Background: Randomization is the foundation of any clinical trial involving treatment comparison. It helps mitigate selection bias, promotes similarity of treatment groups with respect to important known and unknown confounders, and contributes to the validity of statistical tests. Various restricted randomization procedures with different probabilistic structures and different statistical properties are available. The goal of this paper is to present a systematic roadmap for the choice and application of a restricted randomization procedure in a clinical trial.Methods: We survey available restricted randomization procedures for sequential allocation of subjects in a randomized, comparative, parallel group clinical trial with equal (1:1) allocation. We explore statistical properties of these procedures, including balance/randomness tradeoff, type I error rate and power. We perform head-to-head comparisons of different procedures through simulation under various experimental scenarios, including cases when common model assumptions are violated. We also provide some real-life clinical trial examples to illustrate the thinking process for selecting a randomization procedure for implementation in practice.Results: Restricted randomization procedures targeting 1:1 allocation vary in the degree of balance/randomness they induce, and more importantly, they vary in terms of validity and efficiency of statistical inference when common model assumptions are violated (e.g. when outcomes are affected by a linear time trend; measurement error distribution is misspecified; or selection bias is introduced in the experiment). Some procedures are more robust than others. Covariate-adjusted analysis may be essential to ensure validity of the results. Special considerations are required when selecting a randomization procedure for a clinical trial with very small sample size.Conclusions: The choice of randomization design, data analytic technique (parametric or nonparametric), and analysis strategy (randomization-based or population model-based) are all very important considerations. Randomization-based tests are robust and valid alternatives to likelihood-based tests and should be considered more frequently by clinical investigators.

BACKGROUND. We investigated residual β cell function in Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study participants with an average 35-year duration of type 1 diabetes mellitus (T1DM). METHODS. Serum C-peptide was measured during a 4-hour mixed-meal tolerance test. Associations with metabolic outcomes and complications were explored among nonresponders (all C-peptide values after meal 0.2), intermediate (>0.03 to ≤0.2), and low (≥ 0.003 to ≤0.03). RESULTS. Of the 944 participants, 117 (12.4%) were classified as responders. Residual C-peptide concentrations were associated with higher DCCT baseline concentrations of stimulated C-peptide (P value for trend = 0.0001). Residual C-peptide secretion was not associated with current or mean HbA1c, HLA high-risk haplotypes for T1DM, or the current presence of T1DM autoantibodies. The proportion of subjects with a history of severe hypoglycemia was lower with high (27%) and intermediate (48%) residual C-peptide concentrations than with low (74%) and no (70%) residual C-peptide concentrations (P value for trend = 0.0001). Responders and nonresponders demonstrated similar rates of advanced microvascular complications. CONCLUSION. β Cell function can persist in long-duration T1DM. With a peak C-peptide concentration of >0.03 nmol/L, we observed clinically meaningful reductions in the prevalence of severe hypoglycemia. TRIAL REGISTRATION. ClinicalTrials.gov NCT00360815 and NCT00360893. FUNDING. Division of Diabetes Endocrinology and Metabolic Diseases of the National Institute of Diabetes and Digestive and Kidney Diseases (DP3-DK104438, U01 DK094176, and U01 DK094157).

Randomization is considered to be a key feature to protect against bias in randomized clinical trials. Randomization induces comparability with respect to known and unknown covariates, mitigates selection bias, and provides a basis for inference. Although various randomization procedures have been proposed, no single procedure performs uniformly best. In the design phase of a clinical trial, the scientist has to decide which randomization procedure to use, taking into account the practical setting of the trial with respect to the potential of bias. Less emphasis has been placed on this important design decision than on analysis, and less support has been available to guide the scientist in making this decision. We propose a framework that weights the properties of the randomization procedure with respect to practical needs of the research question to be answered by the clinical trial. In particular, the framework assesses the impact of chronological and selection bias on the probability of a type I error. The framework is applied to a case study with a 2-arm parallel group, single center randomized clinical trial with continuous endpoint, with no-interim analysis, 1:1 allocation and no adaptation in the randomization process. In so doing, we derive scientific arguments for the selection of an appropriate randomization procedure and develop a template which is illustrated in parallel by a case study. Possible extensions are discussed. The proposed ERDO framework guides the investigator through a template for the choice of a randomization procedure, and provides easy to use tools for the assessment. The barriers for the thorough reporting and assessment of randomization procedures could be further reduced in the future when regulators and pharmaceutical companies employ similar, standardized frameworks for the choice of a randomization procedure.

