BACKGROUND Eye diseases worldwide, including within the United States, are underdiagnosed and undertreated1. A multitude of factors contribute to this deficiency in eye care including, but not limited to, availability of specialists, transportation and mobility barriers, financial burden, lack of education, and poor patient-physician communication and understanding2,3,4. Teleophthalmology, a paradigm of care delivery in which ocular images are interpreted remotely by an eye specialist, has increased in interest since the COVID-19 pandemic, may offer improved access to necessary eye care5. The need for improved access through teleophthalmology is particularly critical for diabetic retinopathy (DR), the leading cause of new cases of blindness among adults aged 20 to 60 affecting more than 100 million patients worldwide6,7. DR arises when elevated levels of blood sugar resulting from either type 1 or 2 diabetes mellitus damage the blood vessels that supply oxygen and nutrients to the retina, the light-sensing part of the eye. The risk of developing DR is directly related to the length of time that a patient has diabetes and usually does not appear for approximately five years after a type 1 diabetes diagnosis, although it may already be present when type 2 diabetes is diagnosed8. In the absence of glycemic control and/or ophthalmic treatment, the disease may progress through three stages of non-proliferative retinopathy (mild, moderate, severe) before proliferative retinopathy develops. Diabetic macular edema can occur with any stage of retinopathy. If DR is diagnosed early, vision loss may be mitigated or prevented9. An annual fundus examination to screen for DR is critical, however, only about half of all patients with diabetes receive proper screening and less than 40% of patients with a high risk of vision loss ever undergo treatment10,11. In 2010, primary care providers (PCPs) delivered clinical care to approximately 90% of individuals with type 2 diabetes, and the proportion has increased over time12. The importance of primary care practitioners ensuring that their diabetic patients receive recommended eye care is reflected in the Healthcare Effectiveness Data and Information Set (HEDIS). This comprehensive set of quality performance measures across six domains of care guide the primary care of chronic medical conditions like diabetes mellitus and includes assessment of whether a diabetic patient receives diabetic eye screening at least every two years13. Attainment of these quality measures is increasingly important for health-system quality ratings and value-based reimbursement models. Practices are increasingly turning to teleophthalmology programs to aid in this goal of care5,14. Traditionally, DR is diagnosed by an eye specialist via an annual in-person fundoscopic examination. However, with appropriate training, non-ophthalmic clinicians and clinical personnel are able to use a fundus camera to take retinal photos that can then be evaluated by an ophthalmologist typically via a store-and-forward model. DR can be determined with high sensitivity and specificity from fundus photography, and a referral for further ophthalmic evaluation or treatment is made for those patients with retinopathy15. Primary care-based teleophthalmology programs have improved the accessibility and cost-effectiveness of DR screening in both rural and urban settings worldwide and are currently being applied to DR screening more commonly than any other ocular pathology16, 17,18. The ongoing COVID-19 pandemic has exacerbated existing barriers and increased the likelihood of ophthalmic appointment postponement or cancellation rendering teleophthalmology services even more critical to DR screening programs19, 20. The prevalence of diabetes in California is more than 40% above the United States national average21. As a means to improve the ophthalmic health of our patients, the Stanford Teleophthalmology Automated Testing and Universal Screening (STATUS) program was developed as a multi-site teleophthalmology DR screening collaboration between the Byers Eye Institute of Stanford (BEIS) and five affiliated primary care clinics throughout the San Francisco Bay Area. The program was initiated two to six months (depending on the site) prior to the onset of the COVID-19 pandemic in the United States and continued to provide remote eye examinations to patients throughout 2020 and 2021. The goal of the program was to evaluate whether the use of teleophthalmology could increase the percentage of patients screened for DR in collaboration with regional primary care clinics. Here, we examine the ability of the 18-month teleophthalmology program to improve and maintain access to DR eye care prior to and during the COVID-19 pandemic. METHODS Clinic Sites Non-mydriatic fundus cameras were deployed at an academic-affiliated primary care site in Santa Clara, CA in September 2019, and in four additional affiliated primary care sites in Los Gatos, Oakland, Hayward, and Pleasanton, CA beginning in February 2020. The primary care sites ranged from 20 miles (25-minute drive) to 42 miles (45-minute drive) away from the BEIS. Store-and-forward teleophthalmology screening for diabetic retinopathy continued at all five locations throughout the study period which ended April 2021. In order to determine whether the teleophthalmology program impacted the adherence rate to annual diabetic eye exams, HEDIS measures at two primary care sites (Pinole, CA and San Pablo, CA) in the same healthcare system that did not deploy the teleophthalmology system were also assessed. The study was approved by the Institutional Review Board at Stanford University. Patient Image Collection and Assessment Patients 18 years or older with type 1 or type 2 diabetes mellitus without a prior DR diagnosis or a DR exam in the past 12 months were offered the opportunity to have fundus photographs taken at the end of their primary care visit. Fundus imaging was performed by a trained medical assistant using the CenterVue DRS fundus camera (Hillrom Inc., Chicago, IL) at the Santa Clara primary clinic site and the TopCon NW400 fundus camera (Welch Allyn Inc., Skaneateles Falls, NY) at the Los Gatos, Oakland, Hayward, and Pleasanton primary care clinics. If medical assistants deemed the image quality to be poor, they repeated image acquisition and did so up to 4 times. The fundus images were forwarded to vitreoretinal specialists at BEIS who evaluated the images within one week. These fundus images were classified as ungradable (such as when opacity, blurring, or decentration impaired visualization of the fundus), or gradable if quality was sufficient for grading of DR. Images of adequate quality had a DR grade assigned in accordance with the International Clinical Diabetic Retinopathy Disease Severity Scale with moderate and severe categories combined on teleophthalmology evaluation (no diabetic retinopathy/mild non-proliferative diabetic retinopathy/moderate to severe non-proliferative diabetic retinopathy/proliferative diabetic retinopathy)22. Patient images were also assessed for the presence of macular edema or other fundus abnormalities. Patients with images of insufficient quality from one or both eyes were recommended to have the images retaken or present for an in-person eye examination. Diagnosis and stage of DR was determined by the eye with more advanced retinopathy. Those with referral-warranted disease were referred for an in-person exam at BEIS or their local ophthalmologist. A subset of patients (N=26) voluntarily presented for a second teleophthalmology screening one year after their first examination. Patient Data Patient files containing information on labs, orders, clinical notes, and patient information were retrieved from The STAnford Research Repository (STARR), an institutional resource for working with clinical data for research purposes. Data was managed and analyzed using Python (version 3.9.0) with Pandas (version 1.3.0). Patients who underwent fundus imaging without a documented assessment by BEIS specialists were excluded (N = 23). For all patients who were seen at BEIS after a referral for in-person examination, data was manually collected from the electronic health record. For analyses comparing patients prior to and during the COVID-19 pandemic, March 16th, 2020, was used as the start of the pandemic since on that date legal stay-at-home orders were announced in Alameda, Contra Costa, Marin, San Francisco, San Mateo, and Santa Clara counties. Longitudinal HEDIS data were only available for three of the teleophthalmology primary care sites and the two non-teleophthalmology comparison sites; two teleophthalmology primary care sites did not have structured HEDIS data available for analysis.

The authors have declared no competing interest.

The study was funded by an unrestricted grant from Research to Prevent Blindness and a P30 NIH grant to the Byers Eye Institute of Stanford University.

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Ethics committee/IRB of Stanford University School of Medicine gave ethical approval for this work.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

All data produced in the present study are available upon reasonable request to the authors.