The Accuracy of Detecting Diabetic Retinopathy Using Panoptic Ophthalmoscope by Primary Care Doctors At Klinik Kesihatan Sendayan

Abstract

Diabetic Retinopathy (DR) is a microvascular complications of diabetes that may lead to blindness. The prevalence is increasing which demands early detection to prevent further deterioration of retinopathy. However, the gold standard fundus camera is not widely available in primary care. This study assesses the accuracy of PanOptic Opthalmoscope (PO) as a screening tool for DR at primary care level. A cross-sectional study on 66 diabetes patients were enrolled via systematic random sampling. Sample size was calculated using PASS software and OpenEpi method. PO and fundus camera were done on the same day with DR screening. PO were done among six primary care doctors. Fundus camera photo was interpreted by the Ophthalmologist as gold standard. Both findings were compared and analysed. PO has low accuracy (53%) in detecting DR.The sensitivity and specificity of  DR detection was 10.3 % and 86.5% respectively. The positive predictive value (PPV) was 37.5% and negative predictive value (NPV) was 55.2%. The accuracy of Sight threatening abnormalities (STA) was 75.8%. The sensitivity and specificity of STA were 35.3% and 89.8%, respectively. The PPV was 54.4% and NPV was 80%. The prevalence of newly diagnosed DR was 43.9%, while prevalence of STA was 25.8%. DR is significantly associated with patients living with diabetes for more than 10 years (p = 0.007). The accuracy of PO in DR screening is poor, but it is modestly acceptable for STA detection. However, the sensitivity and PPV is low, causing some serious retinopathy cases undetected. The prevalence of DR is higher with increasing duration of diabetes. We recommend the availability of fundus camera in primary care setting.