Follow-Up Compliance for Patients Diagnosed with Diabetic Retinopathy After Teleretinal Imaging in Primary Care

Introduction

Diabetes is an epidemic affecting ∼10% of the U.S. population with 1.5 million new diagnoses each year. ¹ It is associated with many complications, including diabetic retinopathy (DR), which is the leading cause of blindness among working-age adults in the United States. ² Although DR is one the most preventable blinding diseases, ³ <60% of individuals with diabetes receive an annual eye examination. ⁴ Teleretinal imaging in the primary care space has been shown to increase patient compliance with the annual diabetic retinal examination (DRE). ^5,6 Specialized nonmydriatic retinal cameras are used in primary care to capture retinal images of patients during routine appointments. The images are uploaded for remote interpretation by eye specialists, with the diagnosis and recommended management plan returned to the primary care practitioner.

Although teleretinal imaging programs have been reported to increase compliance rates to ∼90%, ^5,6 little evidence currently exists to link increased examination compliance with a reduction in diabetic blindness in the United States. Other countries, most notably the United Kingdom, have successfully reduced diabetic blindness rates through teleretinal evaluation and efficient referral programs. ⁷ However, within the United States, the independence of electronic medical records (EMRs) and siloed delivery between primary care and eye care professionals makes the collection and analysis of teleretinal program efficacy data difficult.

The goal of this study was to determine the rates of follow-up compliance for patients examined through a teleretinal imaging program in a single large primary care network. The reasons for noncompliance and the correlation between primary care screening results and eye specialist diagnosis were assessed.

Materials and Methods

Summit Medical Group (SMG) is a large primary care group in eastern Tennessee comprising 300 providers in 66 office locations across 13 different counties. SMG had extremely low retinal examination compliance rates before 2015 and sought to improve and integrate the care they were providing for their patients with diabetes. SMG implemented the RetinaVue^® care delivery model (Welch Allyn^®, Skaneateles Falls, NY) to provide teleretinal diabetic eye examinations.

Designated clinical team members at SMG primary care clinics use the RetinaVue 100 Imager (Welch Allyn), a handheld retinal camera, to capture nonmydriatic 45° macula-centered retinal images on patients with diabetes who have not had a comprehensive eye examination within the past 12 months. The camera is designed for use by a minimally trained operator and includes features of auto-capture and autofocus, and an image quality indicator, which enhances its use and adoption in the primary care environment. After capture, the image is uploaded for review and remote interpretation by an ophthalmologist using Health Insurance Portability and Accountability Act (HIPAA)-compliant software. The ophthalmologist identifies any pathology present and a follow-up recommendation and management plan for the patient is provided to the referring practitioner. The electronic report is deposited directly in the patient's EMR.

The RetinaVue care delivery model uses the International Clinical Diabetic Retinopathy Disease Severity Scale (ICDRS), which is recommended by the American Academy of Ophthalmology. ⁸ The recommendation for patients without DR or with only mild nonproliferative DR is to have follow-up imaging in 12 months. Patients with moderate nonproliferative DR are recommended to follow up with an ophthalmologist within the next 6 months. Patients with evidence of severe nonproliferative DR, proliferative DR (previously treated or untreated), or any retinopathy with clinically significant macular edema (CSME) are considered to have vision-threatening diabetic retinopathy (VTDR) and are referred for a follow-up, in-person examination with a local ophthalmologist.

SMG receives the RetinaVue report with the referral recommendation within 24 hours of the patient's examination. The report is reviewed by the patient's primary care physician and the physician or clinical designee notifies the patient of a required follow-up appointment, if recommended. Patients who do not schedule or attend ophthalmology follow-up appointments are often contacted again by SMG by phone. Multiple reminder calls may be made to encourage patients to attend their follow-up appointments.

The central institutional review board IntegReview IRB of Austin, TX, approved the study protocol, which complied with the tenets of the Declaration of Helsinki. Informed consent was waived due to the retrospective nature of the study.

