Download resume.yihe.Oct2005 - Vanderbilt Biostatistics Wiki

Yi He
111 Acklen Park Dr. #C207
Nashville, TN 37203
615-275-7315(cell)
[email protected]
HIGHLIGHTS




Ph. D. Candidate in Biostatistics
Extensive involvement in the OPT (Obstetrics and Periodontal Therapy, Phase
III) clinical trial.
Strong mathematical and computing background
Relevant work experiences
STATISTICAL AND COMPUTATIONAL SKILLS


4-year professional statistical software- SAS system (BASE, STAT, MACRO,
GRAPH, IML, ODS)
Extensive statistical programming in R/S-Plus language (functions, modeling, and
packages).
ACADEMIC BACKGROUND
Ph. D. Candidate in Biostatistics
University of Minnesota, Minneapolis, MN
Relevant Coursework
Survival Analysis
Measure Theory
Probability Model (Stochastic Process)
Bayesian Decision Theory / Data Analysis
Analysis of Longitudinal Data
2002-Present
Clinical Trials
Mathematical Statistics
Linear Models
Spatial Biostatistics
Sequential Analysis
Master of Science in Biostatistics
University of Minnesota, Minneapolis, MN
2004
Relevant Coursework
Statistical Genetics and Human Genome
Categorical Data
ANOVA and Experimental Design
Correlated Data
Advanced Statistical Computing (SAS and R)
Master of Science in Applied Mathematics (Numerical Analysis)
University of Minnesota, Minneapolis, MN
2002
1
WORK EXPERIENCE
Biostatistician II
Department of Biostatistics, School of Medicine, Vanderbilt University
Statistical Intern (statistical consultant)
Vanderbilt-Ingram Cancer Center, Vanderbilt University
2005-present
winter 2004 & summer 2005
Research Assistant
2002-present
Division of Biostatistics, School of Public Health, University of Minnesota
Teaching Assistant
2002-2005
Division of Biostatistics, School of Public Health, University of Minnesota
Recitation Instructor
School of Mathematics, University of Minnesota, Minneapolis, MN
2000-2002
PROJECTS
Project: Validating Acute Lung Injury Markers for Diagnosis (VALID) 2005-present
The purpose of the study is to diagnose acute lung injury (ALI) / acute respiratory
distress syndrome (ARDS) in at risk patients and to identify patients with ALI/ARDS
who are at highest risk of adverse clinical outcomes.
Department of Medicine, School of Medicine, Vanderbilt University



Developing multivariable models relating clinical and biomarker characteristics
Predicting each patient’s response
Selecting relevant biomarkers for prediction
Project: Obstetrics and Periodontal Therapy (OPT)
2002-present
Randomized, single-blinded, multi-center, controlled, intervention trial. The purpose is to
determine if treatment of pregnant women with periodontitis affects gestational age.
Coordinating Centers for Biometric Research (CCBR), Division of Biostatistics,
University of Minnesota, Minneapolis, MN

Carried out statistical summary reports for monthly reports, biannual DSMB
(Data and Safety Monitoring Board) reports, such as enrollment, eligibility
information, and statistical models.
2



Carried out statistical modeling such as survival analysis, logistic regression,
Poisson regression, and correlated data analysis.
Attended DSMB meetings.
Created and maintained the official website for OPT study, such as daily updating
of open tables. See http://www.biostat.umn.edu/OPT.
Project: Bioassay Validation and Inverse Prediction
(Summer Intern, statistical consultant)
2005
Department of Biostatistics, School of Medicine, Vanderbilt University, Nashville, TN



Statistical consulting and modeling of dose-response data
Model selection / validation
Inverse prediction
Project: Study of Hyponatremia in Craniofacial Surgery Patients
2005
Department of Pediatrics, Vanderbilt Children’s Hospital, Nashville, TN




Study design based on the current enrolment, sample size, and historical data
Sample size calculation
Statistical analysis plan in the protocol
Interaction with investigators
Project: Real-Time Reverse Transcription Polymerase Chain Reaction (RT-PCR)
(Research Intern)
2004
Vanderbilt-Ingram Cancer Center, Vanderbilt University, Nashville, TN

Statistical modelling of Real-Time RT-PCR data using censored regression after
accounting for detection limits and measurement error.

