The use of digital cameras in a mobile retinal screening environment

Diabet Med. 1999 Aug;16(8):680-6. doi: 10.1046/j.1464-5491.1999.00139.x.


Aims: To assess how successfully digital camera technology might be incorporated into a mobile screening environment.

Methods: One hundred and ninety-seven people had their fundi photographed using a Topcon/Imagenet digital system and 534 using a Canon CR5/Ris-Lite system in addition to concurrent 45 degrees CR4NM Polaroid photography. One hundred and eighteen randomly selected patients were also sent for 7 field stereo photography as a gold standard. An acceptability questionnaire was answered by a random sample of those photographed.

Results: For the detection of any retinopathy, digital pictures had a sensitivity of 0.74 (95% confidence interval (CI) 0.68-0.80) whilst Polaroid was 0.72 (95% CI 0.66-0.78) and for referable retinopathy digital pictures had a sensitivity of 0.85 (95% CI 0.80-0.90) and Polaroid was 0.90 (95% CI 0.86-0.94). A concurrent ophthalmoscopic evaluation improved the sensitivity to 0.92 (95% CI 0.86-0.98) for detection of any retinopathy whilst the sensitivity at the referral level was improved to 0.95 (95% CI 0.91-0.99). Twenty-nine of 176 respondents experienced flash discomfort with the Polaroid system with only four of 154 describing 'some' discomfort from the digital systems which have a lower flash power (10 W vs. 300 W) and a faster recovery time.

Conclusions: This study indicates that digital systems are a feasible and acceptable alternative to Polaroid-based cameras for use in a mobile environment.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Blindness / prevention & control*
  • Confidence Intervals
  • Diabetic Retinopathy / diagnosis*
  • Diabetic Retinopathy / epidemiology
  • England / epidemiology
  • Female
  • Fluorescein Angiography* / instrumentation
  • Humans
  • Male
  • Mass Screening / methods*
  • Middle Aged
  • Mobile Health Units*
  • Sensitivity and Specificity