Table 1. Evidence classification.

I	Integrated analyses	Any meta-analyses of cohort and case-control studies
II	Cohort study design	Comparison of the incidence of a given comorbid condition between patients with and without diabetes. The incidence rate ratio of the condition is estimated or can be estimated regarding diabetes exposure. There is no distinction between prospective and retrospective studies.
III	Case-control study design (longitudinal)	Comparison of the proportion with a history of diabetes among those who did and did not experience a given comorbid condition. The odds ratio of the condition is estimated regarding diabetes exposure. The timing of exposure (diabetes) and the outcome measurement should be separate (longitudinal).
IV	Cross-sectional study design	Comparison of the prevalence of a given comorbid condition between patients with and without diabetes. The prevalence ratio or odds ratio is either estimated or can be estimated.