Are we making good use of our public resources? The false-positive rate of screening by fundus photography for diabetic macular oedema

Raymond L.M. Wong, C. W. Tsang, David S.H. Wong, Sarah McGhee, C. H. Lam, J. Lian, Jacky W.Y. Lee, Jimmy S.M. Lai, Victor Chong, Ian Y.H. Wong

Research output: Journal article publicationJournal articleAcademic researchpeer-review

11 Citations (Scopus)


Introduction: A large proportion of patients diagnosed with diabetic maculopathy using fundus photography and hence referred to specialist clinics following the current screening guidelines adopted in Hong Kong and United Kingdom are found to be false-positive, implying that they did not have macular oedema. This study aimed to evaluate the false-positive rate of diabetic maculopathy screening using the objective optical coherence tomography scan. Methods: This was a cross-sectional observational study. Consecutive diabetic patients from the Hong Kong West Cluster Diabetic Retinopathy Screening Programme with fundus photographs graded R1M1 were recruited between October 2011 and June 2013. Spectral-domain optical coherence tomography imaging was performed. Central macular thickness of ≥300 µm and/or the presence of optical coherence tomography signs of diabetic macular oedema were used to define the presence of diabetic macular oedema. Patients with conditions other than diabetes that might affect macular thickness were excluded. The mean central macular thickness in various subgroups of R1M1 patients was calculated and the proportion of subjects with central macular thickness of ≥300 µm was used to assess the false-positive rate of this screening strategy. Results: A total of 491 patients were recruited during the study period. Of the 352 who were eligible for analysis, 44.0%, 17.0%, and 38.9% were graded as M1 due to the presence of foveal ‘haemorrhages’, ‘exudates’, or ‘haemorrhages and exudates’, respectively. The mean (±standard deviation) central macular thickness was 265.1 ± 55.4 µm. Only 13.4% (95% confidence interval, 9.8%-17.0%) of eyes had a central macular thickness of ≥300 µm, and 42.9% (95% confidence interval, 37.7%-48.1%) of eyes had at least one optical coherence tomography sign of diabetic macular oedema. For patients with retinal haemorrhages only, 9.0% (95% confidence interval, 4.5%-13.5%) had a central macular thickness of ≥300 µm; 23.2% (95% confidence interval, 16.6%-29.9%) had at least one optical coherence tomography sign of diabetic macular oedema. The false-positive rate of the current screening strategy for diabetic macular oedema was 86.6%. Conclusion: The high false-positive rate of the current diabetic macular oedema screening adopted by the United Kingdom and Hong Kong may lead to unnecessary psychological stress for patients and place a financial burden on the health care system. A better way of screening is urgently needed. Performing additional spectral-domain optical coherence tomography scans on selected patients fulfils this need.

Original languageEnglish
Pages (from-to)356-364
Number of pages9
JournalHong Kong Medical Journal
Issue number4
Early online date7 Jul 2017
Publication statusPublished - Aug 2017
Externally publishedYes

ASJC Scopus subject areas

  • Medicine(all)

Cite this