Tailored, institution-specific strategies that address demographic and socioeconomic factors are needed to bridge the gap in diabetic eye screening adherence. According to a large national cohort study, the problem is not solely due to patient factors. Instead, system-level failures appear to play a major role.

Researchers from the University of California San Francisco (UCSF) and UC Irvine conducted a retrospective study (2008-2023) of adults with proliferative diabetic retinopathy (PDR) to evaluate the diabetic eye disease care continuum in patients with type 2 diabetes (T2D). They tracked referrals, screenings, diagnoses, and treatments to identify predictors of adherence and assessed the potential impact of an automated referral system.

The study, published in Diabetes Care, included 8240 unscreened patients with T2D, of whom 43% received a referral, but only 16% completed screening within 1 year. Demographic, provider, and socioeconomic factors influenced adherence, but referrals emerged as the strongest predictor.

Using OpenAI’s GPT-4o, the investigators also analyzed 16,735 physician notes from 250 patients who were referred for screening but missed their visits. The artificial intelligence program identified the most common reasons for missed appointments as health complexities (16%), scheduling or referral issues (16%), care received elsewhere (14%), patient nonadherence (7%), and logistical or transportation challenges (5%). However, for 46% of patients, the notes contained insufficient detail to determine possible causes of missed screenings.

Identifying Logistical Barriers to Care

According to corresponding author Vivek Rudrapatna, MD, PHD, with the Bakar Computational Health Sciences Institute at UCSF, the findings suggest that streamlining referral processes could meaningfully improve adherence.

“Even when patients engage with their primary care teams, many never receive or complete referrals,” Rudrapatna told Medscape Medical News. “Improving adherence requires integrated and automated interventions. For primary care clinicians, endocrinologists, and eye specialists, this means partnering with health IT teams to automate EHR-based referral prompts and teleophthalmology screening directly into diabetes care workflows.”

He emphasized that eye clinics must also have adequate scheduling capacity to manage increased patient volumes.

“Primary care and local eye clinics must work together to ensure referrals and follow-ups are closed-loop, not passive handoffs,” he said.

Rudrapatna added that partnering with social workers and care coordinators can help address social determinants of health, such as health literacy and transportation, both of which affect adherence. Improving diabetic eye screening, he said, requires multilevel collaboration between primary care, endocrinology, and optometry or ophthalmology, supported by technology and tailored to the needs of local patient populations.

Persistent Vision Loss Linked to Noncompliance

The costs of noncompliance can be considerable. A separate study, also published in Diabetes Care, found that prolonged lapses in care for patients with PDR were common and disproportionately affected vulnerable groups. These lapses were also linked to persistent vision loss.

This retrospective national cohort study (2008-2023) analyzed data from 15,211 adults with PDR and ≥ 6 months of follow-up in the Sight Outcomes Research Collaborative. Overall, 71.8% experienced a lapse in care and 14.2% had lapses that lasted more than 24 months.

Lapses were more common among non-Hispanic Black patients, younger adults, and individuals with disabilities, and less frequent among those with poor vision or prior PDR treatment. Older age and previous PDR treatment predicted shorter lapses, whereas living in “distressed areas” predicted longer lapses. Visual acuity worsened after lapses, with greater deterioration following longer gaps in care.

Efforts to Improve Adherence

Kim Donaghue, MB, PhD, past president of the International Society for Pediatric and Adolescent Diabetes and professor at the University of Sydney in Sydney, Australia, said automated rescheduling systems could help reduce lapses in care once PDR is diagnosed.

Donaghue, who was not involved in either study, added that an education package explaining the vital importance of follow-up could also improve adherence.

“Certainly, flagging screening at diabetes consultations is important,” Donaghue told Medscape Medical News. “For this to be effective, though, there also needs to be an easy method of establishing that a referral was actually issued and received by the person with diabetes with an explanation of why it was issued.”

Rudrapatna reported receiving grant funding to UCSF from Genentech, Inc., Merck, Janssen, Takeda, Blueprint Medicines, Mitsubishi Tanabe, and Stryker; receiving payments for lectures from Natera and Ironwood; and having a financial stake in the companies ZebraMD, Acucare AI, and DataUnite.

Donaghue reported no relevant financial disclosures.