Related Experiment Videos
Stochastic curtailing for comparison of slopes in longitudinal studies
M Halperin1, K K Lan, E C Wright
1Department of Statistics, George Washington University, Washington, D.C.
Controlled Clinical Trials
|December 1, 1987
Summary
This study introduces a method for analyzing clinical trial data using stochastic curtailing, which efficiently compares treatment effects over time. It accounts for individual variations and missing data in large trials.
Area of Science:
- Biostatistics
- Clinical Trial Design
- Longitudinal Data Analysis
Background:
- Physiological function change rates are often used as surrogates for serious clinical outcomes.
- Analyzing longitudinal data with individual variability and missing observations presents statistical challenges.
Purpose of the Study:
- To implement stochastic curtailing for comparing slopes in two-treatment clinical trials.
- To provide a statistical framework for analyzing rate of change data with staggered entry and missing values.
Main Methods:
- Assumes a linear change in physiological function over time for each participant, with individual variations in slopes and intercepts.
- Applies stochastic curtailing (Lan et al., 1982) for one-sided slope comparisons.
- Accounts for staggered participant entry and randomly missing response data, assuming large sample sizes and at least two measurements per individual.
Main Results:
- The proposed methods enable efficient analysis of rate of change data in clinical trials.
- The framework accommodates common complexities such as staggered entry and missing data.
Conclusions:
- Stochastic curtailing provides a viable method for analyzing surrogate endpoints based on physiological rate of change.
- The approach is robust to variations in individual responses and data availability in large clinical trials.