skip navigation

Biostatistics Study Examples

Study Examples | Key Publications | Intro

These are among the current Group Health Center for Health Studies (CHS) studies on behavior change:

• Assessing the Impact of Colorectal Cancer Screening
• Statistical Coordinating Center for Breast Cancer Surveillance Consortium
• Understanding Variability in Community Mammography

Assessing the Impact of Colorectal Cancer Screening

• Principal Investigator: Carolyn M. Rutter, PhD
• Co-investigators: Diana L. Miglioretti, PhD; Margaret T. Mandelson, PhD; Jennifer B. McClure, PhD

Microsimulation models are computer models that operate at the level of a smaller individual entity, such as a person, but can simulate large representative populations to draw conclusions. This project began by building a new microsimulation model for the natural history of colorectal cancer (CRC). Now Center researchers are using this model to estimate the effect of screening, polypectomy, and changes in population risk (such as smoking and obesity) on CRC outcomes, gaining insights beyond those from either observational or randomized studies. Their goal is to use the model to compare the impact of CRC screening, polypectomy, and risk factors on both projected and observed population trends in CRC incidence and mortality within the United States population. In collaboration with colleagues in the Cancer Intervention and Surveillance Modeling Network (CISNET), they aim to compare the impact of screening—as practiced—with the potential impact of different guideline-recommended screening strategies on CRC outcomes. This project is funded as part of a four-year, $1 million grant from the National Cancer Institute (NCI) for studying the development, prevention, and treatment of colorectal cancer.

Top

Statistical Coordinating Center for Breast Cancer Surveillance Consortium

• Principal Investigator: Diana L. Miglioretti, PhD
• Co-investigators: William E. Barlow, PhD; Sebastien Haneuse, PhD; Linn A. Abraham, MS; Laura E. Ichikawa, MS

The Statistical Coordinating Center (SCC) organizes the collection, management, and analysis of data from seven mammography registries that make up the Breast Cancer Surveillance Consortium (BCSC), funded by the NCI. The SCC also develops and implements novel statistical methodology to characterize the process of breast cancer screening. It maintains a data management system that facilitates pooled analyses and rapid development of new projects. And it provides scientific guidance for publications, enhances communication and cooperation among BCSC members, and facilitates the use of BCSC data by investigators outside the BCSC. The BCSC aims to make breast cancer screening more effective at detecting tumors while limiting unnecessary follow-up due to false positive examinations. It provides population-based information on breast cancer screening, examines the role of practice patterns and biological factors on outcomes, and evaluates mammography’s efficacy in reducing breast cancer morbidity and mortality.

Top

Understanding Variability in Community Mammography

• Principal Investigator: Joann G. Elmore, MD, MPH, Harborview Medical Center; University of Washington Department of Epidemiology
• Group Health Co-investigator: Diana L. Miglioretti, PhD

Although mammography’s value has been proven, its efficacy depends largely on interpretations, which differ substantially among radiologists. In an attempt to enhance community performance of mammography, this study’s leaders are using hierarchical statistical models to understand the sources of variability among individual radiologists. These models are based on BCSC data from more than two million mammography encounters, as well as a survey of 400 radiologists to determine whether they understand concepts of numeracy (rates, risks, and probability) in general and specifically related to breast cancer screening. Survey data will be linked with BCSC data at an individual radiologist level. The researchers are outlining the individual characteristics of radiologists that are associated with interpretive performance. And they will evaluate the feasibility and impact of an interactive Web-based educational intervention and new audit reporting system for interpretation of mammography in the community.

Top

Study Examples | Key Publications | Intro