Making inferences about projected completors in longitudinal studies.

Area of Science:

• Biostatistics
• Clinical Trials Methodology
• Health Services Research

Background:

• Attrition is a significant challenge in clinical trials, potentially biasing results.
• Existing methods often focus on the entire study population, overlooking specific subgroups.
• Disability research in severe head injuries requires methods to analyze data from individuals completing the study.

Purpose of the Study:

• To develop a methodology for making inferences about projected completers in studies with attrition.
• To address scientific questions focused on participants likely to complete a study or remain on protocol.
• To improve the analysis of clinical trial data, particularly in neurological injury research.

Main Methods:

• Proposing the use of individual dropout inclination measures to identify projected completers.
• Developing a stratified response model based on projected completion status.
• Employing maximum likelihood and restricted maximum likelihood for parameter estimation.

Main Results:

• The methodology allows for targeted inferences on individuals likely to complete the study.
• Prediction measures and accuracy evaluation procedures for observed dropout are presented.
• The approach was illustrated using data from a clinical trial for severe head injury patients.

Conclusions:

• The proposed methodology effectively identifies projected completers in the presence of attrition.
• This approach enhances the ability to answer specific scientific questions in clinical trials.
• The method offers a valuable tool for analyzing disability data in severe head injury populations.