Methodologic research for stochastic models of colorectal cancer in natural history and under screening

Principal Investigator

Name
Hormuzd Katki

Degrees
Ph.D.

Institution
NCI

Position Title
Senior Investigator

Email
katkih@mail.nih.gov

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-308

Initial CDAS Request Approval
Sep 27, 2017

Title
Methodologic research for stochastic models of colorectal cancer in natural history and under screening

Summary
We propose general methodologic research to address shortcomings of stochastic models that represent transitions from healthy states to intermediate states to cancer, and to death. We plan to use the PLCO colorectal datasets, and especially the Study of Colonoscopy Utilization (SCU), as examples to apply our methodology to. This research is purely methodologic and will appear in journals for statistics, biostatistics, or epidemiologic methodology.

Aims

We plan to develop methods for addressing censoring of states due to interventions, measurement errors in state determination, effect of screening frequency and intervals, frailty in transition probabilities, modeling differences between the prevalence screen and subsequent incidence screens, and semi-Markov transition probabilities. All of these methods will incorporate available covariates such as age, race, sex, BMI, and others.

Collaborators

May Ishaq al-Husseini (Ph.D. student, Department of Statistics, George Washington University)
Qing Pan (Associate Professor, Department of Statistics, George Washington University)
Li Cheung (Staff Scientist, Division of Cancer Epidemiology and Genetics, NCI)
Rajeshwari Sundaram (Senior Investigator, Biostatistics and Bioinformatics Branch, NICHD)