Retinal photography with a non-mydriatic retinal camera in people with diabetes

Ellery B, Milverton J, Newton S, Morona J, Gum D, Parsons J, Vogan A, Fischer S, Merlin T
Record ID 32015000104
Original Title: Application 1181
Authors' objectives: To determine whether retinal photography with a nonmydriatic retinal camera (RP-NMRC) is safe, effective and cost-effective for the detection of diabetic retinopathy (DR) in patients with diagnosed diabetes. The assessment was undertaken to inform whether RP-NMRC for the described population should be publicly funded in the primary care setting
Authors' results and conclusions: Safety - No studies on comparative safety of retinal photography with a non-mydriatic retinal camera (RP-NMRC) were identified. Accuracy to detect any diabetic retinopathy (DR) - Meta-analysis of 13 studies compared RP-NMRC with slit lamp biomicroscopy (SLBM) and/or comprehensive eye examination (CEE) to detect any level of DR. Overall pooled results showed that RP-NMRC can accurately confirm the presence of any DR (sensitivity 91.2%, 95%CI 81.7, 96.1; positive likelihood ratio (LR+) 3.88, 95%CI 2.79, 5.40) with a trade-off in ability to rule out DR (specificity 76.5%, 95%CI 67.4, 83.6; negative likelihood ratio (LR–) 0.11, 95%CI 0.05, 0.24). For the detection of any DR by RP-NMRC, an analysis conducted according to use or non-use of mydriasis showed no appreciable difference: sensitivity 89.4% (95%CI 72.2, 96.5) versus 91.7% (95%CI 79.2, 97.0), respectively; specificity 74.2% (95%CI 63.5, 82.7) versus 74.4% (95%CI 61.2, 84.2), respectively. There were no studies identified that reported on the relative accuracy of RP-NMRC versus ophthalmoscopy by a GP, with CEE as the reference standard. Overall conclusion on comparative effectiveness - RP-NMRC is a more effective tool for triaging patients with diabetes for further assessment with CEE than no eye examination or ophthalmoscopy delivered in primary care. Patients who receive a false negative result from RP-NMRC, like those who do not have an eye examination or who receive a false negative result from ophthalmoscopy in primary care, are at risk of blindness as they will not be referred to an eye specialist for a CEE following screening, and treatment may be delayed until their disease is symptomatic. It is likely, however, that the false negative rate with RP-NMRC will be smaller than not having an examination at all. Patients who receive a false positive result for DR by RP-NMRC are not likely to be negatively affected (except with regard to inconvenience), as they will be referred to an eye specialist for a CEE, at which point DR will be excluded. Overall cost-effectiveness - RP-NMRC is likely to be a cost-effective option for diagnosing DR in patients with diabetes who would not otherwise receive regular eye examinations. The sensitivity analyses confirm that the results of the economic model comparing RP-NMRC testing with the primary comparator (no testing) are reasonably robust. The model is most sensitive to the cost of treatment and the quality-of-life weight applied to the advanced sight-threatening DR health state, but the ICER remains below $45,000/QALY gained in all modelled scenarios. The introduction of RP-NMRC will only be effective if provision is made to ensure compliance with regular testing, appropriate follow-up of results, and prompt treatment of sight-threatening DR when indicated.
Project Status: Completed
Year Published: 2014
English language abstract: An English language summary is available
Publication Type: Not Assigned
Country: Australia
MeSH Terms
  • Diabetic Retinopathy
  • Diagnostic Techniques, Ophthalmological
  • Photography
  • Ophthalmology
  • Mydriatics
  • Predictive Value of Tests
  • Retina
Organisation Name: Adelaide Health Technology Assessment
Contact Address: School of Public Health, Mail Drop 545, University of Adelaide, Adelaide SA 5005, AUSTRALIA, Tel: +61 8 8313 4617
Contact Name:
Contact Email:
Copyright: Adelaide Health Technology Assessment (AHTA)
This is a bibliographic record of a published health technology assessment from a member of INAHTA or other HTA producer. No evaluation of the quality of this assessment has been made for the HTA database.