Overview

The Biostatistics Branch develops statistical methods and data resources to strengthen observational studies, intervention trials, and laboratory investigations of cancer. The Branch also provides information on cancer rates for generating etiologic leads and developing appropriate study designs, and it plans and conducts independent and collaborative descriptive and analytic studies of cancer etiology. Members of the Branch consult and collaborate with scientists throughout the NCI and NIH, and with investigators and public health officials at other government agencies and academic and research institutions in this country and abroad.

The Branch offers opportunities for post-doctoral research in statistical methods for epidemiologic research, genetic epidemiology, and risk assessment, and for projects in analytic and descriptive epidemiology. Post-doctoral fellows also have an opportunity to collaborate on important epidemiologic studies, many of which include molecular and genetic components.

Methodologic Research

Branch scientists develop methods for increasing efficiency and improving analysis of case-control and cohort designs. High priority is being given to methodologic issues in genetic epidemiology. Recently, Branch members developed methods for estimating risk associated with an identified autosomal dominant gene from family studies in which one member volunteers to be genotyped and to provide disease histories of first-degree relatives (kin-cohort design). Other research focuses on various methods for characterizing exposure history and overcoming problems associated with measurement error and missing information.

Descriptive Studies of Cancer

The Branch studies geographic patterns and temporal trends in cancer incidence and mortality rates to generate etiologic hypotheses, evaluate consistency with other hypotheses, and identify cancer sites warranting special study. Etiologic leads may also be suggested from studies of temporal trends in cancer subsites and histologic subtypes. Branch members develop statistical methods to interpret geographic variation, which they apply in collaborative studies to explain variations in breast cancer rates due to differences in the prevalence of known risk factors. Recently, the Branch published a U.S. atlas of cancer mortality data spanning the years 1950 to 1994, which identified geographic areas with high rates of specific cancers and spatial and temporal patterns that may suggest particular carcinogenic exposures. The atlas is available in printed and electronic formats (http://www3.cancer.gov/atlasplus/). Branch scientists also develop statistical methods to analyze associations of age, calendar time, and year of birth with cancer rates. These age-period-cohort analyses found a decrease in risk of breast cancer among black and white women born since 1948, despite later age of childbearing than in earlier cohorts

Cancer Risk Assessment

A major research thrust within the Branch is quantifying risk of developing cancer from exposure to known carcinogens. Studies of underground miners yielded risk estimates of lung cancer from exposure to radon, and data from a combined analysis of 11 such studies worldwide indicated that 10-12% of all U.S. lung cancer deaths each year may be attributable to indoor radon exposure. A recent analysis of data from eight case-control studies of exposure to radon in the home yielded risk estimates consistent with those from studies of underground miners. Branch scientists also developed a model for predicting the risk of breast cancer over a defined time interval for a woman with specific risk factors, such as a strong family history. This model has been used to assist in counseling women and to define eligibility criteria for intervention studies, such as the NCI Breast Cancer Prevention Trial. The model has also be used to help assess risks and benefits of tamoxifen for preventing breast cancer in high risk women. Recent work quantified the risk of melanoma from exposure to ultraviolet radiation based on personal history of residence.

Collaborative Studies

Branch investigators are key participants in large, complex interdisciplinary studies of cancer etiology and risk assessment. In a recent example, they were major collaborators in a case-control study of children with acute lymphoblastic leukemia, which found that risk was not associated with exposure to electromagnetic fields (EMF). Statistical elements of the study included assessing the validity of exposure measurements, designing and analyzing reliability studies, and evaluating dose-response relationships. Branch investigators led analyses of hormone assays to determine their reliability for studies aimed at understanding the etiology of hormone-related cancers. Branch investigators collaborate on more than one hundred interdisciplinary projects with scientists at the National Cancer Institute and worldwide.