Participants reported their date of birth, sex, race/ethnicity, education level, health insurance type, employment status, marital status, and preferred transportation methods to medical appointments.

Medical and ocular conditions, prescription eyeglasses use, and family history of glaucoma and blindness were documented.

Participants were asked if they ever had a dilated eye exam and if so, approximately when was their last dilated eye exam and their transportation method to this appointment. Participants were also asked if they are currently under the care of an eye doctor and would they like to follow-up with him or her if referred to ophthalmology.

Eye health screenings were conducted in community rooms at NYCHA developments or DFTA senior centers. Visual acuity was measured by trained community health workers using the Snellen eye chart at 20 feet with correction.^16,20 Trained ocular technicians measured intraocular pressure (IOP) in both eyes using the Ic100 iCare rebound tonometer (iCare, Helsinki, Finland).¹⁶ In a darkened space, 2 nondilated fundus images of the optic disc and macula of each eye were taken by a trained ocular photographer using an autofocus, hand-held fundus camera (Volk, Pictor Prestige, Mentor, Ohio, USA). Within 24 hours, these images were uploaded by the ocular photographer to a HIPAA-compliant cloud-based database (Forum, Zeiss Oberkochen, Germany). The images were read and graded by 1 study ophthalmologist specializing in retina and 1 specializing in glaucoma.²¹ Final reading results were based on the worse eye and classified as normal or abnormal with no significant findings, abnormal with significant findings, or unreadable.

Those with visual acuity 20/40 or worse in either eye, IOP 23–29 mmHg, or an unreadable image “failed” the eye health screening and were scheduled to see the study optometrist within 3 weeks in the same community-based location. Participants with an IOP ≥30 mmHg were “fast-tracked” and immediately scheduled with ophthalmology for an in-office glaucoma evaluation within 2 weeks. Those with an abnormal fundus image with significant findings were also referred directly to ophthalmology.

For those who failed the screening and attended the optometric exam, an ocular technician performed an autorefraction (QuickSee Autorefractor, Plenoptika, Cambridge, MA, USA) to assess for refractive error. The optometrist performed a Trial Lens refraction and conducted a nondilated anterior segment exam using a portable slit lamp and a posterior segment exam using a direct ophthalmoscope (Welch Allyn Panoptic 3.5V, Skaneateles Falls, NY) to detect ocular abnormalities. The optic nerve exam for diagnosing glaucoma and glaucoma suspect was based on the American Academy of Ophthalmology (AAO) Preferred Practice Pattern Guidelines^®.²² The optometrist also diagnosed cataracts, retinal abnormalities, and other ocular conditions and referred participants to ophthalmology for an in-office dilated eye exam and further ocular testing/imaging as deemed necessary, and the “reasons for referral” results have been previously published.^23–24

All participants referred to ophthalmology for an in-office eye exam, either due to an IOP ≥30 mmHg, an abnormal fundus image finding, or ocular conditions confirmed by the optometrist, were scheduled for an initial in-office eye exam by the study staff within 6 months (+4 months). All eye exam dates were documented in the Research Electronic Data Capture (REDCap) and study staff reviewed EPIC electronic medical records (EMR) each time a participant was scheduled in-order to document appointment adherence over 10 months.

Participants in the Navigator Intervention group received additional telephone-based support over a 10-month period from trained patient navigators, specifically with scheduling, rescheduling, and eye exam appointment reminders, and if requested, transportation assistance was provided. Navigator Intervention participants who did not attend the initial eye exam were contacted by the patient navigators and rescheduled. Reasons for not attending were captured in the REDCap call log to assess barriers to adherence.

Participants in the Usual Care group referred to ophthalmology were only called to share abnormal image findings and schedule the initial in-office eye exam; they did not receive ongoing eye exam scheduling or transportation support after the initial appointment. Usual Care participants were not contacted if they missed the in-office eye exam.

Attendance at the initial in-office eye exam is the primary outcome of the trial.¹⁶ Participants were considered adherent if they attended the recommended initial in-office eye exam within 10 months of the community-based screenings since appointments were scheduled within 6 months after screenings, and an additional 4-month window was permitted in case of rescheduling. Participants who chose to receive follow-up eye care from their own ophthalmologist and did not sign a medical release form were excluded from the analysis.

