Incidence and baseline predictors of discontinuation in DRCRnet studies

To evaluate the incidence and baseline predictors of participant discontinuation in Diabetic Retinopathy Clinical Research Retina Network (DRCRnet) studies METHODS: Using the publicly available DRCRnet data, discontinuation rate and its 95% confidence intervals (95% CI) were calculated. Univariable and multivariable logistic regression models were used to assess predictors for participant discontinuation, including demographic, clinical, and study design characteristics, and within-study correlation was accounted for using generalised estimating equations.

Among 7353 participants (mean age 60 years, 48.0% female, 64.1% non-Hispanic (NH)-White, 17.8% NH-Black, 14.2% Hispanic), 1110 participants (15.1%, 95% CI 14.3-15.9%) discontinued study participation over a median 24-month follow-up. Multivariable analysis showed significantly higher risk of discontinuation among Hispanic (aOR 1.73, 95% CI 1.38-2.18) and NH-Black (aOR 1.36, 95% CI 1.14-1.62) participants, those with age <50 years (aOR 1.46, 95% CI 1.21-1.77), 50-59 years (aOR 1.26, 95% CI 1.05-1.51), or ≥70 years (aOR 1.20, 95% CI 1.03-1.39), participants with higher baseline HbA1c (aOR 1.09 per unit increase, 95% CI 1.06-1.12), or baseline blood pressure ≥140/90 mmHg (aOR 1.24, 95% CI 1.09-1.41), and those in studies with follow-up length 24 to <36 months (aOR 1.64, 95% CI 1.02-2.64) or 36 to 48 months (aOR 2.68, 95% CI 1.87-3.84).

Approximately 15% of participants in the eligible DRCRnet studies discontinued participation. Racial and ethnic identity, age, glycaemic and blood pressure control, and study follow-up duration are significant predictors of participant discontinuation and should be considered when designing future studies.