Domain	Signalling items	Authors' judgement for 'yes'
1. Study participation	(a) Adequate participation in study by eligible individuals	The sampling frame and recruitment are adequately described, including methods to identify the sample sufficient to limit potential bias (number and type used, e.g. referral patterns in health care)
	(b) Description of target population	Source population for cohort with diabetic retinopathy (DR) is clearly described
	(c) Description of baseline study sample	Number of people with DR at baseline is clearly described
	(d) Adequate description of recruitment process	Way of establishing the source population, selection criteria and key characteristics of the source population clearly described
	(e) Adequate description of period and place of recruitment	Time period and place of recruitment for both baseline and follow‐up examinations are clearly described
	(f) Adequate description of inclusion/exclusion criteria	Definition of DR and other inclusion and exclusion criteria clearly defined
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
2. Study attrition	(a) Adequate response rate for study participants	Response rate (i.e., proportion of study sample completing the study and providing outcome data) is adequate
	(b) Description of process for collecting information on participants who dropped out	Attempts to collect information on participants who dropped out are described (e.g. telephone contact, mail, registers)
	(c) Reasons for loss to follow‐up provided	Reasons on participants who dropped out are reported
	(d) Adequate description of participants lost to follow‐up	Key characteristics of participants lost to follow‐up are described
	(e) No important differences between participants who completed the study and those who dropped out	Study authors described differences between participants completing the study and those who did not as not important or information provided to judge the differences
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
3. Prognostic factor measurement	(a) Clear definition of prognostic factor (PF) provided	Measurements for prognostic factors (PFs) are provided
	(b) Method of PF measurement is adequately valid and reliable	Measurements techniques for prognostic factors are described and likely to be valid and reliable, e.g., standardised, repeated
	(c) Continuous variables are reported	Standard categories for prognostic factors / cut‐offs
	(d) Method and setting of measurement of PF is identical for all participants	Measurements of PFs are the same for all study participants
	(e) Adequate proportion of study sample has complete data for PF	Adequate proportion of the study sample has complete data for PF variable
	(f) Appropriate methods of imputation used for missing PF data	Appropriate methods of imputation are used for missing PF data
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
4. Outcome measurement	a) Clear definition of outcome provided	Measurement of proliferative diabetic retinopathy(PDR)/high‐risk characteristics (HRC) is defined
	(b) Method of outcome measurement is adequately valid and reliable	Measurement of PDR/HRC has to be a part of a diagnostic assessment
	(c) Method and setting of outcome measurement is identical for all participants	Measurements of PDR/HRC are the same for all study participants
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
5. Adjustment for other prognostic factors	(a) All other important PFs measured	Important confounders are: HbA1c and duration of DM
	b) Clear definitions of important PFs measured provided	Measurement of confounders has to be clearly described
	(c) Measurement of all important PFs adequately valid and reliable	Measurement of confounders is valid and reliable
	(d) Measurement and setting of PF measurement identical for all participants	Measurements of confounders are the same for all study participants
	(e) Appropriate methods are used to deal with missing values of PFs	Strategy to impute missing confounder data is described
	(f) Important PFs accounted for in study design	Methods section of the publication describes strategy to account for confounders
	(g) Important PFs accounted for in analysis	Important confounders are accounted for in multivariable logistic regression and Cox proportional hazards models
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'
6. Statistical analysis and reporting	(a) Sufficient presentation of data to assess adequacy of analytic strategy	Mean or median values, including confidence intervals or standard errors or standard deviations provided
	(b) Strategy for model building appropriate and based on a conceptual framework or model	The selected statistical model is adequate for the design of the study
	(c) Selected statistical model adequate for design of study	Mainly incidence rates, uni‐ and multivariate logistic regression, Cox proportional hazard model
	(d) No selective reporting of results	There is no selective reporting of results
	Domain overall risk of bias	High: most items are answered with 'no'; Low: all items answered with 'yes'; Moderate: most items are answered with 'unclear'