Applications of the Wei-Lachin Multivariate One-Sided Test for Multiple Outcomes on Possibly Different Scales
Authors
Discussion
Loading...
Member takes
Implication
Methodological study demonstrates enhanced statistical power for the Wei-Lachin test across multi-scale endpoints, highlighting significant sample size reductions in clinical trials.
Key Points
- To describe the application, power, and sample size characteristics of the Wei-Lachin multivariate one-sided test for assessing simultaneous beneficial effects across multiple outcomes measured on diverse scales.
- Evaluated the one-degree-of-freedom Wei-Lachin directional test across combined outcomes including differences in continuous means, binary proportions, and time-to-event lifetimes.
- Calculated statistical power and sample size requirements compared against separate one-sided tests with multiple testing adjustments, multivariate omnibus tests, and composite endpoints.
- For two mean outcomes with unit variance and a correlation of 0.5, achieving 90% power via two separate one-sided tests at the 0.025 level required 64% more sample size, rendering the Wei-Lachin test 39% more efficient.
- The Wei-Lachin test proved 32% more efficient than a two-degree-of-freedom T2-like omnibus test.
- The multivariate test demonstrated superior statistical power compared to standard analyses based on single composite outcomes.
Cite This Study
John M. Lachin (2014) studied this question.