Predicting the Need for Low Vision Rehabilitation Services

Abstract

Purpose:: To determine the independent predictors of rehabilitation needs of people with low vision.

Methods:: Patients attending low vision clinics completed the Impact of Vision Impairment questionnaire (IVI) and provided general sociodemographic and clinical information. Rasch analysis was used to generate person measures for the IVI overall and three domain scores. Four models of linear regression analysis (LRA) were used to determine independent predictors of rehabilitation needs.

Results:: 477 patients (56% women) with a mean age 72 yrs (SD±15.3) were recruited. Most (74%) had moderate or severe vision loss (presenting visual acuity (PVA) <6/18) and Age-Related Macular Degeneration (43%). Age, duration and main cause of vision impairment, PVA, the impact of comorbidities on daily living and reliance on family or friends for meal provision were univariately associated with poorer IVI overall and domain scores (p<0.05). Patients requiring family assistance with chores also had poorer scores on all the IVI scores except for the emotional well-being domain. Females performed significantly worse on the mobility and independence domain and participants from culturally and linguistically diverse backgrounds recorded poorer emotional well-being domain scores (p<0.05). LRA consistently ranked PVA, the effect of comorbidity on life activities and reliance on family or friends for meals as the highest independent predictors of need in each of the four IVI dependent scores. The IVI overall regression model explained 29% of the variance. PVA made the largest unique contribution (ß=.39 and R²= 13.96) followed by the effect of comorbidity on life activities (ß=.17 and R²=2.59) and reliance on family or friends for meals (ß=.15 and R²=1.52). The reading and accessing information domain model replicated these values. The emotional well-being and mobility and independence models explained 16.7% and 24.6% of the variance respectively. PVA, the effect of comorbidity on life activities and reliance on family or friends for meals remained the highest three predictors in each (ß=.19 and R²=3.27, ß=.22 and R²= 4.51, ß=.12 and R²= 1.30) and (ß=.31 and R²= 8.67, ß=.21 and R²= 4.10, and ß=.15 and R²= 1.50) respectively.

Conclusions:: Rasch-calibrated person measures confirmed PVA as the strongest consistent predictor of overall quality of life and three domain scores in people with low vision. Comorbidity and ‘needing home help’ also featured. Critically, the association between poorer vision and decreased independence indicates that PVA can be used by clinicians to determine who should be referred to low vision rehabilitation services.

Keywords: low vision • quality of life