More than half of patients with photographic or exam-based evidence of diabetic retinopathy were unaware of their diagnosis in included US studies, and additional gaps persisted through screening, referral, and follow-up, according to a systematic review published in BMJ Open.
Researchers used a cascade-of-care framework to characterize diabetic retinopathy care in the US, from diabetes diagnosis to diabetic retinopathy–related visual impairment and blindness. The review included English-language US-based studies published from 2018 through 2025, with searches of Ovid MEDLINE and Ovid Embase conducted from September to November 2025 and updated from March to April 2026. Reference lists and Centers for Disease Control and Prevention publications were also reviewed.
The review included 46 sources: 27 cohort studies, 15 cross-sectional studies, 2 systematic reviews/meta-analyses, and 2 CDC records. Systematic reviews were used only to inform estimates of diabetic retinopathy–related blindness, where primary US data were limited.
The predefined cascade stages were diagnosis of diabetes, adherence to diabetic retinopathy screening, diagnosis of diabetic retinopathy, adherence to diabetic retinopathy care, and diabetic retinopathy–related blindness. Because of heterogeneity in study populations, outcome definitions, and methods, the researchers did not perform a meta-analysis. Two reviewers screened studies and assessed risk of bias using the Newcastle-Ottawa Scale and Risk of Bias in Systematic Reviews tool. The researchers reported that they did not prioritize studies by design, risk-of-bias assessment, or relevance when summarizing results.
Across included studies, 55% to 89% of patients with photographic evidence of diabetic retinopathy were unaware of their diagnosis, with a median estimate of about 70%. The researchers noted that the available studies could not determine whether lack of awareness reflected patients who were never examined by an eye care professional, were examined but not informed of their diagnosis, or were informed but did not understand that they had diabetic retinopathy.
Screening adherence was also inconsistent. Annual diabetic retinopathy screening adherence ranged from 3% to 69%, with a median of about 47%. Biennial screening adherence ranged from 16% to 79%, with a median of about 59%. The American Academy of Ophthalmology and the American Diabetes Association recommend at least annual diabetic retinopathy screening beginning at diagnosis for patients with type 2 diabetes and 5 years after diagnosis for patients with type 1 diabetes; longer screening intervals may be appropriate for selected patients with adequate glycemic control and no evidence of retinopathy on prior examinations.
Gaps continued following diagnosis. In primary care settings, studies reported that 31% to 63% of patients diagnosed with diabetic retinopathy were linked to ophthalmic care, with a median of about 52%. In emergency department settings, follow-up rates were 53% for nonproliferative diabetic retinopathy referral and 62% for proliferative diabetic retinopathy referral. Among studies evaluating diabetic retinopathy follow-up, 55% to 78% of patients had a lapse in recommended care.
The researchers’ quantitative cascade illustrated the scale of patient attrition across the care continuum. Starting from an estimated 38.1 million US adults with diabetes, about 28.4 million were aware of their diabetes diagnosis. About 16.9 million were estimated to adhere to biennial diabetic retinopathy screening, about 3 million were diagnosed with and aware of diabetic retinopathy, about 1.5 million were linked to diabetic retinopathy care in primary care settings, and about 464,000 were adherent to subsequent follow-up. The researchers cautioned that these proportions were derived from ranges reported across heterogeneous studies rather than from a single longitudinal cohort.
For diabetic retinopathy–related blindness, the researchers used 2 systematic reviews because primary US estimates were limited. They estimated that approximately 0.4% of patients with diabetic retinopathy in the US had diabetic retinopathy–related blindness and 2.64% had moderate to severe visual impairment. The researchers described these as derived estimates rather than pooled national prevalence estimates.
The review also highlighted potential reasons for disengagement. Lack of awareness of diabetic retinopathy diagnosis was associated in prior work cited by the researchers with longer diabetes duration, elevated A1C, and not having an eye examination in the prior year. Barriers to follow-up and treatment identified across the reviewed literature included younger age, systemic racism, language barriers, transportation difficulties, and lack of insurance. The researchers noted that follow-up adherence may be especially difficult because diabetic retinopathy management often requires more frequent visits than screening, and patients needing ongoing retinal care may also have more advanced diabetes or comorbidities competing for attention.
The cascade model was intended to be hypothesis-generating and was not designed to determine which stage contributes most to blindness, the researchers wrote. They hypothesized that lack of diagnosis awareness and loss to follow-up may contribute to adverse visual outcomes through different mechanisms, but said the available literature did not permit direct comparison or causal attribution.
Several limitations should be considered. Some included studies analyzed data collected before 2018, with the earliest data from 2005, meaning some estimates may not reflect current care patterns. Study populations, methods, and outcome definitions varied substantially. Some included studies used overlapping national data sets, including National Health and Nutrition Examination Survey data, which the researchers said may have reduced variability in some estimates. Evidence was also sparse for some cascade stages, including awareness of diabetic retinopathy diagnosis.
The researchers pointed to several potential interventions to address these gaps, including patient navigation, teleophthalmology-based services, artificial intelligence–assisted diagnosis tools, rural health clinics, and community-designed digital health tools.
The researchers concluded that awareness of diabetic retinopathy diagnosis, linkage to care, and follow-up adherence were among the lowest points of engagement in the care continuum and should be priorities for future research and intervention.
Disclosures: The study was supported by the National Eye Institute, an unrestricted/challenge award to Yale Eye Center from Research to Prevent Blindness, and the James G. Hirsch Endowed Medical Student Research Fellowship. The researchers reported no competing interests.
Source: BMJ Open
