Diagnostic test accuracy of diabetic retinopathy screening by physician graders using a hand-held non-mydriatic retinal camera at a tertiary level medical clinic

Purpose
Diabetic retinopathy (DR) is a common complication of diabetes mellitus (DM), leading to sight loss if not detected and treated in time. Different models of DR screening (DRS) have been implemented in many parts of the world. In resource poor low and middle-income countries (LMICs) development of a DRS model is complex, due to lack of resources and skilled human resource. Therefore task-shifting and using locally adaptable technology is useful to initiate DRS. The evidence on diagnostic test accuracy (DTA) of DRS utilising retinal photographic studies by non-ophthalmologist personnel in LMIC settings is scarce. This study aims to demonstrate the functional and technical feasibility of using a hand-held non-mydriatic digital camera in a LMIC non-ophthalmic setting. We assessed the DTA of DR detection by general physicians using this method compared to the local clinical reference standard of mydriatic indirect ophthalmoscopy and bio-microscopic examination by a retinologist. This study was a component of a doctoral research project and Ethics approval was obtained from both Ethics Review Committees of the London School of Hygiene & Tropical Medicine-United Kingdom and the National Eye Hospital-Sri Lanka.
Setting/Venue
Nine general physicians from a tertiary level institution in Sri Lanka underwent a competency-based training programme following written informed consent, delivered by two retinologists. The training included the following: capturing retinal fields using a hand-held non-mydriatic fundus camera (Zeiss-Visuscout100®-Germany), identification of signs of DR using images and DR grading according to an adopted classification system based on the United Kingdom - National Screening System. The screening intervention validation study was conducted in a public sector tertiary level out-patient medical clinic in Sri Lanka.
Methods
A sample size of n=506 participants was chosen, in order to estimate the sensitivity within a margin of error 10% (based on 95% confidence intervals [CI]), with an expected sensitivity of 70% and prevalence of moderate non-proliferative DR among people with DM (PwDM) of 20%. This included an additional 25% to take account of ungradable images (i.e., 18 years of age, not previously screened or treated for DR, presenting at the medical clinic. Two-field retinal imaging was used to capture fundus images before and after pupil dilatation, using a hand-held non-mydriatic digital retinal camera. The images were captured and graded by two trained, masked independent physician graders. The DTA (sensitivity, specificity, and predictive values) of different levels of DR was assessed comparing physician’s grading with a retinologist’s clinical examination by mydriatic bio-microscopy, to identify any level of DR and referrable level DR (moderate non-proliferative DR and levels above).
Results
Eligible PwDM identified from medical clinical records and response rate was 84.7% (700/826). Mean age of the participants was 60.8 years (SD ±10.08). Ungradable image proportion in non-mydriatic imaging was 43.4% and decreased to 12.8% following pupil dilatation. Sensitivity of detection of any level of DR using non-mydriatic imaging (ungradable accounted as screen positives) was 82.7% (95% CI 78.4-86.5%) in grader-1 and 78.3% (95% CI 73.7-82.5%) in grader-2 and specificity values dropped to 70.4% (95% CI 67.6-73.1%) in grader-1 and 76.2% (95% CI 73.6-78.7%) in grader-2 in non-mydriatic imaging. As a pragmatic approach for a resource poor non-ophthalmic setting, we reported the DTA of DRS using non-mydriatic imaging and dilatation of the pupils of only those who have ungradable images (two-step process). In this sub-analysis, the eye which was ungradable even following mydriasis were considered as screen positives. We derived a sensitivity of referable level of DR 81.1% (95% CI 72.9-87.9%) for grader-1 and 82.1% (95% CI 74.0-88.6%) for grader-2. The specificity values were 95.4% (95% CI 94.2-96.5%) for grader-1 and 97.1% (95% CI 96.1-97.9%) for grader-2 in this approach. We observed an improved level of PPV (59.7-70.2%) and NPV (98.4-98.5%) in this strategy.
Conlusions
In this study we demonstrated that the diagnostic test accuracy of the physician graders was closer to the standard practice of national level screening programs in other settings. We conclude that 2-field retinal imaging using a hand-held digital camera at a medical clinic, by physician graders, with dilatation of pupil of those who have ungradable images, provides a valid modality to identify referable level of diabetic retinopathy. This strategy is an accurate screening method of detection of a referable level in a health care facility-based people with diabetes who are at risk of developing sight threatening diabetic retinopathy.