Biostatistics and Clinical Outcomes

Statistical Program

The Biostatistics and Clinical Outcomes group provides statistical collaboration to clinical, genomic, and basic science research programs in the Department of Anesthesiology. We collaborate with investigators in the Anesthesiology Department to contribute to data query, experimental design, data analysis, and scientific interpretation of clinical and basic research studies. We participate in the development of study protocols and grant proposals, for which we help investigators to form research ideas, analytic strategies, and power estimates for the proposed studies. We also contribute to manuscript development, particularly the content of the statistical methods and the presentation of the results. Our mission is to promote research of the highest quality by encouraging application of sound statistical principles, by conducting accurate and reproducible data analysis, and by developing productive collaborations with clinicians, research faculty, and other members of the research community. We use the Ticket Track System to serve as a channel for communication between Anesthesiology investigators and our group, and project tracking. We encourage investigators to submit their request four weeks before the deadline to ensure sufficient time for communication, discussion, and analytical work before any deadlines.

Expertise: Our statisticians have expertise on analyzing a wide variety of data including clinical, genetic, and genomic data. Both Dr. Phillips-Bute and Mr. White have many years of experience in analyzing clinical data using a wide array of contemporary analysis methods involving statistical modeling, hypothesis testing, and parameter estimation to problems of data interpretation. They are also familiar with the clinical variables frequently used by investigators in the Anesthesiology department. Dr. Li has led many genetic studies for human complex diseases in her career. She has strong expertise in study design for genetic research, linkage screens, family-based association studies, genome wide association studies (GWAS), copy number variant (CNV) analysis, and gene expression studies. Our group works with the data analyst, Mrs. Betsy Hale, in Anesthesiology IT group closely to access the data needed for the research projects. Mrs. Hale is an experienced data analyst who is very knowledgeable with the clinical database environment at Duke. She leads the first line of dataset assembling and basic data quality check before delivering the dataset to the statistician or investigator.

Training Program

Part of our mission is to promote a sound understanding of statistical principles. We view the one-on-one exchange that clarifies relationships between study goals, experimental design, and statistical inference as a vital and rewarding aspect of collaboration. Effective collaboration, however, acknowledges that the value of quantitative methods must emerge from the context of investigation. To this end, we actively mentor research fellows and clinical faculty in the statistical aspects of their chosen project. Our program also participates in ongoing discussions of statistical issues at conferences and seminars with faculty, fellows, and medical students.

Research Program

Our members are closely connected with many clinical or basic research projects within the department. During the planning stage of a project, we contribute to power calculation and sample size estimation to assess the potential of the proposed study design. Our experience with the structure of different patient databases and protocols for populating them helps ensure the feasibility of retrospective studies and inform the design of prospective ones. We play a pivotal role in establishing the data analysis plan, conducting the statistical analysis, and generating summary reports and figures. We help the interpretation of research results and the dissemination of new insights through peer-reviewed publication.

We contribute to clinical research studies such as protection from perioperative organ injury, alternative analgesic techniques, and utility of transesophageal echocardiography. We continue to support basic research studies involving mechanisms underlying perioperative effects on long-term quality of life and genetic associations with cardiac surgery outcomes. In addition to these studies, we contribute to the development of research proposals in support of the department’s strong commitment to externally funded research. Over the past year, we worked on many research projects involving investigators at all levels. Our collaborative efforts have yielded many co-authored publications.

Effective collaboration requires critical evaluation and appropriate application of contemporary statistical methods. We emphasize constant learning of new statistical methodology and software as well as new method development motivated by research projects that we are involved. Areas of interests include applications of factor analysis to cognitive outcomes data, association methods for genome wide association data, family-based association method for sequencing data, and feature identification and classification in high-dimensional data such as microarray, proteomic, and metabolomics data.

We continue to serve as members of Data Safety and Monitoring Boards for several study protocols within the department. This ongoing vigilance toward unanticipated outcomes helps ensure safe care for patients while seeking the best treatments.

Goals

Our goals are to continue to promote greater understanding of statistical principles, and to conduct high quality data management and analysis that lead to high quality translational research at Duke. We will continue to improve our workflow on project tracking and analysis pipeline for genetic and genomic projects. We will continue to emphasize the documentation of our work to ensure the reproducibility of the research project. Particularly, for genomic research projects, we will ensure our conducts following the guideline of Duke translational medicine quality framework (TMQF).

Working closely with the Clinical Anesthesia Research Endeavors group, which oversees clinical research within the department, we look forward to engaging in new clinical and basic research initiatives and to enhancing awareness of statistical principles and their role in understanding quantitative relationships.

As research in the Department of Anesthesiology advances, we in the Division of Biostatistics and Clinical Outcomes will continue to apply our complementary mix of quantitative skills and strong commitment to effective collaboration to ensure the department’s statistical needs are met. We look forward to productive collaborations on challenging research initiatives.

Yi-Ju Li, PhD

Yi-Ju Li, PhD

Program Leader
Associate Professor
Department of Biostatistics and Bioinformatics
Duke Molecular Physiology Institute
Office: 919-684-0604

Biostatisticians:

Yi-Ju Li, PhD, Associate Professor
Mary Cooter, MS, Biostatistician II
Ishwori Dhakal, MS, Biostatistician II
Wenjing Qi, PhD, Postdoctoral Associate