Department of Biostatistics
Development of Statistical Models
Statistical methods developed by biostatisticians in support of
ongoing cancer studies at The University of Texas M. D. Anderson
Cancer Center have included:
- statistical inference for self-designing clinical trials
- Bayesian strategies for monitoring multiple outcomes in clinical trials
- strategies for dose-finding and safety monitoring
- logistics of delaying accrual in Phase I clinical trials
- graphical methods for evaluating covariate effects in the Cox
proportional hazards model
- simulation study of hazard function estimation, including the
effects of optimal band-width selection and correction for
boundary effects
- assessment of long-term survival using the parametric
likelihoods for multiple, nonfatal competing risks and death
- employment of Bayesian optimal designs in population models of
hematologic data for bone marrow transplantation studies
- determination of appropriate sample size for the primary objectives
- efficient bootstrap resampling for Cox proportional hazards models
- GEE and Bayesian mixed models for time-to-event in family studies
Recent design and analysis projects have included:
- permutation tests for comparing marginal survival functions with clustered failure time data
- implications for clinical trial sample size from group sequential strategies in two-armed bandit problems
- estimating equations and Bayesian random effects modeling in the genetic analysis of age at menopause
- comparison of methods of measuring HER2 in metastatic breast cancer
- methods for epidemiologic evaluation of smoking and prostate cancer
- analysis of genetic susceptibility and survival in breast cancer
- multivariate models in an epidemiologic case-control study
- design considerations for efficiency in prostate cancer chemoprevention trials
- extensions and applications of event charts
- evaluating ten years of translational research in prediction of cancer development from oral leukoplakia
- clinical trial monitoring under the new FDA regulations
- parametric and non-parametric methods to analyze carcinogen effects in the distal and proximal regions
of the colon
- Bayesian applications to meta-analysis
- cost-effectiveness of testing for breast-ovarian cancer susceptibility genes
- two-stage designs for clinical trials based on safety and efficacy
- evaluating multiple treatment courses and effects in clinical trials
- standardized methods of measurement of the future liver remnant prior to extended liver resection
Last updated: September 30, 2002
For questions concerning these pages, contact the
Department of
Biostatistics Webmaster.
|