This review assessed the effects of different screening intervals for diabetic retinopathy, on the incidence of sight-threatening diabetic retinopathy or vision loss. It included 25 studies, 15 of which were screening studies; the others were models of the natural history of the disease, or economic models. This summary describes the systematic review of clinical effectiveness, using the 15 included screening studies.
The review question was not clearly stated; the inclusion criteria for study design and participants were not stated. A limited search was undertaken to identify the relevant studies. It was unclear whether language restrictions were applied and there were no attempts to identify unpublished studies, so some relevant studies may have been missed. It was unclear if study selection was duplicated to reduce the potential for reviewer error and bias. Data extraction was undertaken by two reviewers, with disagreements resolved by a third reviewer, reducing the potential for error and bias. The quality of the included studies was not formally assessed, but the authors stated that more credit was given to well-designed, large, prospective studies with appropriate measures of outcomes.
Most studies did not formally assess the relationship between the screening interval and the incidence of diabetic retinopathy. They were conducted in the USA, Europe or Australia. Most participants were Caucasian, with type 1 or type 2 diabetes, and their ages ranged from under 18 years (children) to 99 years, where stated. The studies varied in their population and method of screening, but their results were fairly consistent; most studies supported a screening interval of around two years. A brief narrative synthesis was presented.
The authors concluded that, despite the between-study variation, the overall outcome was that two-year screening intervals, for people with no diabetic retinopathy at diagnosis, were not associated with a high incidence of sight-threatening diabetic retinopathy. This conclusion was based on the extremely few patients advancing from no diabetic retinopathy, to sight-threatening diabetic retinopathy, within two-to-three years.
There were several methodological limitations to the review. The included studies were reasonably consistent in their findings, but the potential for missed studies, and the unclear quality of those that were included, limit the reliability of the results of the review.