We performed a RetinaVue Network database search to identify all SMG patients evaluated for DR between June 2015 and October 2017. The medical record number of each patient was obtained from the RetinaVue diagnostic report and used to search the SMG EMR. The medical record for each patient was searched for the presence of ophthalmology notes to document a follow-up appointment examination and for any retinal pathology identified. In cases where there was no documentation of follow-up in the primary care medical record, SMG clinical staff reached out to local eye care providers in an attempt to obtain any missing records. If records could not be obtained, the medical record was searched to identify reasons for noncompliance.

Results

A total of 3,360 diabetic patients were examined at SMG using the RetinaVue care delivery model during the 2.5-year study period. During that time, 110 (3.3%) patients were referred for follow-up eye examinations for VTDR. Ninety-four of 110 (85.5%) patients were Type 2 diabetics and 64 of 110 (58.2%) were male. The average age of referred patients was 60 (range 26–85). Of those referred, 27 were referred for proliferative DR, 12 were referred for severe nonproliferative DR, 26 were referred for moderate nonproliferative DR with CSME, and 45 were referred for mild nonproliferative DR with CSME, as detailed in Table 1. There were an additional 264 patients with mild nonproliferative DR and 32 patients with moderate nonproliferative DR, reflecting an 8.8% prevalence of nonreferable eye disease in the patients evaluated. Sixty-two of 110 (56.3%) patients referred attended follow-up appointments within 3 months and 83 of 110 (75.5%) patients referred attended follow-up appointments within 1 year of referral.

There were 27 patients who were referred but did not have documentation of a follow-up appointment within 1 year of the teleretinal examination. Of these patients, 9 (33.3%) followed up after 1 year, 4 (14.8%) were deceased or left the practice, 4 (14.8%) had a subsequent teleretinal examination that no longer documented referable DR, and 3 (11.1%) had a significant acute health issue that impacted follow-up (e.g., hospitalization). No specific reason was determined in 7 (25.9%) of the referred patients who did not obtain an eye examination.

Each patient's diagnosis from the follow-up comprehensive eye examination was compared with the initial diagnosis by teleretinal imaging. Of the 83 referrals who attended follow-up within 1 year, 52 (62.7%) had VTDR by ICDRS criteria, 22 (26.5%) had DR that was not vision threatening, 5 (6.0%) did not confirm DR, 4 (4.8%) had other retinal disease findings or no diagnosis documented. It is important to note that ophthalmologists in community practices may follow different conventions for assessing and grading DR, particularly within the mild and moderate levels. RetinaVue, P.C. ophthalmologists strictly grade per the ICDRS, which may have led to some of the discrepancies in grading between the remote interpretation and the follow-up assessments. Correlation between the teleretinal and follow-up diagnosis is detailed in Table 2.

Discussion

SMG patients receiving teleretinal imaging through the RetinaVue care delivery model had a moderate rate of compliance with follow-up examination attendance within 3 months (56.3%) and a high rate of compliance within 1 year (75.5%). This rate is higher than that noted by similar studies, which range from 32.8% to 60.0% ^{9 –13} (Table 3).

Several studies have examined patient follow-up rates for up to 2 years after the referral appointment. ^10,12,13 We chose to limit our follow-up compliance to 1 year, as patients who did not follow up during this period would be indicated to have a subsequent annual teleretinal examination. Proliferative DR and CSME often require prompt treatment to prevent permanent vision loss. A 2-year follow-up period may be relevant for research purposes, but patients with VTDR may have significant vision loss if allowed 2 years to attend a follow-up visit. In our study, there were 9 patients who followed up after 1 year but within 2 years, which increased overall follow-up examination compliance to 83.6%.

Other publications assessing rates of follow-up conducted in the southeastern United States focused on health disparity patient groups, including rural and underserved communities, and veterans. These populations may lack resources such as health insurance and transportation, making them more likely to miss follow-up appointments. SMG's relatively high compliance rate may be attributed to higher health literacy within the population, a more rigorous and standardized referral process within the practice, the availability of local eye specialist practices, or other factors not determined in our study.

