Social inequities and visual function in pediatric ophthalmology populations: a small cross-sectional study


 Objectives
 The relationship between visual function and social determinants of health in pediatric ophthalmology patients is unclear. Our study evaluated the feasibility of screening for social determinants of health in ophthalmology clinics, assess the prevalence of these risk factors, and to clarify the relationship between social determinants of health and visual function.
Methods
 An institution approved survey on social determinants of health risk factors was completed by 145 patients from five pediatric ophthalmology outpatient clinics in British Columbia. Medical charts of survey participants were reviewed to determine diagnoses and level of visual function. Descriptive statistics, univariate, and multivariate analysis were performed.
Results
 Socioeconomic risk factors were present in all pediatric ophthalmology settings. 56% (n=81) of participants reported having at least one risk factor. Characteristics of poverty, including annual household income, size of support network, adverse childhood experiences, and level of parental education, were not associated with level of visual function. Food insecurity, housing instability, low income, and lack of social support were all associated with a higher adverse event scores (p<0.05). Patients who experienced food insecurity were more likely (OR 7.14, 95% CI 1.47-38.44) to have an adverse childhood experience score of four or more compared to patients who did not have any risk factors.
Conclusions
 Our study found no relationship between the level of visual function and social determinants of health risk factors. Given that all pediatric ophthalmology populations experience socioeconomic poverty, the need to establish social screening and social work support is pertinent in all pediatric ophthalmology clinics.


Introduction
Approximately 19 million of the world's children are visually impaired(1) and 6.8% of children under the age of 18 in the United States have a diagnosed eye and vision condition.(2) Good vision is crucial for a child's physical, emotional, social, and developmental well-being as the impact of having a vision disorder at any age can be felt for the remaining years of life. (3)(4)(5) At the same time, the World Health Organization (WHO) has identi ed social determinants of health (SDOH), which include income, food security, housing stability, academic attainment, and healthcare access, to have a far greater in uence on overall health outcomes than the medical care being provided.(6, 7) Children are especially vulnerable to the effects of SDOH, as the number of SDOH risk factors, and the length of exposure to those risk factors, have a cumulative effect as they become older. (8) While there have been numerous population-based studies on the prevalence and demographic risk factors of visual impairment in children, the relationship between poverty and vision remains unclear. (3,(9)(10)(11)(12)(13)(14)(15) We conducted a cross-sectional study that included ve different pediatric ophthalmology populations across the province of British Columbia (BC) to 1) determine whether screening for SDOH and ACEs was feasible in outpatient pediatric ophthalmology settings, 2) report the prevalence of negative socioeconomic experiences in different pediatric ophthalmology populations, and 3) investigate associations between SDOH and level of visual function in children with a diagnosed vision disorder. Findings from this study will help identify gaps in quality of care, guide allocation of social work resources, and demonstrate the importance of social screening that may reveal health inequities and accessibility barriers in pediatric ophthalmology.