The work was presented at Joint Statistical Meeting (JSM) in August 2005,
Minneapolis, Minnesota.
This was a joint work with Professor Bonnie LaFleur.
Project: Patient Satisfaction Study
2004
Randomized, controlled, intervention trial. The purpose is to determine whether use of
photo ID role cards can improve patients’ satisfaction of the physicians and hospital.
Department of Medicine, Hennepin County Medical Center (HCMC), Minneapolis, MN

Conducted study design in statistical point of view with the principal
biostatistician of the department.
3


Designed the block-wise Randomization algorithm. Within each color unit,
patients are randomized to a strategy arm by the time order of their visit to
hospital (HCMC) either in period 1 or period 2.
Calculated Sample Size by ANOVA F-test.
Project: High Blood Pressure Intervention Study
2004
Randomized, single-blinded, multi-center, controlled, intervention trial. The study’s
primary aim is to determine whether treating self-identified African American patients
with the special intervention will achieve a high rate of BP control at the levels specified
as (SBP <140 mm Hg and DBP <90 mm Hg for uncomplicated hypertension, <130/80
mm Hg in the presence of diabetes or chronic kidney disease).
Department of Medicine, Hennepin County Medical Center (HCMC), Minneapolis, MN



Designed the Randomization Schemas. All randomizations will assign
participants to study group in a 1:1 ratio, within randomly permuted blocks of
sizes 4, 6 and 8.
Calculated Sample Size by approximate 2-sample t-test.
Participated in the statistical analysis plan design including primary analysis and
secondary analysis.
Project: International Variability of Childhood Cancer in Relation to Diet
2003
The primary purpose of the study is to generate hypothesis about the association between
various childhood cancer incidences and nutrition consumption in different countries.
Division of Epidemiology / Clinical Research, Department of Pediatrics, University of
Minnesota, Minneapolis, MN
 Used the Food and Agriculture Organization (FAO) database and generated
nation-level diet data for forty-four countries using R.
 Conducted Poisson Regressions.
THESIS WORK
Master thesis: Multivariate Normal Mixture Models in Cluster Analysis to Detect
Gene Differential Expression from Microarray Experiments
2003-2004
Advisor: Dr. Wei Pan
Dissertation: Improved Markov Chain Monte Carlo Methods for Variance
Components in Hierarchical Linear Models
Advisor: Dr. James S. Hodges
4
PUBLICATIONS AND RESEARCH PROJECTS




He, Y., Pan, W. and Lin, J. (2005). Cluster Analysis Using Multivariate Normal
Mixture Models to Detect Differential Gene Expression With Microarray Data.
submitted.
He, Y. and Hodges, J.S., and Carlin, B.P. (2005) Re-considering the
parameterization in Hierarchical models. submitted.
LaFleur, B. and He, Y. (2005) A likelihood-based approach to modeling real-time
RT-PCR data with right censoring. To be submitted soon.
He, Y. and Hodges, J.H. (2005) A note on the posterior estimate for variance
parameters in Hierarchical models. To be submitted soon.
Working papers



He, Y., LaFleur, B., and Billheimer, D., and Hodges, J.S. (2005) Bayesian
analysis of real-time RT-PCR data.
He, Y. and Hodges, J.S. (2005) Modeling covariance matrices using simplex
parameterization in shrinkage regression model.
He, Y., Hodges, J.S., and Xiao, G. (2005) Modeling micorarray data using
biological information.
HONORS AND AWARDS



Biostatistics Fellowship, Division of Biostatistics, 2004
The Graduate Scholarship, Graduate School, 2000-2004
Summer Fellowship, School of Mathematics, 2001, 2002
PROFESSIONAL MEETINGS



OPT study DSMB meeting
Radisson Hotel, Minneapolis, MN.
International Biometric Society, East Region (ENAR)
Hilton Hotel, Austin, TX
Joint Statistical Meeting, American Statistical Association
Minnesota Convention Center, Minneapolis, MN.
October 2004
March 2005
August 2005
PRESENTATIONS

He, Y. and Hodges, J.S., and Carlin, B.P. (2005) Simplex parameterization in
Hierarchical models.
ENAR, 2005
5



LaFleur, B. and He, Y. (2005) A likelihood-based approach to modeling realtime RT-PCR data with right censoring.
JSM, 2005
PROFESSIONAL MEMBERSHIP
American Statistical Association (ASA)
ENAR
6