Longitudinal ophthalmic outcomes after tertiary referral for patients experiencing homelessness

Introduction

People experiencing homelessness (PEH) face significant barriers to eye care, resulting in higher rates of preventable visual impairment. Common causes of poor vision in this population include uncorrected refractive error (RE), followed by cataracts, corneal and external eye disease, glaucoma and retinal disease (1). A meta-analysis found that 25% of PEH had non-RE ocular pathology (1). Studies have shown that PEH have worse visual acuity (VA), higher rates of vision-threatening conditions, and are more likely to need surgery compared to their housed counterparts (1-3). Barriers to PEH ophthalmic care include transportation difficulties, lack of insurance, limited awareness of available resources, and negative prior experiences with healthcare providers (4-7).

Most prior studies have focused on initial screening in traditional clinics, homeless shelters, temporary/mobile clinics or non-ambulatory locations such as homeless encampments (1,8). These studies commonly include follow-up interventions, such as referrals to nearby tertiary eye care centers for those identified with pathology. Only a few have tracked whether these patients present to their tertiary clinic appointment; these rates range from approximately 32–59% for PEH (9-11). This rate is lower than their housed counterparts; housed patients are almost twice as likely to follow-up and complete recommended surgery compared to the PEH patients (2). There is limited research on what happens after PEH are referred to tertiary care. While a few studies document cataract surgery in this population, none have assessed long-term compliance or outcomes across treatments such as multiple intravitreal injections, laser procedures, and non-cataract surgeries over an extended period (2,10,12).

While it is difficult to get specific numbers, adults without health insurance is estimated to be 21.3% in Harris County, one of the highest in USA (13). PEH tend to have less insurance coverage compared to the general population, especially in states like Texas, which is a Medicaid Non-Expansion State (14). The Health Care for the Homeless Grant was established in 1988 by the Harris Health System in Houston, Texas (15). The grant provides health, mental health and dental services through six clinic-based and three shelter-based locations. Qualified patients receive full coverage for clinical visits and surgical procedures in the Harris Health System. Ben Taub Hospital, a safety-net hospital and part of the Harris Health System, provides ophthalmology services staffed by resident and attending physicians from Baylor College of Medicine. The emergency department (ED) in the hospital is also covered by Baylor ophthalmologists 24 hours a day.

The primary objective of this study was to assess the one-year compliance rate of PEH in a tertiary eye center. Secondary objectives were to evaluate changes in VA over the study period, track the number and type of ophthalmology procedures and characterize referral sources. We present this article in accordance with the STROBE reporting checklist (available at https://aes.amegroups.com/article/view/10.21037/aes-2026-1-0003/rc).

Methods

This retrospective, single-center cohort study examined follow-up patterns among PEH treated by the ophthalmology service at Ben Taub Hospital between May 2024 and May 2025.

Eligible patients had at least one documented ophthalmology encounter and used Healthcare for the Homeless Grant for at least one visit during the study period. Because housing status is not systematically documented in the electronic health record, enrollment in the Homeless Grant served as a proxy for current or recent homelessness. PEH without insurance or using a different insurance plan were excluded from the study.

Patients were considered compliant if they met both of the following criteria: (I) attendance at least 75% of their scheduled visits between May 2024 to May 2025, (II) documentation of formal clinic discharge or a future appointment scheduled beyond May 2025. Patients who did not meet both criteria were deemed non-compliant. Patients who were discharged to community health providers (CHP) after their initial visit were not included in compliance calculations.

Referral sources for the patients included referrals from CHP optometrists, primary care physicians, and self-referral. Established patients in the clinic that continued to follow-up during the study period were also included in the study.

