Systematic Assessment of Clinical Methods to Diagnose and Monitor Diabetic Retinal Neuropathy

Table 1

Strength of evidence (modified GRADE criteria).

Quality level	Definition
Compelling	Statistically significant relationship between a positive investigation result and the subsequent development of diabetic retinal pathology is shown in the study design. For a metric to be compelling, the study had to demonstrate a statistically significant predictability of change in the metric with the development of diabetic retinopathy
Indicative	Studies where the metric is used are either not shown to be statistically significant in the prediction of diabetic retinopathy but is still a useful clinical measure for presence and/or deterioration of retinopathy, or reviewers did not agree on the evidence of its compelling value
Moderate	Some evidence of a statistically significant relationship between a positive investigation result and the subsequent development of diabetic retinal pathology but problems with study design or applicability
Weak	Equivocal, unconvincing, and statistically insignificant correlation between investigation in the study and development of diabetic retinopathy
None	No evidence of correlation between investigation and development of diabetic retinopathy or contradictory evidence

A version of the GRADE criteria developed by Guyatt et al. [28–32] was created by the authors specific to ophthalmic interventions. Many ophthalmic clinical investigations are reliable and compelling due to their basis in psychophysics or their capacity for direct observation, and therefore, the terminology ‘compelling’ is used rather than ‘strong’, although definitive diabetic retinopathy cannot currently be demonstrated either via observation or through psychophysics with a single test.