tjbencomo / survival-talk-pntlab

Best practices for survival analysis at PNT Lab

survival-talk-pntlab

Best practices for survival analysis given at the Mayo Clinic Precision Neurotheraputics Lab (Summer 2019).

survival_talk.pdf contains the lecture slides and tutorial.Rmd is a case study on prostate cancer with code.

Goals

The talk is meant to introduce non-statisticians to several topics needed for survival analysis including:

• Multivariable regression strategies
• Do's and don'ts of variable selection
• Maximizing sample size with imputation
• Interpretation and reporting of statistical results
• Awareness of problems with observational data (covariate imbalance, confounding, etc.)

Dependencies

Make sure the following R packages are installed before running the notebook

install.packages(c("rms", "rpart", "dplyr", "ggplot2", "mice", "stringr", "tidyr"))

About the Authors

Tomas Bencomo is a B.S. candidate in Computer Science at Stanford University. His research interests include using informatics to accelerate biological discoveries, building tools to improve physician decision making, and using evidence based medicine to evaluate medical practices.

Kyle W. Singleton is a Research Fellow at the Precision Neurotheraputics Lab at Mayo Clinic Arizona. He received his PhD in Biomedical Informatics from UCLA. His research focuses on radiomics methods using Magnetic Resonance Imaging (MRI) to guide treatment decisions for glioblastoma patients.