Materials And Methods
To determine the effects of SDOH in pediatric ophthalmology patients and families, an institution-approved survey was built using current best practices re ected in the literature and included questions on demographics; social and material capital; healthcare utilization and access; food, housing, and nancial security; and an optional adverse childhood experiences questionnaire. Visually Impaired Program (VIP), Prince George (PG) Pediatric Ophthalmology Outreach Clinic, and a private practice Community Clinic (CC). Patients were recruited from multiple sites to increase the generalizability of our ndings. To be eligible to for the study, participants had to be English-speaking caregivers attending an appointment with their child. At the time of recruitment, participants underwent an informed consent process prior to completing a survey about ACEs and SDOH risk factors. Ophthalmological information including visual acuity (which was measured using Snellen, Teller or Allen visual acuity charts depending on the child's ability) and eye condition were extracted from the charts of participants who completed the survey. This prospective cross-sectional study obtained ethics approval from the University of British Columbia's Clinical Research Ethics Board (H19-03822) and was conducted in accordance with the principles of the Declaration of Helsinki.
Patients were classi ed into three mutually exclusive groups for visual function based on their best-corrected visual acuity and the etiology of their eye disease: normal (or functional) vision (visual acuity of 20/50 or better, which is also the provincial standard for driving);(17) impaired vision (visual acuity between 20/60 and greater than 20/200) and blind (visual acuity of 20/200 or less).
Patients were also categorized into seven mutually exclusive categories for their diagnoses based on the etiology that was most responsible for their visit.
These seven categories include: external (eyelid, orbit and nasolacrimal system) disease, anterior segment (cornea, conjunctiva, iris, and lens), neuroophthalmology, retina, refractive error, strabismus, and screening. The screening category referred to patients who had a normal eye exam. If a patient had multiple diagnoses, (ie. refractive error and optic disc atrophy), the diagnosis most responsible for their visit was used for classi cation.
All statistical analyses were performed using GraphPad Prism Version 8.4.3 (GraphPad Software, Inc., CA, US) and RStudio Version 1.2.5033 (RStudio, PBC, MA, USA). If a participant did not answer a particular question, the missing data was not included in the analysis, and the sample size was adjusted accordingly. Descriptive analyses were used to summarize categorical variables of interest in frequency and percent. Principle component analysis was used to determine which of the numerous variables captured from the survey best described the variation in the dataset. Pearson's chi-square tests were then used to investigate the relationship between clinic site, visual impairment, and demographics of the patient population including annual household income, number of people the family can turn to for support, number of ACEs, and highest level of parental formal education. P-values were adjusted using the Bonferroni method to account for multiple comparisons. Throughout analysis, p<0.05 was considered statistically signi cant. If initial Pearson's chisquared tests demonstrated signi cance, a post hoc multiple comparisons chi-squared test was applied to elucidate the source of signi cant differences.
Multiple logistic regression analysis was also used to examine the relationship of a priori speci ed SDOH variable of interest and having an ACEs score of four or more. Crude and adjusted odds ratios, as well as 95% con dence intervals (CIs) were reported.

Results
Our survey's respondents came from diverse regions of the province and re ected a heterogeneous distribution of ocular conditions. Of the 145 participants, 45 patients were from BCCH, a tertiary care ophthalmology clinic, while 32 patients were from VIP, a clinic for children with low vision. Twentysix patients came from DEEC, a pediatric ophthalmology clinic operating in one of Canada's poorest neighbourhoods, while 21 patients were from PG, a rural outreach clinic that serviced patients from northern regions of BC. Finally, 21 patients from CC, which was a general pediatric ophthalmology clinic in the community. Most of the patients had a family size of four to six members (62.1%) and about a third (37.2%) of the patients identi ed as being in a visible, ethnic or cultural minority. Over a quarter (25.5%) of patients had parents whose highest level of formal education was a high school diploma or less. 28.3% of respondents had an annual household income below $40,000, which is the poverty income line for a family of four in BC.

Differences between clinics
Apart from the Visually Impaired Program, in which patients were referred to by a pediatric ophthalmologist after con rmation of having a diagnosed visual impairment, the majority of patients seen at the other four clinics had normal vision (Supplementary eFigure 1). The DEEC and CC had patient populations of similar visual function, with refractive errors being the most common diagnosis at these two clinics and neither having patients who were blind.
Likewise, BCCH and PG also had similar patient population visual function pro les, with both having a wider variety of vision problems (Supplementary eFigure 2).
Although the DEEC and CC had patients of similar visual function and diagnoses, compared to patients from the CC, patients at the DEEC were more likely to experience low household income (p=<0.0001), few social supports (p<0.01), high ACEs (p=<0.0001), and low levels of parental education (p<0.001) ( Figure 1). In contrast, while the BCCH and PG pairing also had near-identical patient population pro les in terms of visual function and diagnoses, there were no differences in negative social determinants experienced by patients between these two clinics ( Figure 1).