Keenum et al. found that attendance at follow-up examinations correlated with increased age, knowing one's hemoglobin A1C level, and obtaining assistance in setting up the referral appointment. ¹⁰ Chasan et al. found that patients with a history of missing scheduled outpatient visits had a much greater chance of noncompliance with referral appointments as well. ¹³

In our study, we could determine reasons for noncompliance in 20 of 27 (74.1%) patients who did not attend a referral appointment. Seven patients were lost to follow-up for reasons that could not be determined from the medical record. Some noncompliant patients did have their disease monitored and managed by repeat telemedical imaging at SMG; however, this took place outside of the recommended time window, putting the patients at risk for visual impairment. Three of the patients with severe diabetic eye disease had comorbidities, including end-stage renal disease, cancer, and acute cardiac issues. These conditions may have increased the risk of noncompliance due to in-patient hospital admission. Patients with comorbidities may have difficulty prioritizing, scheduling, and attending follow-up eye care, so it is critical that primary care practices track these patients to assure recommended follow-up examinations.

Even with higher rates of follow-up, documentation of eye specialist examinations in the SMG medical record was often incomplete. Eye care specialists frequently sent an initial follow-up visit report but did not send further records of ongoing care, treatment provided, or visual outcomes. In several cases, patients were referred for another eye care service (e.g., cataract extraction) while the patient was undergoing treatment for DR. In specific cases, documentation of the other procedure was the only evidence of follow-up for DR in the primary care record.

When eye care specialist follow-up was documented in SMG's medical record, the follow-up diagnosis closely matched the recommendation from teleretinal screening. More than 60% of patients were confirmed to have VTDR and ∼90% of patients examined had some level of DR. These findings indicate that it is critical for primary care providers to ensure that each referred patient follows up with an eye care specialist, as the patient more likely than not requires frequent monitoring or treatment of retinal disease.

The limitations to this study were that it was restricted to one practice group and included a limited number of patient referrals for analysis. The eye care specialist follow-up information in the primary care record was, in many cases, limited; thus, actual compliance might have been higher but was missed due to incomplete documentation. As noted earlier, the compliance with follow-up appointments evaluated in this study may not be replicable in a patient population with lower socioeconomic status or with limited access to eye care providers. Finally, this study was not able to report on long-term treatment and visual outcomes, and only assessed whether patients who needed referral for specialist eye care attended their initial recommended follow-up appointment.

Conclusions

This study demonstrated that most patients evaluated for DR and referred from a telemedical care delivery model in a large primary care practice attended their follow-up appointments with an eye specialist. However, importantly, almost 25% of patients did not have the recommended eye examination within 1 year of the diagnosis of VTDR. Review of the SMG EMR identified reasons for missing referral appointments in majority of these patients. Eye specialist diagnosis confirmed VTDR in >60% of the patients and some level of DR in ∼90% of those referred.

Primary care practices providing telemedical care for DR evaluation represent a large and growing patient-centered health care delivery model that has the potential to improve DRE rates and visual outcomes for the large and growing population of patients with diabetes in the United States. To date, these efforts have been driven by individual practice decisions and by the requirement for large managed care health plans (e.g., Medicare Advantage plans) to meet Healthcare Effectiveness Data and Information Set (HEDIS^®) metrics for their quality ratings. However, as camera technology improves, providing less expensive and scalable methods for DR evaluation, more patients with vision-threatening disease will be identified in primary care practices and telemedical networks.

Our study and others demonstrate the need to implement EMR systems to track patients through follow-up and long-term management and to provide patients and referring physicians with reminders if follow-up does not occur or is delayed. Collaboration and communication between the primary care provider and the eye care specialist during and after the referral process into the eye care network for treatment will be extremely important and help to ensure that patients with diabetes are receiving the appropriate care. If providers are made aware of required eye treatments, they can help to reinforce patient compliance during routine primary care visits. Strong collaboration between provider networks will be required to improve quality patient management and reduce long-term visual impairment from diabetic eye disease as telemedical diagnosis and management becomes the norm and ultimately the standard of care, as it is already in many countries around the world.