Biostatistics Branch Research Areas
Investigators in the Biostatistics Branch (BB) apply advanced statistical methods and data resources for many types of studies, described below.
Descriptive Epidemiology Studies
Descriptive epidemiology studies characterize cancer incidence and mortality temporal trends, age-specific rates, geographic distribution of cancer, race and ethnic differences in cancer rates, and birth cohort effects. BB investigators utilize novel analytic tools to obtain etiologic clues from descriptive studies, for example:
- Identify novel data sources through national and international collaboration (e.g. ACS, IARC, Danish breast cancer group)
- Integrate novel analytic tools to obtain etiologic clues from descriptive studies (e.g. Modeling age-period-cohort effects and etiologic heterogeneity)
- Perform record linkage studies, using SEER and other population-based databases (e.g. SEER-AIDS and SEER-Transplant studies, NHANES-NDI)
Analytical Studies
Increased participation in domestic and international consortia has yielded large-scale datasets and biospecimens, supporting various types of analytic studies:
- Cohort, case-control, and cross-sectional designs
- Incorporating chemical mixtures for characterizing chemical exposure in cancer risk prediction
- Genome-wide association studies and next-generation sequencing
- Assessment of emerging technologies and defining methods for integrating them into etiologic studies
- NGS, microbiome, metabolomics etc.
Collaboration Across Major DCEG Studies
BB investigators collaborate in major studies across DCEG, providing statistical and analytical leadership. Selected studies include:
- Agricultural Health Study
- Prostate, Lung, Colorectal and Ovarian Cancer (PLCO) screening trial
- NIH-AARP Diet and Health Study Cohort
- U.S. Radiologic Technologists Cohort
- Human Papillomavirus (HPV) Natural History Study
- HPV and Vaccine Dose Studies
- Non-Hodgkin Lymphoma Case-Control Study
- New England Bladder Cancer Case-Control Study
- Testicular Cancer Case-control Study
- Fanconi Anemia (an inherited bone marrow failure syndrome)
Translational Research
BB investigators strive to apply what is learned from research discoveries and findings to practical applications in support of public health. Examples of these efforts include:
- Development and validation of models for minority populations
- Refinement of screening guidelines based on HPV testing
- Evaluation of potential utility of genetic and other biomarkers
- Development and validation of cancer risk models for breast cancer risk tool, melanoma risk tool, colorectal, thyroid, and cervical cancer
- Evaluation of specific public health applications
- Risk-benefit calculations for the use of Tamoxifen
Methodological Research
BB maintains active methodological research programs in a wide range of topics that directly relate to analytic problems encountered in DCEG and at NCI, more generally. Investigators publish their methodological research in leading statistical, biostatistics, and bioinformatics as well as epidemiology, genetics, and clinical journals. Examples of research topics include:
- New approaches for cancer risk prediction that are used for counseling and screening recommendations
- Sampling-based designs and analysis approaches for generalizing the results of a cohort study to a representative population
- Development of new methods for assessing complex environmental exposure (possibly, longitudinally) on cancer incidence
- Novel approaches for the analysis of microbiome, metabolomic, and genomic biomarkers
- New methods for characterizing both germline and somatic mutational patterns in the genome
- Modeling natural history of cancer progression through precursors and stages of cancer using stochastic modeling
- Novel approaches for cancer surveillance using age-period-cohort and other methodology
- Statistical methods that account for errors in linkage when analyzing cohort data in which the outcomes are obtained through linkage to a large registry
- Novel design strategies for pooling or combining samples for disease screening or prevalence estimations