Covariate-Adjusted Difference in Proportions from Clinical Trials Using Logistic Regression and Weighted Risk Differences

Abstract

Risk differences and associated confidence intervals are fiequentiy the basis for statistical testing in clinical trials. Ile analysis is often complicated by the presence of measured baseline covariates related to response that may be used to improve the precision of the treatment comparison by covariate adjustment in the statistical analysis. We use a clinical trial example and supporting simulations to show that logis-Miaomiao Go, PhD Covariate-adjusted Difference in Proportions From Clinical Trials Using Logistic Regression Candidate Pjzer Inc. Groton. Connecticut; Department Risk Differences and Weightedtic regression can be used to estimate the risk difference and compare its performance to common weighted-difference methods. We also examine when a useful improvement in precision can result fiom covariate adjustment, the use of a continuous rather than categorized covariate, and the consequences of including an unpredictive covariate.

Efficacy and Safety of MEDI2070, an Antibody Against Interleukin 23, in Patients With Moderate to Severe Crohn’s Disease: A Phase 2a Study

“…Analyses of the primary and CDAI-based dichotomous efficacy outcomes were performed using a logistic regression model, 24 with treatment assignment and number of prior TNF antagonists (1 vs >1) as covariates. Significance was tested using a 2-sided test of a ¼ .10 as prespecified in the analysis plan; however, 95% confidence intervals (CIs) for the absolute treatment differences for all efficacy outcomes are reported.…”

Treatment of atopic dermatitis with tralokinumab, an anti–IL-13 mAb

“…Covariate-adjusted differences in percentages with 95% CIs were calculated for binary end points. 15 Results from visits after a participant had received prohibited medications were excluded from the analysis. For continuous end points, no imputation for missing data was applied.…”

A randomised phase IIb study of mavrilimumab, a novel GM–CSF receptor alpha monoclonal antibody, in the treatment of rheumatoid arthritis

“…Two sensitivity analyses were performed for change from baseline DAS28−CRP to allow for patients withdrawing from treatment (see online supplementary material). ACR20/50/70 response rates, DAS28-defined remission (<2.6) and response rates at each visit were analysed using logistic regression, with results presented as differences in response rates (95% CI; p value) 28. Individual ACR components were analysed using the same method as for DAS28−CRP analyses.…”