The timely detection and treatment of diabetic retinopathy is an important component of diabetes management.  At RPAH Diabetes Centre, fundal examination and grading of retinopathy (modified Airlie House classification) are routinely performed. In Australia this function is increasingly undertaken by outside ophthalmologists and optometrists (O-O&O). This trend is fostered by the availability of retinal photography and by emphasis on completing cycles of care at the general practice level.  In our experience, retrieving eye data from O-O&O is time consuming and information gathered is often difficult to interpret.  The aim of this study is to quantitate the deficiencies of this process.  We also examined whether the data from O-O&O are representative.

Eye data from 638 patients who attended O-O&O in the last three years were extracted from the Diabetes Centre electronic database. To minimise bias, each O-O&O was limited to 15 patients (mean 2.3) and only the first letter for each patient was analysed.  This resulted in 310 letters (260 Ophthalmologists and 50 Optometrists) eligible for coding according to their utility in indicating severity of retinopathy, presence of macular pathology and which eye(s) were affected. 

The comparison of letters from O-O&O is shown in Figure 1.  In letters from 16% of ophthalmologists and 36% of optometrists, severe vision-threatening and milder retinopathy cannot be distinguished. Only ~60% of either group commented on macular pathology and a minority of documents specified the affected eye(s). Figure 2 demonstrates that for any duration of diabetes, more ophthalmologists’ patients had retinopathy than a comparable RPAH cohort (n=6496) while the optometrists’ patients had the lowest prevalence of retinopathy.

We conclude that in addition to the heavy burden of retrieving data from O-O&O correspondence, there is also difficulty in deriving precise and clinically relevant information. This hampers efficient clinical decision-making such as establishing urgency of intervention and individualising glycaemic targets.  A national standardised reporting pathway between health professionals and using a uniform and preferably numerical grading system can overcome many of these problems. Moreover, information gathered from O-O&O should not be considered representative of the whole population because of possible distortion to estimates of disease burden.