Back to Explorer

Adjusting for Covariates in Randomized Clinical Trials for Drugs and Biological Products

FinalCenter for Drug Evaluation and Research Center for Biologics Evaluation and Research05/26/2023
Covariate Adjustment

Description

This guidance describes FDA’s current recommendations regarding adjusting for covariates in the statistical analysis of randomized clinical trials in drug development programs. This guidance provides recommendations for the use of covariates in the analysis of randomized, parallel group clinical trials that are applicable to both superiority trials and noninferiority trials. The main focus of the guidance is on the use of prognostic baseline covariates to improve statistical efficiency for estimating and testing treatment effects. This guidance does not address use of covariates to control for confounding variables in non-randomized trials, the use of covariates in models to account for missing outcome data (National Research Council 2010), the use of covariate adjustment for analyzing longitudinal repeated measures data, the use of Bayesian methods for covariate adjustment, or the use of machine learning methods for covariate adjustment.

Key Topics

Terms and concepts identified from this document

Scope & Applicability

Product Classes

2
Drugs

Clinical investigations of drugs, including human drugs and biological products

Biological Products

Requires analytical comparability per ICH Q5E

Stakeholders

3
Sponsor

Entity responsible for submitting applications under section 524B

Review Division

Sponsors should consult regarding hybrid or augmented data approaches

Sponsors

Assist sponsors in the nonclinical evaluation

Regulatory Context

Regulatory Activities

1
Randomized Clinical Trials

statistical analysis of randomized clinical trials in drug development programs

Document Types

2
Statistical Analysis Plan

Document for protocol execution; The sponsor should develop a statistical analysis plan that is consistent with the trial protocol; Deviations from this plan must be justified

Protocol

Defines the standard of veterinary practice and limits for anesthetic regimens

Attributes

8
Confidence intervals

sponsors should include confidence intervals on all reported results

Prognostic baseline variables

Leveraging Prognostic Baseline Variables to Gain Precision

Statistical Efficiency

improve statistical efficiency for estimating and testing treatment effects

Average Treatment Effect

Linear model is an acceptable method for estimating the average treatment effect

Unconditional Treatment Effect

The average treatment effect is an example of an unconditional treatment effect

Conditional Treatment Effect

subgroup-specific conditional treatment effect can differ from the unconditional treatment effect

Non-collapsibility

This is termed non-collapsibility, which is distinct from confounding

Standard errors

computing standard errors in stratified randomization

Technical Details

Testing Methods

10
Covariate Adjustment

use of baseline covariate measurements for estimating and testing population-level treatment effects

Linear Models

adjustment for prognostic baseline covariates often leads to improved precision

Nonlinear Models

includes generalized linear models with nonlinear link functions

Logistic Regression

used in studies with binary outcomes; Nonlinear models such as logistic regression are commonly used; Randomization Does Not Justify Logistic Regression

Linear Model

Covariate adjustment through a linear model is an acceptable method for estimating the average treatment effect

Nonlinear Model

Covariate adjustment with nonlinear models is often used when the primary outcome is not continuous

Ordinary Least Squares

the model is estimated using ordinary least squares

Huber-White sandwich standard error

Agency recommends use of a robust standard error method such as the Huber-White sandwich standard error

Nonparametric Bootstrap Procedure

An appropriate nonparametric bootstrap procedure can also be used

Cochran-Mantel-Haenszel

Cochran-Mantel-Haenszel methods are acceptable for the analysis of clinical trial data with binary endpoints

Processes

1
Randomized Clinical Trials

Sponsors can adjust for baseline covariates in the analyses of efficacy endpoints in randomized clinical trials

Clinical Concepts

6
Prognostic Baseline Covariates

baseline covariates that are likely to be associated with the primary endpoint

Type I Error Rate

minimal impact on bias or the Type I error rate

Efficacy Endpoint

Sponsors can adjust for baseline covariates in the analyses of efficacy endpoints

Binary Outcome

analysis of clinical trial data with binary endpoints

Time-to-event Outcome

In trials with time-to-event outcomes, the hazard ratio can also be non-collapsible

Missing Data

The Prevention and Treatment of Missing Data in Clinical Trials

ICH References (2)

ICH E9

Statistical Principles for Clinical Trials; Discourages deterministic procedures due to high risk of bias; Notes that the use of Bayesian methods in clinical trials may be considered.

ICH E9(R1)

Statistical Principles for Clinical Trials: Addendum: Estimands and Sensitivity Analysis in Clinical Trials

Related MFDS Guidelines

Korean regulatory guidelines covering similar topics

See Also (8)

Adjusting for Covariates in Randomized Clinical Trials for Drugs and Biological Products | Guideline Explorer | BioRegHub