Introduction: Observational studies of SARS-CoV-2 vaccine effectiveness depend on accurate ascertainment of vaccination receipt, date, and product type. Self-reported vaccine data may be more readily available to and less expensive for researchers than assessing medical records. Methods: We surveyed adult participants in the COVID-19 Community Research Partnership who had an authenticated Electronic Health Record (EHR) (N = 41,484) concerning receipt of SARS-CoV-2 vaccination using a daily survey beginning in December 2020 and a supplemental survey in September–October 2021. We compared self-reported information to that available in the EHR for the following data points: vaccine brand, date of first dose, and number of doses using rates of agreement and Bland–Altman plots for visual assessment. Self-reported data was available immediately following vaccination (in the daily survey) and at a delayed interval (in a secondary supplemental survey). Results: For the date of first vaccine dose, self-reported “immediate” recall was within ±7 days of the date reported in the “delayed” survey for 87.9% of participants. Among the 19.6% of participants with evidence of vaccination in their EHR, 95% self-reported vaccination in one of the two surveys. Self-reported dates were within ±7 days of documented EHR vaccination for 97.6% of the “immediate” surveys and 92.0% of the “delayed” surveys. Self-reported vaccine product details matched those in the EHR for over 98% of participants for both “immediate” and “delayed” surveys. Conclusions: Self-reported dates and product details for COVID-19 vaccination can be a good surrogate when medical records are unavailable in large observational studies. A secondary confirmation of dates for a subset of participants with EHR data will provide internal validity.

The impact of selection bias on the results of clinical trials has been analyzed extensively for trials of two treatments, yet its impact in multi-arm trials is still unknown. In this paper, we investigate selection bias in multi-arm trials by its impact on the type I error probability. We propose two models for selection bias, so-called biasing policies, that both extend the classic guessing strategy by Blackwell and Hodges. We derive the distribution of the F-test statistic under the misspecified outcome model and provide a formula for the type I error probability under selection bias. We apply the presented approach to quantify the influence of selection bias in multi-arm trials with increasing number of treatment groups using a permuted block design for different assumptions and different biasing strategies. Our results confirm previous findings that smaller block sizes lead to a higher proportion of sequences with inflated type I error probability. Astonishingly, our results also show that the proportion of sequences with inflated type I error probability remains constant when the number of treatment groups is increased. Realizing that the impact of selection bias cannot be completely eliminated, we propose a bias adjusted statistical model and show that the power of the statistical test is only slightly deflated for larger block sizes.

To evaluate changes in retinal thickness and morphology using OCT in youth with type 2 diabetes (T2D) and to identify systemic biomarkers correlating with these changes. Retrospective subgroup analysis of a prospective study. Participants who underwent OCT imaging in the Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) trial and its follow-up study TODAY2. In 2010–2011 (TODAY) and 2017–2018 (TODAY2), 6 × 6-mm macular volume OCT scans were acquired, segmented, and analyzed to generate total retinal thickness, inner retinal thickness, and outer retinal thickness. The main retinal morphologies graded were intraretinal cystoid spaces, subretinal fluid, and posterior vitreous detachment (PVD). Changes in total and individual retinal layer thickness and development of abnormal vitreomacular morphology between TODAY and TODAY2. Participants had a mean age of 17.9 ± 2.4 years and glycated hemoglobin (HbA1c) of 8.2 ± 2.8% in TODAY and a mean age of 25.0 ± 2.4 years and mean HbA1c of 9.5 ± 2.8% in TODAY2. Longitudinally between assessments, there were overall decreases in outer retinal thickness from 167.2 ± 11.5 microns to 158.4 ± 12.8 microns (P < 0.001) and in photoreceptor thickness from 30.3 ± 2.9 microns to 29.8 ± 4.1 microns (P = 0.04) in the central subfield, while in the inner subfield, we noted a decrease in outer retinal thickness from 150.5 ± 10.1 microns to 144.9 ± 10.5 microns (P < 0.001) and an increase in inner retinal thickness from 136.9 ± 11.5 microns to 137.4 ± 12.6 microns (P = 0.01). Multivariate analysis showed that in the center subfield, HbA1c increases were associated with increases in total retinal thickness (r: 0.67, P = 0.001), whereas fasting glucose was positively correlated with inner retinal thickness (r: 0.02, P = 0.02). In the inner subfield, both systolic (r: −0.22, P < 0.001) and diastolic (r: −0.22, P = 0.003) blood pressures were negatively correlated with total retinal thickness. There was an increase in PVD (18.9%) and cystoid spaces (4.2%). Youth with T2D develop retinal thickness changes on OCT, including increases in total retinal and inner retinal thickness in the center subfield that correlate with HbA1c and fasting glucose, respectively. Taken together with the increased prevalence of abnormal vitreomacular morphology in this cohort at risk, these findings emphasize the importance of controlling risk factors to prevent the development of sight-threatening retinal complications.