Statistical analysis

Demographic characteristics, diagnoses, procedures, referrals and VA data were summarized using univariate analysis. VA was converted into logarithm of the minimum angle of resolution (logMAR) by using a standard conversion table. Compliance status was compared with gender using the Fisher-exact test, with referral source using the Chi-square test, and with age. The worse eye was selected for each patient for the initial VA, and final VA using the Mann-Whitney U test. Multivariate logistic regression was performed with predictors identified in the univariate screening. Due to the non-parametric distribution of the data, the Wilcoxon signed-rank test for paired samples was used to assess the difference between initial and final VA in the same patients over time. A two-sided P value of <0.05 was considered statistically significant. All statistical analyses were done using Microsoft Excel (Version 2506, Microsoft Corporation, Redmond, WA, USA).

Ethical considerations

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional ethics board of Baylor College of Medicine (ESP1: H-57931), and individual consent for this retrospective analysis was waived.

Results

A total of 68 patients were identified through a retrospective review of the electronic records. Figure 1 illustrates the breakdown of patient appointments. CHP optometrists referred 13/68 patients (19%) for testing appointments, including optical coherence tomography (OCT) of the macula and of the retinal nerve fiber layer and Humphrey Visual Field testing. These patients completed their tests and left the clinic without encountering any medical personnel. The remaining 55/68 (81%) received a standard ophthalmic work-up including VA measurement, intraocular pressure (IOP) measurement, pupillary examination, confrontational visual fields and extraocular movement testing. They were also examined with slit-lamp biomicroscopy and dilated fundus exam. While the majority of the first encounters occurred in the outpatient ophthalmology clinic, encounters in the ED and inpatient units were also included. Patients were then either referred back to CHP if stable or given a follow-up appointment. If needed, patients were scheduled for intravitreal injections, in-office laser procedures or surgeries at their initial or follow-up appointments.

Figure 1 Eye care referral and treatment for persons experiencing homelessness (PEH). Screening of PEH at stand-alone clinics, homeless shelters, temporary/mobile clinics and through street medicine programs. A total of 68 patients needing a higher level of care were referred to the ophthalmology service at Ben Taub Hospital between May 2024 and May 2025. Of these 13 (19%) were testing-only visits without an encounter with a physician. The remaining 55 patients (81%) were PEH on the Homeless Grant and were evaluated by an ophthalmologist. 10/55 (18%) were deemed stable and were referred back to community health provider (CHP) optometrists. 4/55 (7%) were lost to follow-up (LTFU) after the initial visit. The remaining 41/55 (75%) continue follow-up (F/U) at the Ben Taub clinic. Among these patients, 23/41 (56%) underwent a procedure either in clinic or in the operating room. Figure created by Miranda Pruitt using information provided by the authors. F/u, follow-up.

Table 1 summarizes the demographic characteristics of the 55 patients seen by the ophthalmology service during the study period. The median number of follow-up visits was 4.0, with an interquartile range (IQR) of 4.5. Table 1 also shows the five most common diagnoses.

Ten patients were discharged back to CHP after their initial visit. For the remaining patients, compliance was defined as sustained follow-up after the initial visit. Out of the 45, 41 (91%) of patients followed up at least once after the initial visit. The compliance of patients by referral source and patient status is shown in Table 2. Based on this definition, 23/45 (51%) were compliant, while 22/45 (49%) were non-compliant. In univariate analysis only referral source and number of clinic visits were statistically significant with patient compliance (Table 3). In multivariable logistic regression (Table 4), the number of clinic visits remained significantly associated with compliance [odds ratio (OR) =1.60, 95% confidence interval (CI): 1.14–2.23, P=0.006]. Referral source was also significant (OR =2.23, 95% CI: 1.01–4.91, P=0.047).

Among the patients who continued to follow-up throughout the year, 23 (56%) patients required at least one procedure. Procedures were defined as either intravitreal injection or laser treatments performed in clinic, or surgical interventions done in the operating room. A total of 76 separate procedures were performed. Table 5 summarizes the types and frequencies of the procedures performed. Some patients received multiple procedures or combined surgeries. Surgeries were only counted as a separate procedure if performed on different dates. For example, one patient who underwent two pars plana vitrectomies on different dates was counted twice. Conversely, a combined phacoemulsification and pars plana vitrectomy performed during a single visit was recorded only once, under the latter procedure.

