Return-to-baseline multiple imputation for missing values in clinical trials.

Area of Science:

• Biostatistics
• Clinical Trial Methodology
• Data Imputation Techniques

Background:

• Return-to-baseline is a common imputation method for handling intercurrent events in clinical trials.
• Existing methods can inflate data variability and introduce bias when missingness is related to observed outcomes.

Purpose of the Study:

• To establish criteria for effective return-to-baseline imputation methods.
• To propose a novel, improved return-to-baseline imputation technique.

Main Methods:

• Developed a new return-to-baseline imputation method based on defined criteria.
• Conducted simulations to compare the novel method against traditional approaches.
• Evaluated imputation performance based on data distribution, bias, and variance.

Main Results:

• The proposed imputation method results in completed data with proper distributions.
• Simulations demonstrate reduced bias and variance in estimators compared to traditional methods.
• The new method shows superior performance when missingness depends on observed values.

Conclusions:

• The novel return-to-baseline method addresses limitations of current practices.
• This approach provides more accurate and reliable data imputation for clinical trials.
• The method is easily integrated into existing multiple imputation software.