Visual impairment and the prevalence of ocular pathology in homeless children and adults globally: a systematic review

Author: Sayal AP, Slomovic J, Bhambra NS, Popovic MM, Lichter M.

Geographical coverage: The United States, Canada, the United Kingdom, Germany, Nigeria, Nepal, and Australia.  

Sector: Epidemiology

Sub-sector: Burden of disease

Equity focus: Homeless

Study population: General population  

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: Homelessness is a global issue in developing and developed countries. Research indicates that homeless individuals experience a higher prevalence of ocular morbidity, glaucoma, cataracts and reduced visual acuity. Poor visual acuity is also associated with diminished income and a reduced quality of life, which may compound the challenges faced by homeless individuals. No study has summarised the overall impact of homelessness on ocular outcomes globally.

Objectives: The aim of this systematic review are three-fold: to synthesise the current research on ocular history, visual impairment, and the prevalence of ocular pathology in homeless individuals globally; to comment on the limitations of the currently available literature; and to suggest how future research could be improved to encourage accuracy in prevalence estimates and to improve accessibility of eye care services for this vulnerable population.

Main findings:

Overall, visual impairment in homeless individuals is higher than the general population, with uncorrected refractive error being leading cause of visual impairment. Homeless adults have significantly more visual impairment and ocular pathology than homeless children.

A total of 24 studies were selected for inclusion (and two were considered together, as they used the same sample). 23 study samples were included in this review that collectively reported on 5,774 homeless individuals. Of the study sample, 71.1% were male and 28.9% were female. The weighted mean age was 37.4 years. Eleven articles were from the United States, six from Canada, two from the United Kingdom, one from Germany, one from Nigeria, one from Nepal, and one from Australia. 32.9% of homeless individuals did not graduate high school, 59.5% were single, 25.5% were divorced, 40.9% earned less than $500 a month, and 55.2% were homeless for longer than one year. 22.0% saw an eye care provider (either optometrist or ophthalmologist) in the past year, 20.9% were referred to an eye care specialist after participation in ocular screening, and 17.0% had their prescription glasses on their person at the time of the study.

Of the 23 articles, all were assessed to have a low risk of bias. The most commonly identified source of bias was a lack of randomisation in sampling the homeless population.

In studies focusing on primary outcomes, authors found that 36.8% of homeless individuals expressed dissatisfaction with their vision. Additionally, 26.8% reported a history of eye disease, 26.3% had uncorrected vision errors, 25.6% were functionally visually impaired, 9.2% had undergone at least one eye surgery or procedure, and 4.0% had non-refractive visual impairment.

Upon examination, auhtors observed that 25.1% of the homeless population had some form of eye disease. This included diseases of the cornea and external eye (13.4%), glaucoma (7.4%), cataracts (6.3%), retinal diseases (5.3%), ocular motility disorders (4.7%), trauma (2.3%), neuro-ophthalmological conditions (1.7%), and oculoplastic conditions (0.7%).

The review concluded that, compared to children, homeless adults had significantly higher rates of visual impairment (p<0.001), uncorrected refractive error (p<0.001), eye disease (p<0.001), cataracts (p<0.001), retinal disease (p<0.001), and neuro-ophthalmological conditions (p<0.001).

Review authors note that future studies need to include more data from under-represented groups, minimise selection bias, work to develop standardised data collection tools, and add comparator groups to establish more sound conclusions. Ultimately, this study highlights the need for programmes offering refractive services for homeless populations and encourages global policy changes that reduce the rate of preventable visual impairment. Future studies should also identify if the differences identified are consistent in developing countries.

Methodology:

The following criteria were applied by the authors to select studies for inclusion: (1) any English-language study published in a peer-reviewed journal; (2) articles that only reported on homeless children or homeless adults as study groups of focus; and (3) studies that included an assessment of ocular pathology conducted by an eye care provider or any quantitative assessment of visual impairment. Nonoriginal articles, abstracts, correspondences, letters to the editors, editorials and repeat data were excluded.

Authors conducted a search on OVID Medline, EMBASE and Cochrane CENTRAL for peer-reviewed studies written in English. Additionally, authors reviewed references of included articles and conducted a search of Google Scholar. Studies were selected through a two-phase process. Full-text review of identified studies were performed by two independent reviewers. The quality of the included articles was appraised by two reviewers independently using a risk of bias tool for prevalence studies modified by Hoy et al., which has been shown to have a high inter-rater agreement.

Results were reported using descriptive statistics. Synthesis of data from the included articles was conducted for outcomes reported by at least two articles. Authors calculate the averages for primary and secondary outcomes and weighted based on sample size. A main analysis reported primary outcomes as a total including all homeless individuals, whereas a subgroup analysis reported on primary outcomes stratified by age (that is, homeless children vs adults). Statistical comparisons between these two subgroups were made if at least two articles reported on an outcome in each group. Given the categorical nature of the data, these comparisons were made using the Fisher’s exact test, and a p value of less than 0.05 was deemed statistically significant.

Applicability/external validity: The authors point out that this is the first systematic review on this topic, and as such, there was limited literature for comparison. Consequently, it was not feasible to ascertain the applicability and validity of the review findings.

Geographic focus: Out of the 23 studies included, two were carried out in Nigeria and Nepal. However, the authors caution that due to the scarcity of studies in developing countries, their prevalence estimates may not be globally applicable.

Summary of quality assessment:

A number of weaknesses in the approaches employed to select, include and critically appraise relevant studies were identified. The search was limited in terms of language and publication status, and it is unclear how inclusive it was in terms of period of coverage. No table was included summarising the characteristics and results of included studies. In terms of the analysis, the review offered limited consideration of heterogeneity and did not attempt to ascertain how the exclusion of studies with a greater risk of bias might affect the reported results. For these reasons, we attributed a low confidence in the findings of this review.