Evaluating six-step parsimonious quadratic logistic regression for risk prediction

Principal Investigator

Name
Stuart Baker

Degrees
Sc.D.

Institution
National Cancer Institute

Position Title
Mathematical statistician

Email
bakers@mail.nih.gov

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-676

Initial CDAS Request Approval
Jun 12, 2020

Title
Evaluating six-step parsimonious quadratic logistic regression for risk prediction

Summary
The goal of the study is to evaluate the six-step parsimonious quadratic logistic regression model for risk prediction (that I developed) using the NLST x-ray arm data as the training dataset, the PLCO lung cancer data as the independent validation dataset, and Yunnan Tin Miners Cohort Study data as the second independent validation dataset. Dr Ping Hu in BRG/DCP/NCI along with programmers from IMS created a NLST and PLCO data set for this type of evaluatino and has generously offered to share it with me if I get this approval.

Aims

Fit the risk prediction model to the NLST x-ray arm data and evaluate performance using PLCO lung cancer data and data from Yunnan Tin Miners Cohort Study.

Collaborators

Dr. Ping Hu. Biometry Research Group, Division of Cancer Prevention, National Cancer Institute