Implementing AI-based Glaucoma Screening Within Federally Qualified Health Centers

Overview

The goal of this clinical trial is to learn if implementation of an eye screening program at Federally Qualified Health Center (FQHC) clinics provides results that participants may have glaucoma, and/or other eye conditions (diabetic retinopathy, cataract, visual acuity impairment). The glaucoma screening will incorporate use of an artificial intelligence (AI)-assisted screening tool. This project is called AI-RONA. The main questions it aims to answer are: * How does this eye screening program compare to the rate of glaucoma and other eye conditions detected at other FQHC clinics where the screening program has not been implemented? * Do particpants who screen positive for these conditions adhere to the physician's recommendation for a follow-up examination by an optometrist or ophthalmologist? * Are referral rates for a follow-up comprehensive eye exam by an optometrist or ophthalmologist similar to those implemented by an ophthalmologist using telemedicine (that is, using the results of the screening to make a diagnosis remotely)? * What is the cost-effectiveness of the AI-assisted screening program in diagnosing glaucoma as compared to a physician-guided program? * Are participants completing the screening satisfied with it? * Are physicians at the FQHC clinics administering the screening satisfied with it? Participants will: * Undergo an ocular screening whose goal is to detect glaucoma, diabetic retinopathy, cataract, and/or impairment in visual acuity. If the screening indicates that participants may have these conditions, participants will be referred for a comprehensive eye examination by an optometrist or ophthalmologist. * Following the screening, participants and physicians will complete a survey on their satisfaction with the program.

Federally Qualified Health Center (FQHC) clinics are excellent candidate clinics for eye disease screening programs for adults, including glaucoma and also other glaucoma associated diseases (GAD), diabetic retinopathy, cataract and refractive error. These clinics are safety-net primary health care clinics in the United States designed to serve medically underserved areas and populations. The physician will provide services regardless of the patient's ability to pay, using a sliding-scale fee based on the ability to pay. Over half the patient population at FQHCs nationally is Black or Hispanic, with 82% of patients uninsured or federally insured. While FQHCs provide primary medical care in many health domains, one domain that is not adequately addressed by FQHC clinics is eye care. A recent report by the National Academies of Science, Engineering and Medicine indicated that primary eye care services (through optometrists or ophthalmologists) are rarely available at FQHC clinics; the report provided an estimate that less than 3% of FQHC patients actually receive vision care services at these clinics, representing 0.89% of FQHC clinic visits. The screening methods used in this protocol are a modified version of our prior imaging-based ophthalmologist guided telemedicine screening protocol utilized in the previous CDC- funded Alabama Screening and Intervention for Glaucoma and eye Health through Telemedicine (AL-SIGHT) study. In AL-SIGHT we used a telemedicine program to screen for glaucoma in 3 FQHCs located in a low-income, under-resourced region of the state with a high proportion of Black residents. Just as we did in AL-SIGHT, investigators in the current protocol will target screening only for those FQHC patients at-risk for glaucoma. The eligibility definition for screening in the current project is listed under Eligibility. After screening, patients with a screening diagnosis of GAD, diabetic retinopathy, cataract, or visual acuity impairment will be referred for an in-person follow-up exam by an ophthalmologist or optometrist located in the FQHC clinic's region. In the investigators' previous project, participants had excellent adherence to the follow-up appointment with the optometrist or ophthalmology at a rate of 76.7% of patients attending follow-up. Thus, Investigators are hopeful that in the current project they will also have this high level of adherence. There are two major differences between our previous project (AL-SIGHT) and the current project AI-RONA. In AL-SIGHT, we used research staff members to do the screening at the FQHCs. In AI-RONA investigators will now use implementation science to do the screening. This means that investigators will train the FQHC clinic staff to do the screening after a detailed training and certification program. A second way AI-RONA differs from AL-SIGHT is that investigators will use an AI-assisted, remote screening algorithm to generate the screening results for glaucoma. Investigators will also focus on both the effectiveness of the program (disease detection, rate of referrals) and primary care providers (PCPs) and patients' attitudes about AI-RONA. The latter will be achieved by administering a questionnaire to PCPs and to participants about their satisfaction with the screening program. The AI approaches to be deployed for AI-RONA have been previously developed and extensively tested by the AI-researchers at the University of California, San Diego, who are also investigators in AI-RONA. It was found that glaucoma testing using AI based on retinal images taken of patients' eyes lead to clinician-level accuracy. These AI methodologies have been tested on diverse populations and publicly available datasets to facilitate their accuracy. The eye screen protocol that investigators are using includes the following. All assessments are performed on each eye separately: (1) retinal imaging including fundus photography and optical coherence tomography (OCT), (2) visual acuity measurement, (3) refractive error measurement.