The most frequent procedures were intravitreal injections and cataract surgeries. Among the 23 patients who underwent procedures, the median number of procedures per patient was 3.0 (IQR: 3.0–3.0).

Tables 6,7 evaluate the VA outcomes. In part A, the right and left eye of 41 patients with at least one follow-up were assessed. There was no statistically significant change in VA for either eye. Part B presents a sub-group analysis of 35 eyes from 23 patients (combining both right and left eyes) that underwent treatment with a procedure. This group demonstrates a statistically significant improvement in VA equivalent to three lines on a vision chart (equivalent to 0.30 logMAR). In contrast the non-treated subgroup had a difference of −0.01 logMAR. The Wilcoxon signed-rank test for the worse eye in these patients yielded a test statistic of 27 for N=23 (P=0.02), indicating statistical significance.

Regarding referral sources, most patients were established with the clinic, i.e. existing clinic patients (21, 38%), followed by ED/hospital consultants (20, 36%), CHP optometrist referrals (10, 18%), primary care physicians (3, 6.0%), and self-referrals (1, 1%).

Discussion

This study demonstrates the effectiveness of a grant for PEH in the field of ophthalmology at a tertiary ophthalmology clinic. PEH enrolled in the Harris Health Homeless Grant program achieved sustained ophthalmology follow-up over one-year, with a compliance rate of 51%. To our knowledge, this is the first study to evaluate ophthalmology follow-up and clinical outcomes for PEH in a tertiary center.

Compliance with medical care among PEH is thought to be low, particularly for specialty services. Prior studies have shown that only 32–58% of patients referred after community screenings completed their initial appointment at a ophthalmology tertiary center (9-11). Long-term follow-up data is largely unavailable. Therefore, a one-year compliance rate of 51% is encouraging, especially considering the barriers this population faces in accessing care.

Based on multivariable logistic regression, each additional clinic visit increases the odds of compliance by 60% (Table 4). This is likely to reflect that patients with more advanced pathology, requiring intervention and frequent follow-ups, are more motivated to remain engaged with care. Referral source was also associated with compliance. Those referred by optometry had higher follow-up rates than those presenting via ED, suggesting prior establishment in outpatient care facilitates continuity.

Although it is not possible to determine whether ED patients had grant coverage at presentation, care coordinators report that the majority of PEH in the ED are uninsured. These patients likely begin the process of obtaining the grant while in the ED or during inpatient admissions, and complete the process as outpatients. Thus, there may be two subgroups of PEH using the grants. The first are patients already established in the clinic or referred to the clinic by outpatient sources who maintain follow-up. The second are patients presenting via the ED with acute episodes, who are less likely to follow-up. Access to insurance, while important, is not the only barrier for PEH. Previous studies have shown the difficulty in getting PEH to present for initial appointments, even when insured (7,11,16). One study demonstrated that follow-up rates to a tertiary center clinic improved when patients received both counseling and transportation support (16). A similar outreach during follow-up clinic visits could potentially improve compliance in this population.

Thirteen patients used the grant exclusively for testing appointments, arranged through CHP. Although excluded from analysis, these encounters demonstrate that some PEH are willing to travel for diagnostic services without direct provider contact, underscoring that when structural barriers are reduced, many PEH are motivated to engage in their eye health.

The common diagnosis in the study corresponded well to the non-RE ocular pathology observed in PEH (1). The overall change in VA between the initial and final visit was not statistically significant, largely due to the presence of advanced, irreversible disease like proliferative diabetic retinopathy and glaucoma (Tables 6,7). However, patients who underwent ophthalmic interventions achieved an average of 0.30 logMAR (equivalent to 3-line) improvement in VA. This suggests that in a subset of PEH, appropriate intervention can make a visual improvement, which is significant in this population (10). Future studies need to see if this trend continues in larger population size.

