Purpose
Diabetes is the leading cause of blindness among United States (US) adults 40 years and older. Diabetes and diabetic complication rates have been shown to be higher in medically underserved populations. Gaining insight into diagnosis and ophthalmic care of minority, low-income, and uninsured patients will provide a further basis to effectively prevent, detect, and treat diabetic eye disease.
Methods
The HealthLNK database was used to identify approximately 2 million unique patients who visited one of the participating institutions from 2006-2012. HealthLNK includes electronic medical record (EMR) data from 6 federally qualified health centers (FQHCs), and 6 hospitals, including 4 academic medical centers in the Chicago area. Diabetic patients were defined by having ICD-9 codes for diabetes (250.xx) and/or diabetic complications (357.2, 362.01-362.07, and 366.41). From this population, patients with diabetic retinopathy (362.0-362.10, 362.1, 362.10, 362.14, 362.16, 362.2, 364.42) were elucidated. Diabetic retinopathy patients with CPT codes related to diabetic retinopathy treatment (67015, 67025, 67028-67031, 67036, 67039-67043, 67105, 67108, 67113, 67210, 67227, 67228) were further categorized. Insurance status was also determined within each subgroup.
Results
Of the 1,933,082 patients in the HealthLNK database, 171,427 were identified as diabetics (representing a total prevalence of 8.9%). 12,014 patients had diabetic retinopathy (7.0% of diabetics). 2,143 patients had CPT codes related to diabetic retinopathy treatment (17.8% of all retinopathy patients). There were differences in the prevalence of both diabetic retinopathy in diabetics (5.7% vs. 9.0%, p<0.01) and subsequent treatment (15.4% vs. 20.2%, p<0.01) when comparing "Medicaid/Financial Means Tested/Uninsured" vs. "Medicare/Privately Insured" patients.
Conclusions
The prevalence of diabetic retinopathy and procedures varied by insurance status, suggesting screening and treatment disparities may exist in this population. Future work will need to be done to elucidate the significance and reasons for these differences. This also work provides rationale for targeted screening and treatment strategies. The study also demonstrations the effectiveness of using large multi-center EMR data such as HealthLNK to identify healthcare disparities and design solutions to bridge this gap.