xiang zhang
Email: abqb0f@r.postjobfree.com
Address: *** ****** *****, ******
City: Lexington
State: KY
Zip: 40503
Country: USA
Phone: 859-***-****
Skill Level: Assistant
Salary Range: $90,000
Willing to Relocate
Primary Skills/Experience:
See Resume
Educational Background:
See Resume
Job History / Details:
Xiang Zhang
300 Alumni Drive, Apt166
Lexington, KY 40503
abqb0f@r.postjobfree.com
EDUCATION Ph.D, Statistics, University of Kentucky, Lexington, KY, Expected May 2013
M.S, Statistics, University of Kentucky, Lexington, KY, August 2010
B.S, Statistics, University of Science and Technology of China, China, July 2008
Summary
Qualifications
* 2 years' experience in clinical research in an academic setting.
* 3 years' experience in Bayesian methods in an academic setting.
* Knowledge of various research designs and statistical methodologies in clinical
research.
* Experience in the design and development of statistical analysis plans (SAP).
* Ability to utilize/understand electronic data processing systems in gathering,
storing, managing, retrieving and analyzing large data sets.
* Ability to teach and clearly communicate statistical techniques in layman's
terms to non-statistical staff.
* Ability to plan and resource long-range projects while assuming responsibility
for daily details with a high degree of accuracy.
Computational
Skills
* Proficiency in SQL(Oracle/MySQL) with 3 years' experience.
* Proficiency in SAS/R with 5+ years' experience.
* Knowledge of SPSS, Matlab, Python.
* Proficiency in Microsoft Office tools.
WORK
EXPERIENCE
Research Assistant 05/2011 - current
Institute for Pharmaceutical Outcomes and Policy, University of Kentucky
* Work on the large health claims database, e.g., UnitedHealth Group Claims
Database (I3 Invision Datamart), Kentucky Medicare/Medicaid database. Con-
duct data extraction, manipulation, cleaning and validation using SAS MACROS
and SAS procedures as well as DATA step.
* Provide hands-on statistical analysis including descriptive statistics, logistic/probit
models, multivariate model regressions, Cox proportional hazard models, and
GLM models.
* Implement appropriate sample size and power calculations while assuring accu-
racy for data undergoing statistical analysis.
* Consult and assist project teams on selecting appropriate study design and
statistical methodology.
* Present research study results with corresponding statistical tables and figures.
* Assist in suggesting alternative analysis strategies when changes are required.
* Produce well-reasoned proposals and research reports, independently wrote sta-
tistical methodology part of 2 collaborative manuscripts, 1 published publica-
tion.
Teaching Assistant 08/2008 - 05/2011
Department of Statistics, University of Kentucky
* Primary instructor for undergraduate class containing 70 students, taught sta-
tistical concepts, statistical methods (different tests and OLS) and programming
(SAS/R), created syllabus, homework problems and exams.
* Recitation instructor for graduate classes containing 40 - 50 students and under-
graduate classes containing 70 students, in charge of the programming (SAS/R/SPSS)
and problem-solved sessions for the course.
PROJECT Value Based Insurance Design in pharmaceuticals 05/2011 - 11/2011
* Objective: To examine the effect of copay reductions for prescription medica-
tions, on prescription utilization and medication adherence as well as use of
other healthcare services.
* Data: about 10,000 objects from the UnitedHealth Group Claims Database.
* Methods:
- Designed the intervention group (patients who experience prescription co-
pay reductions) and comparison group (patients who do not experience
prescription copay reductions).
- Conducted several sensitivity analysis for different purpose, e.g. categoriz-
ing copay reductions to better specify price elasticity, excluding patients
taking multiple medications.
- Applied Mann-Whitney analysis to test the differences in the change scores
of prescription utilization.
- Applied logistic regression to predict post-period prevalence .
- Applied Cox regression to examine the length of therapy for both inter-
vention and comparison groups.
- Applied multivariate generalized estimating equation models to determine
the effects of prescription copay reductions on utilization and spending.
Acute liver failure (ALF) associated with acetaminophen prescribing and other risk
factors 11/2011 - 10/2012
* Objective:
- Primary Objective: to determine the incidence of ALF in patients pre-
scribed opioid/acetaminophen combination products versus opioids alone.
- Secondary Objectives: to determine the incidence of ALF in patients tak-
ing opioid/acetaminophen combination products prescribed with varying
daily doses of acetaminophen, as well as identification of risk factors for
developing ALF in patients prescribed these products.
* Data: over 2 million objects from the UnitedHealth Group Claims Database.
* methods:
- Helped Defining the including and excluding criteria in the study.
- Applied Cox regression/Logistic regression to determine the incidence of
ALF in patients prescribed opioid/acetaminophen combination products
versus opioids alone.
- Applied Cox regression/Logistic regression to determine the incidence of
ALF in patients taking opioid/acetaminophen combination products pre-
scribed with varying daily doses of acetaminophen, as well as identification
of risk factors for developing ALF in patients prescribed these products.
Evaluation of Pharmacological Management for Type 2 Diabetes Post-Myocardial
Infarction 11/2012 - current
* Objective: Previous research suggests elderly patients with type 2 diabetes mel-
litus (T2DM) are likely discharged without anti-hyperglycemic medication after
an admission for myocardial infarction (MI). Our study aimed to evaluate anti-
hyperglycemic medication use post-MI in a commercially insured population.
* Data: about 25,000 objects from the UnitedHealth Group Claims Database.
* Methods:
- T2DM was defined based on diagnosis codes (Chronic Condition Data
Warehouse definition) and pharmacy claims for an anti-hyperglycemic
drug.
- Identified patients who experienced MI (ICD-9 code 410.xx) during our
study period and evaluated anti-hyperglycemic treatment before and after
the cardiac event.
- Described medication utilization and investigated changes in medication
use from pre- to post-MI.
- Logistic regression and resulting odds ratios (OR) with associated 95%
confidence intervals (CI) were used to evaluate factors predicting treatment
post-MI.
PUBLICATION * Zhang, X. and Zheng, Y. (2012). A note on spatial-temporal lattice model-
ing and maximum likelihood estimation, Statistics and Probability Letters, 12,
2145-2155.
* Zhang, X. and Zheng, Y. (2012). Nonparamtric Bayesian inference of multi-
variate density estimation using feller priors, submitted to Journal of Nonpara-
metric Statistics.
* Amanda J. Sowell, George A. Davis, Xiang Zhang, Jeffery C. Talbert, Can-
dace J. Brancato, Heather Bush, Daniel A. Lewis. (2012). Acute liver fail-
ure associated with acetaminophen prescribing in combination opioid products,
Submitted to Annals of Pharmacotherapy.