brain Injury associated Visual Impairment - Impact Questionnaire (BIVI-IQ) was developed to assess the impact of post-stroke visual impairment. The development of the questionnaire used robust methods involving stroke survivors and clinicians. The aim of this study was to assess the validity of the BIVI-IQ in a stroke population.

Methods: Stroke survivors with visual impairment were recruited from stroke units, outpatient clinics and non-healthcare settings. Participants were asked to complete questionnaire sets on three separate occasions; the BIVI-IQ at each visit with additional questionnaires at baseline and visit 2. Vision assessment and anchor questions from participants and clinicians were collected. The analysis included assessment of missing data, acceptability, Rasch model analysis, test-retest reliability, construct validity (NEI VFQ-25, EQ-5D-5L) and responsiveness to change.

Results: 316 stroke survivors completed at least one questionnaire of the 326 recruited. Mean age was 67 years and 64% were male. Adequate fit statistics to the Rasch model were reached (χ² = 73.12, p = 0.02) with two items removed and thresholds of two adjusted, indicating validity and unidimensionality. Excellent test-retest reliability was demonstrated (ICC = 0.905) with a 3-month interval. Construct validity was demonstrated with a strong significant correlation to the NEI VFQ-25 (r = 0.837, p < 0.01). The BIVI-IQ also demonstrated responsiveness to change with significant differences identified between groups based on participant and clinician anchor questions (X² = 23.29, p < 0.001; X² = 24.56, p < 0.001).

Conclusion: The BIVI-IQ has been shown to be valid and practical for 'everyday' use by clinicians and researchers to monitor vision-related quality of life in stroke survivors with visual impairment.

Keywords: Brain injury; Impact; PROM; Quality of life; Stroke; Visual impairment.

Fig. 1

Flow of participants through different validation analyses

Fig. 2

Reasons for non-return of all three questionnaire packs. Refused follow-up refers to patients who have repeatedly have cancelled appointments or expressed they no longer wish to attend appointments; No longer wished to take part they will continue to attend clinic appointments but do not wish to continue to take part in research

Fig. 3

Types of visual impairment present at baseline assessment

Fig. 4

Percentage of response type to each item of the QQ-10 (n = 271) in relation to the BIVI-IQ

Fig. 5

Probability curves for misfitting items 5 and 13 from initial Rasch model analysis. Each curve represents the probability (y-axis) of the selection of each response option (0 = none, 1 = some, 2 = a lot, 3 = unable to do/limits activity)

Fig. 6

Person item distribution for the 13-item BIVI-IQ. A minus logit is indicative of a smaller level of impact on quality of life (persons) and a more demanding activity (items). A positive logit is indicative of a larger level of impact on quality of life (persons) and a less demanding activity (items)

Fig. 7

Box-plots showing the change in BIVI-IQ dependent on the perception of change reported by participants. A lower score indicates smaller level of impact and a higher score a larger level of impact

Fig. 8

Box-plots showing the change in BIVI-IQ dependent on the clinician summary of the objective clinical assessment. A lower score indicates smaller level of impact and a higher score a larger level of impact