Referrals were largely appropriate, as only 18% of patients were discharged from the clinic after their initial evaluation. The remaining patients demonstrated a high burden of ocular pathology, since milder conditions such as early glaucoma, mild to moderate non-proliferative diabetic retinopathy without macular edema, and early or moderate age-related macular degeneration were referred back to CHP for future management.

This study demonstrates the feasibility of managing complex medical and surgical pathology including non-cataract ocular disease in PEH. Previous studies have described successful surgeries in isolated cases but lacked longitudinal data (7,10). Additionally, the high number of retina patients in this study is unique. A prior retina-focused study referred patients without documentation of the treatments received afterwards (7). Our cohort included patients undergoing advanced procedures including one patient who underwent a tap and multiple injections with antibiotics for endophthalmitis in one eye, and later received anti-vascular endothelial growth factor (VEGF) injections for macular edema in the fellow eye. Another patient with a history of anterior uveitis underwent two corneal transplants, followed by a repair for an open-globe and corneal transplant dehiscence secondary to trauma. Multiple patients underwent combined cataract and pars-plana vitrectomy for advanced tractional retinal detachment.

Often a multidisciplinary team worked creatively to accommodate patient needs. One inpatient PEH with severe bilateral cataracts had very poor vision. He received cataract surgery in one eye before hospital discharge, due to the coordination between the primary medical team and ophthalmology. The patient later returned for surgery in the fellow eye, ultimately improving from count fingers to 20/20 in both eyes.

A major strength of the study is the strict inclusion and exclusion criteria, ensuring that only PEH patients were included in the study. Documentations of missed and upcoming appointments helped accurately assess compliance.

There are also limitations to this study. There is possible selection bias, as individuals utilizing the Homeless Grant may differ from the larger PEH population. Documentations needed for the application include official identification, a letter from a shelter or agency dated within 60 days indicating that the person is a PEH, current pay stubs, child support documentations, tax returns (if self-employed), proof of identification for dependents, immigration status for the patient and dependents and, a Medicare Asset Form. After this there is an interview that is completed on a face-to-face basis. Although there are social workers that can help with the application, given the extensive documentation needed to qualify, only individuals that are motivated and can navigate the system are able to successfully receive the grant. This grant also excludes portions of the PEH population like unsheltered PEH, who have been documented to have trouble following-up in clinic and utilize the ED for medical issues (17,18).

Other limitations of the study include the retrospective design, small sample size, and potential under capture of uninsured PEH and PEH using a different insurance plan. Therefore, this data concerning compliance and intervention may not be generalizable to all PEH.

Conclusions

PEH enrolled through the Homeless Grant in Harris Health demonstrated meaningful compliance with long-term ophthalmology follow-up across one year. A subgroup of patients experienced significant visual gains after undergoing procedures. These results suggest that targeted, structured funding and referral programs can improve access and outcomes for PEH requiring specialty eye care.

Acknowledgments

The authors would like to acknowledge Miranda Pruitt for helping create Figure 1 based on information provided by the authors.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://aes.amegroups.com/article/view/10.21037/aes-2026-1-0003/rc

Data Sharing Statement: Available at https://aes.amegroups.com/article/view/10.21037/aes-2026-1-0003/dss