Differences between visual function
Excluding the VIP clinic in which all patients were blind or had impaired vision, 1.8% (n=2) of the study population were blind and 3.5% (n=4) had impaired vision. The top three causes of blindness for the entire study population were neuroophthalmological (63.6%, n=14), retinal (31.8%, n = 7) and anterior segment (4.5%, n =1) diseases (Supplementary eFigure 3). This is consistent with literature reports that have listed cortical visual impairment, optic nerve hypoplasia, retinopathy of prematurity, optic atrophy, albinism and coloboma as being the top causes of childhood blindness.(21-23) Common causes of impaired vision included retinal (50%, n=8) and neuroophthalmological (37.6%, n=6) diseases. Interestingly, there were no associations between the level of visual function and SDOH risk factors, including annual household income, social support, ACEs or level of parental education. (Figure 2).

Relationship between ACEs and SDOH
Of the 145 participants, 103 (71.0%) answered the ACEs questionnaire. Four (food security, housing stability, income, and social support) of the ve SDOH risk factors had a statistically signi cant (p < 0.05) negative association with ACEs (Table 1). There was no signi cant relationship between the level of parental education and ACEs. After adjusting for SDOH co-variates, only one of the ve SDOH risk factors (food insecurity) was associated with a statistically signi cant greater likelihood (OR 5.5, 95% CI 1. 24-25.25) of an ACEs scores of four or more compared to patients who did not have any SDOH factors ( Table 2).

Discussion
Visual impairment is one of the most prevalent causes of childhood disability, (2) and while social determinants of health are known barriers to ophthalmic follow-up care, (24)(25)(26) there have been mixed reports of their in uence on visual function in the pediatric ophthalmology population. (3,9,(12)(13)(14)(15) Our study demonstrates that social screening for SDOH and ACEs can be feasible in an outpatient setting, and that appropriate resources can be made available in the clinic to address concerns arising from patients with signi cant negative SDOH and/or high ACEs scores. The results of this study found socioeconomic risk factors, such as low income, social isolation, lack of a xed address, and limited access to nutritious foods, to be present in all pediatric ophthalmology settings, and ophthalmologists should be cognizant of this when considering barriers to care and treatment compliance. Statistically signi cant relationships between childhood socioeconomic vulnerability and ACEs were also detected on univariate and multivariate analyses.
We did not nd any associations between visual function and SDOH risk factors, a nding that has also been reported in the literature.(9) Patients with visual impairment were no more likely to come from households of lower education, of lower income, or of lower social network than those with no visual impairment. Moreover, when we compared SDOH between clinics that had patients with similar visual function, no relationships between the level of visual impairment or diagnoses and the various SDOH risk factors were observed. For example, the DEEC and CC had near-identical types of patients in which patients from these two clinics were usually referred by family doctors, pediatricians or optometrists and often did not require tertiary care, yet there were signi cant differences in social determinants experienced between the two clinic patient populations.
The strengths of our study include the successful implementation of our screening questionnaire for SDOH risk factors in different pediatric ophthalmology outpatient clinic settings as well as the heterogeneity of pediatric eye conditions captured. Some limitations in our results include the relatively small sample size, and the possibility of recall and reporting bias in self-reported survey responses. In addition, while this study examined numerous social and nancial factors related to health, not all possible SDOH risk factors were included. Finally, because participants represented in our convenience sample were those who were present for their appointments, patients who face additional accessibility barriers to attending appointments may not have been represented in our study, and thus generalizing our results should be done with caution.
Overall, poverty can signi cantly alter a child's health and life trajectory, and public and professional stakeholders need to recognize the impacts of material deprivation, social isolation, and adverse childhood experiences on pediatric ophthalmology patients and their families. While SDOH risk factors were found in all our pediatric ophthalmology clinic settings, they were not associated with level of visual function or diagnosis. Future directions include larger studies to con rm the ndings observed and evaluating the utility of social work support in pediatric ophthalmology outpatient clinics.