3 results match your criteria: "50296Yale University School of Public Health[Affiliation]"
Simulating time-to-event data subject to competing risks and clustering: A review and synthesis.
Stat Methods Med Res
February 2023
Department of Biostatistics, 50296Yale University School of Public Health, New Haven, CT USA.
Article Synopsis
- Simulation studies are essential for testing statistical models used in analyzing complicated survival data, especially when dealing with competing risks and clustering.
- The article guides researchers on generating competing risks data and inducing cluster-level correlation, specifically for randomized clinical trials involving binary treatments.
- It reviews various methods, including the frailty model, probability transform, and Moran's algorithm, to create complex survival outcomes while maintaining specified levels of correlation.
A comparison of analytical strategies for cluster randomized trials with survival outcomes in the presence of competing risks.
Stat Methods Med Res
July 2022
310058Department of Biostatistics, 50296Yale University School of Public Health, New Haven, CT, USA.
Article Synopsis
- The study investigates how different survival models perform in cluster randomized trials, especially when there are competing risks.
- It finds that using a sandwich variance estimator helps maintain accuracy with larger clusters, but can lead to bias in smaller samples.
- The analysis shows that the marginal Fine and Gray model often has better power than other models, particularly when competing events are frequent, and applies these findings to a real trial on elder injury reduction.
A Bayesian approach for estimating the partial potential impact fraction with exposure measurement error under a main study/internal validation design.
Stat Methods Med Res
March 2022
Department of Biostatistics, 50296Yale University School of Public Health, New Haven, CT, USA.
Article Synopsis
- The partial potential impact fraction measures how many disease cases could be prevented by changing people's modifiable behaviors while keeping other risk factors the same.
- When exposure data is inaccurate, it can skew the estimates of this fraction, which means we need ways to fix these errors.
- This study proposes a Bayesian method to adjust for these inaccuracies, particularly using data from the health professionals follow-up study, to accurately estimate the impact of changing diets on colorectal cancer rates.