Peer Review File: Available at https://aes.amegroups.com/article/view/10.21037/aes-2026-1-0003/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://aes.amegroups.com/article/view/10.21037/aes-2026-1-0003/coif). J.C. received consulting fees and payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing or educational events from Novartis, Allergan, OD-OS, Erasca, B&L, Iveric Bio, Eyepoint, Neurotech, Astellas, Genentech, AbbVie and Ocugen Salutaris. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the institutional ethics board approval of Baylor College of Medicine (ESP1: H-57931), and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Sayal AP, Slomovic J, Bhambra NS, et al. Visual impairment and the prevalence of ocular pathology in homeless children and adults globally: a systematic review. Can J Ophthalmol 2021;56:158-65. [Crossref] [PubMed]
2. Bernhisel A, Hekzco JB, Stagg B, et al. Comparing Eye Care at a Tertiary Eye Care Clinic to a Free Homeless Clinic. Invest Ophthalmol Vis Sci 2019;60:4460.
3. Miller J, Samanta A, Nelson P, et al. The SMART Study [The Street Medicine Auto-Refractive Technology Study]. Baltimore; 2023 [cited 2025 Jul 20]. Available online: https://nhchc.org/wp-content/uploads/2023/06/NHCH-Conference-2023_-SMART-Study-1.pdf
4. Kimmel A, Qin G, Braun R, et al. Barriers to Healthcare Access: Insights from a Survey of Persons Experiencing Homelessness [Internet]. [Cincinnati]: University of Cincinnati; 2025 [cited 2025 Jul 30]. Available online: http://rave.ohiolink.edu/etdc/view?acc_num=ucin1747666219622721%0Ahttps://rave.ohiolink.edu/etdc/view?acc_num=ucin1747666219622721
5. Liu JM, Chang M, Liu FM, et al. Perceptions of providers and unhoused patients on access to eye care in Santa Clara County: A qualitative study. Optom Vis Sci 2025;102:464-71. [Crossref] [PubMed]
6. Vinnett A, Tariq Z, Alvarez JA, et al. Eye Disease Prevalence and VF-14 Validation Among Patients Experiencing Homelessness and Presenting for Ophthalmic Examination in Baltimore, Maryland. Transl Vis Sci Technol 2023;12:7. [Crossref] [PubMed]
7. Barnes JB, Barnes SS, Small CR, et al. Mobile eye screenings for Hawaii’s homeless: results and applications. Clin Optom 2010;2010:73-7.
8. Campbell C, Samanta A, Reppa C. Different Models of Ophthalmology Care for Unhoused Patients. Glob Ophthalmol Summit [Internet]. Portland; 2024. Available online: https://www.globalophthalmologysummit.org/2024-eposters?wix-vod-video-id=e4dc575346194719ab3dd2053f7bf817&wix-vod-comp-id=comp-m0qxjpe14
9. Russo D, Bass O. Visual and Ocular Conditions among a Homeless Adult Population of Boston. Optom Vis Sci 2021;98:362-6. [Crossref] [PubMed]
10. Samanta A, Miller J, Jacobs M, et al. The Street Medicine Auto-refraction Technology Study: A Model for Screening and Treating Refractive Error in Persons Experiencing Homelessness. Am J Ophthalmol 2025;270:183-91. [Crossref] [PubMed]
11. Hennein L, Spaulding KA, Karlegan V, et al. Follow-Up Rates at a Free Ophthalmology Clinic at a Homeless Shelter. J Acad Ophthalmol (2017) 2021;13:e51-6. [Crossref] [PubMed]
12. Slomovic J, Hanna V, Chaban Y, et al. Delivering eye care to homeless and marginally housed populations during the COVID-19 pandemic: a pilot study. Can J Ophthalmol 2023;58:136-42. [Crossref] [PubMed]
13. Houston State of Health Data Portal [Internet]. 2025 [cited 2026 Apr 21]. Available online: https://www.houstonstateofhealth.com/indicators/index/view?indicatorId=90&localeId=2675
14. Health Insurance at HCH Programs, 2023. 2024 [cited 2026 Apr 21]; Available online: www.nhchc.org
15. Harris Health System. Health Care for The Homeless. 2023.
16. Hennein L, de Alba Campomanes AG. Association of a Health Coaching and Transportation Assistance Intervention at a Free Ophthalmology Homeless Shelter Clinic With Follow-up Rates. JAMA Ophthalmol 2021;139:311-6. [Crossref] [PubMed]
17. Withers J. Street medicine: an example of reality-based health care. J Health Care Poor Underserved 2011;22:1-4. [Crossref] [PubMed]
18. Frankeberger J, Gagnon K, Withers J, et al. Harm Reduction Principles in a Street Medicine Program: A Qualitative Study. Cult Med Psychiatry 2023;47:1005-21. [Crossref] [PubMed]