Participants who attended the initial in-office eye exam were recommended to follow-up at a variety of intervals depending on their ocular condition(s), including within 4 weeks, 1–2 months, 3–5 months, 6–11 months, or 12 months. Ocular diagnoses were confirmed, and treatment recommendations were captured from the EMR and documented in REDCap by study staff for statistical analysis.

Study data were collected and managed using the REDCap database hosted at Columbia University Mailman School of Public Health and are summarized using means and standard deviations for continuous variables and frequencies, and percentages for categorical variables. Data analysis was performed in IBM SPSS Version 28 (IBM Corp. Armonk, NY).²⁵

Primary outcome measure was adherence to the initial in-office eye exam. The randomization groups (Navigator Intervention vs. Usual Care) were compared in terms of demographics and SDOH variables (ethnicity, race, age, sex, employment, education level, marital status, insurance type, transportation needs). Predictors of adherence variables included randomization group, demographic characteristics, SDOH, self-reported medical and ocular history, and eye health screening results.

A 2-sample t-test was used to compare mean age between those who attended and those who did not attend the initial in-office eye exam. For categorical variables, chi-square tests were conducted to assess differences in adherence rates. Variables that demonstrated statistical significance were then included in a multivariate stepwise logistic regression model to assess their association with adherence, while adjusting for other variables that were also significant. The stepwise logistic regression model was used because a large number of variables were significant in the univariate analyses. At each step, variables were added based on the alpha-to-enter significance level of 0.05 and the alpha-to-remove significance level was set at 0.1 to exclude variables in the final model. For all analyses, P values of < .05 were considered statistically significant and all tests performed were 2-sided. Odds ratios (OR) were estimated with 95% confidence intervals (CI).

A cost analysis was conducted alongside the study to identify staff time and resources of delivering the eye health screenings and optometric exams in a real-world community setting. The cost analysis methods have been previously published and are described in detail.¹⁷

Cost measures included staff time costs for the on-site eye health screenings, equipment, supplies, optometrist eye exams, personal protective equipment, and travel costs to the sites. Staff time costs were calculated using the Activity-Based Costing method, a micro-costing approach which quantifies the monetary value of personnel required to deliver the screening based on the time and resources required.²⁶ During the study, staff logged time spent on implementation tasks in minutes. Since this recorded time did not include time spent managing the program and overhead (indirect costs), we multiplied their recorded time by a factor of 3 to arrive at a more realistic estimate. All time costs were converted to dollars by multiplying time spent performing each task by hourly wage rates, obtained from 2021 US Bureau of Labor Statistics (BLS) data for the specific types of staff involved (ophthalmic medical technician $19.77/hour, optometrist $60.31/hour, and ophthalmologist $129.85/hour).²⁷ The ophthalmologist time was required for image reading only, since they were not otherwise involved in delivering the intervention. All staff time costs were then inflated by 30.4% to account for fringe benefits. Costs for equipment (without depreciation value), travel, and supplies were obtained from financial accounting records. There were no rental costs for any of the spaces, therefore rental costs and depreciation were not included in the cost analysis. Since the goal of the cost analysis was to estimate the costs of delivering the eye health screening in a real-world community setting, we excluded all costs associated with research, including but not limited to, IRB approval and maintenance, research training, research data collection, investigator meetings, obtaining informed consent, research-related travel, and principal investigator time spent on study oversight.

We calculated the costs of delivering the Navigator Intervention by tracking the time patient navigators spent coordinating eye exam appointments and arranging transportation for participants referred to ophthalmology using a REDCap call log. Patient navigators and their approximate wage rate plus fringe benefits was costed using the same approach as previously summarized for screening personnel. For those referred to ophthalmology after the screenings, navigators and study staff logged start and end time in minutes using the REDCap call log during all contact with study participants. Time costs were calculated using the Activity-Based Costing method.²⁶ Since this recorded time did not include time spent managing the program and overhead (indirect costs), we multiplied their recorded time by a factor of 1.5 to arrive at a more realistic estimate. All time costs were converted to dollars by multiplying time spent performing each task by hourly wage rate for the patient navigators and study staff ($25.78/hour including fringe benefits), which was obtained from 2021 US Bureau of Labor Statistics (BLS) data.²⁷ Costs of patient navigators were compared to the staff time contacting Usual Care participants.

For those referred to ophthalmology, we calculated the cost per case for any detected eye disease and cost per participant who attended the in-office eye exam and compared costs between Navigator Intervention and Usual Care participants.