Medi‐Cal.

Study characteristics
Methods	Study design: parallel‐group RCT Unit of randomization: individual Number randomized, total: 358 Number analyzed: 240 for DR Per group: 121 intensive, 119 standard Sample size calculation: based on 1% difference between groups in glycated hemoglobin and 30% dropout rate, power > 95%, 2‐sided probability of 0.05, yielding 150 participants per arm in parent study
Participants	Country: USA (2 centers) Study period: enrollment July 1995 to June 1999 Type of diabetes: type 2 Participants' status at baseline:in retinopathy substudy except as noted Age, years, mean (SD): 55.0 (11.6) intensive, 55.5 (12.9) standard, 55.2 combined Gender, % women: 75% intensive, 69% standard, 72% combined Race/ethnicity:  White, non‐Hispanic 41% intensive, 42% standard, 42% combined Hispanic 41% intensive, 35% standard, 38% combined Black 14% intensive, 13% standard, 14% combined Duration of diabetes, years, mean (SD): 8.8 (6.8) intensive, 10.6 (8.5) standard, 9.7 combined Smokers, % current: 14.7% intensive, 13.9% standard, 14.3% combined Blood pressure, mean (SD): SBP 136 (2) mmHg intensive, 134 (1) mmHg standard, 135 mmHg combined DBP 81 (9) mmHg intensive, 76 (1) mmHg standard, 78 mmHg combined Hypertension, %:  not reported Glycated hemoglobin, mean (SD): 9.6 (1.8) intensive, 9.4 (1.5) standard, 9.5 combined BMI, mean (SD): in parent study 33.1 (0.8) intensive, 31.5 (0.8) standard, 32.3 combined Severity of retinopathy: None 61% intensive, 49% standard, 55% combined Mild NPDR 38% intensive, 38% standard, 38% combined Severe NPDR 3% intensive, 11% standard, 6% combined Medical history: Myocardial infarction, %: not reported Stroke, %: not reported Equivalence of groups at baseline: well balanced Inclusion criteria: glycated hemoglobin > 7.5% Exclusion criteria: not reported
Interventions	Intervention 1: intensive case management + primary care, including home glucose monitor, individualized education regarding diabetes, diet, exercise, and self‐care Intervention 2: standard, primary care only Length of follow‐up: Planned: ≥ 1 year Actual: ≲ 3 years; mean 2.1 years
Outcomes	Primary outcome, as specified for this review: incidence of DR Secondary outcome, as specified for this review: none Other diabetic retinopathy outcome(s): combined incidence and progression of DR Retinopathy diagnosis and monitoring: "All photographs were labeled with only the patient's identification number and were sent for reading in Santa Barbara ... and graded by an experienced endocrinologist who, before the study, had readings verified by and ophthalmologist until agreement was virtually 100%." Participants (eyes) examined for retinopathy outcomes: 82 for incidence, 100 for combined incidence and progression Intervals at which DR outcomes assessed: 6 months and yearly for fundus photos Cost of interventions: not reported Quality of life: not reported Adverse outcomes: hypoglycemia Other outcomes reported from the study: change in glycated hemoglobin, body weight, lipids, blood pressure
Notes	Source(s) of funding: State of California Medi‐Cal Managed Care Division and Center for Disease Control, Prevention, CTRC at Harbor‐UCLA Medical Center, Lifescan (meters and supplies) Declaration of interest: not reported Length of run‐in period: not applicable Class(es) of antihypertensive agents assigned: none Degree of blood pressure control achieved: SBP and DBP each decreased by about 3 mmHg from baseline in intensive arm compared to < 1 mmHg change in standard arm Trial registration: not reported and not found
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information provided.
Allocation concealment (selection bias)	Unclear risk	No information provided.
Masking (performance bias and detection bias) Primary outcomes	Unclear risk	Not possible to mask caregivers or participants to intervention arm. Photographs read centrally by an endocrinologist for DR outcomes.
Incomplete outcome data (attrition bias) Primary outcome	High risk	Of 240 participants at 2 centers with fundus cameras, 200 had at least 1 follow‐up photo. Of 61 participants without DR at baseline in intensive arm, incidence reported for 48 (79%); of 49 participants without DR in standard arm, incidence reported for 34 (69%).
Selective reporting (reporting bias)	Unclear risk	Combined incidence and progression of DR not reported in an analyzable form. Otherwise, all outcomes mentioned in 2 publications were reported.