Office for Research and Graduate Studies (ORAGS)

Biostatistics Core Facility

Scientific Objectives:  The mission of the Biostatistics Shared Service is to promote clinical, epidemiological, and laboratory investigations of cancer through applications of statistical science of the highest quality. The University of Maryland Greenebaum Cancer Center (UMGCC) statisticians strive to provide a central resource of state-of-the-art statistical science that is easily accessible to all members of the Program in Oncology. Members of the Service collaborate with study investigators in the design and analysis of (1) Phase I, II and III clinical trials of cancer; (2) Laboratory-based studies, including in vitro and animal experiments; (3) Epidemiological studies, including population studies on cancer disparities intervention research. Aside from the Shared Service activities, statisticians carry out methodological research in biostatistics to provide innovative solutions to problems arisen from investigations in the Cancer Center.  This is important to maintain our ability to provide statistical science of the highest quality. However, this activity is not explicitly supported by the Cancer Center Support Grant (CCSG) and not included in our request for funding.

Services and Technologies Provided:

The Biostatisticians collaborate with UMGCC investigators and provide them access to state-of-the-art statistical science, expert statistical analysis and sound statistical reporting in all areas of statistical experimental design, analysis and modeling and stochastic modeling, including:

Study Design. This applies to both laboratory experiments (e.g., xenograft models) and clinical trials. Statistical evaluation of sample sizes needed to achieve study objectives, endpoint and control selection, randomization and statistical power analysis and simulation.

Interim Analyses and Monitoring for Clinical Trials. Interim analyses consistent with statistical principles for multiple testing and sequential analyses.

Statistical Analysis. Analysis plan is provided in the protocol (before data collection takes place) and followed. Statisticians assess whether the data satisfy model assumptions implicit in the analysis software to ensure the most appropriate statistical methods are used for analysis.

Statistical Programming. Major statistical software are available including SAS, Splus, Stata, StatXact and East. The Biostatistics Shared Service also develops computer programs for complex statistical problems.

Biostatistics Training. Introductory seminars are provided on biostatistics tailored to cancer clinical investigators and fellows in hematology and oncology. The goal is to enhance understanding of biostatistical concepts and thus to improve clinical protocol.

Reports and Publications. Statistical design and analyses reports are communicated to investigators via memoranda. The memoranda for a project document study design, data source, statistical methods and analysis summaries and is written in a form appropriate for the statistical section for that project (e.g., a manuscript or a grant application).

Facilities:
The Biostatistics Shared Service is located on the 9th floor of the UMMC, N9E17

Specialized Software: for translational research and clinical trials are available to enhance our support of the cancer center’s research programs: DSTPLAN (developed by MD Anderson Cancer Center, TX) for sample size determination of common clinical trials with fixed sample designs; SCPRTbin and SCPRTnm are menu-driven programs that calculate phase II & III trial sample sizes and group sequential designs and their discordance probability based on exact binomial distribution and normal approximations (Tan and Xiong, Statistics in Medicine, 15:2037; Tan, Xiong, Kutner, Biometrics, 54:682). CombDesign is a computer program developed by Drs. Tan, Fang and Tian, to assist the experimental design for combination studies so that experimental resources are utilized most efficiently. SimV1 is an Splus program especially suited for analyzing percent data such as percent of apoptosis, bioavailability and percent changes (Song P, Tan M.  Marginal models for longitudinal proportional data. Biometrics 56:496-502, 2000.) Splus program to adjust p-value for making inference on secondary endpoint for a sequentially designed trial (Liu A, Tan M, Boyett JM and Xiong X.  Testing secondary hypotheses following sequential clinical trials.  Biometrics, 56: 123-127, 2000).  Cox regression analysis of multivariate failure time data (MULCOX2 by Lin: Computer Methods and Programs in Biomedicine, 40: 279-93, 1993).  XgStat is an Splus program for small sample inference for incomplete (including informatively censored) longitudinal data in xenograft models developed by Dr. Tan and his colleagues. (Tan M, Fang H, Tian G and Houghton P. Biometrics, Sept. Issue, 2002). It also includes a beta version for sample size planning/experimental design for combination studies (testing synergism) and a SAS macro for generating statistical analysis by directly importing an Excel spreadsheet that most scientists are familiar with, developed also by Dr. Tan and his colleagues.

Personnel:

Ming T. Tan, Ph.D., is professor of Epidemiology and Preventive Medicine and Head of the Division of Biostatistics of the University of Maryland Greenebaum Cancer Center (UMGCC). He has been on the faculty since 2002. He is a member of the center’s Experimental Therapeutics Program. He was previously a senior member (faculty) at St. Jude Children's Research Hospital Cancer Center and biostatistics director of St Jude's Developmental Therapeutics for Solid Malignancies Program (1997-2002), assistant and associate professor of Biostatistics and Epidemiology at The Cleveland Clinic Foundation (1990-1997). He received his Ph.D. in statistics from Purdue University, Indiana.

Hongbin Fang, Ph.D., is a UMGCC faculty biostatistician and an assistant Professor of Epidemiology and Preventive Medicine. After he received his Ph.D. in statistics from Hong Kong Baptist University in 1998, Dr. Fang worked on statistical methods on survival models as related to AIDS research at University of Missouri as a postdoctoral fellow and then joined the Department of Biostatistics at St. Jude Children's Research Hospital in 2000 as a postdoctoral research associate.

Olga Goloubeva, Ph.D., M.Sc., is a UMGCC senior cancer biostatistician. Dr. Goloubeva received her Ph.D. in engineering. She continued her graduate training in mathematical statistics and in 1999 earned a Master's degree in statistics from Dalhousie University, Halifax, Canada. She was assistant professor in the Department of Mathematics and Computer Studies at the Mount Saint Vincent University, Halifax, Canada (1993-1999). In 1999, she joined St. Jude Children's Research Hospital Cancer Center as a biostatistician, where she collaborated with investigators in virology, diagnostic imaging, and radiation oncology. In 2001, Dr. Goloubeva moved to Dana Farber Cancer Institute, Boston, MA, where she collaborated as a biostatistician in CLL Consortium, and the Leukemia Committee for ECOG, and the AIDS Immunology Group. She joined UMGCC in 2003 and has been involved in collaborative research throughout the center, including areas in breast, prostate, lung cancer, myeloma, radiation oncology, and cancer disparities.

Guo-Liang Tian, Ph.D., is a UMGCC faculty biostatistician and an instructor of Epidemiology and Preventive Medicine. Dr. Tian was trained under Dr. K.T. Fang, an internationally renowned statistical scientist in multivariate analysis and experimental design, at Hong Kong Baptist University. He earned his Ph.D. degree in 1998 from Academic Sinica and his postdoctoral degree from Peking University in 2000. He joined the Department of Biostatistics at St. Jude Children's Research Hospital Cancer Center in May of 2000 as a post-doctoral research associate. His research areas include generalized mixed-effects models for longitudinal data, hierarchical modeling, and applied Bayesian methods in biostatistical models.

Policies of Operation:

Statistical support is available to all UMGCC members. A high priority is given to peer-reviewed, funded research projects and the preparation of grant applications. Support for other projects will be provided as time permits. The initiation of biostatistics consultation and collaboration from the Shared Service is via completing a biostatistics project request form. A statistician will contact the principal investigator upon receipt of the completed form.  To forge collaboration, members of the Biostatistics Shared Service attend the cancer center program working meetings, seminars and journal clubs in the specific program they are assigned to support. The interactions from these meetings help the statistician understand the scientific questions and current issues in the field and provide the appropriate statistical method to address the right question.

Printer friendly Email this page