Biostatistics

The Biostatistics Core serves as the statistical resource for the SMH-IRC. It provides consultation in research design, statistical analysis, statistical concepts, database technology and management, and it takes the lead in network and computer systems development in each of the SMH-IRC Programs. Members of the Biostatistics Core meet with SMH-IRC investigators and project staff monthly to promote the development of new research protocols, to assure the quality of existing data collection and management procedures and to review manuscripts in preparation. Statistical assistance, coupled with a framework for its directed use, is designed to foster greater sophistication in measurement and clinical research methodology and to lead to more powerful testing of hypotheses in the investigation of heterogeneity and treatments in schizophrenia.

The tasks of the Biostatistics Core can be summarized under the broad categories of: 1. Statistics, experimental design and measurement, 2. Data Management 3. Network Administration and Computer Systems Support. Additionally, members of the Biostatistics Core publish in peer reviewed journals and make presentations at professional meetings.

Statistics

•Engages in statistical collaboration and consultations in statistics, research design, randomization, sample size, statistical power, the choices and strategies of statistical analyses and the assessment of the reliability of ratings and measurements
•Assists in formulating and expressing research hypotheses in statistical terms
•Consults in protocol planning, grants preparation, manuscript preparation, data analysis, and the monitoring of ongoing studies. A sign-off procedure for manuscript submissions is in place.
•Provides interpretations of statistical results and formats for presentation as posters, manuscripts and lectures
•Provides training and continuing education in statistical methods and statistical concepts
Statistical research interests include structural and functional imaging, recurrent events, longitudinal data analysis and analysis of multiple outcome measures in clinical trials.

Data Management

•Object-Relational Database Management System Design and Implementation: develops and maintains centralized clinical and research data management systems with semi-custom or custom enterprise applications using programming languages such as Visual Basic.
•Legacy Data Conversion: transfers legacy data into conventional database programs (e.g. KMAN to MS Access)
•Legacy Data Migration: converts existing data structures into relational database formats to be used in a protocol driven data management systems
•Quality Assurance and Quality Control: performs data cleaning, validation, reconciliation, and incident reporting
•Data Entry, Maintenance, Audit Trails: performs data entry, data edits/modifications, and maintains audit trails
•Data Extraction: Export data sets from the main repositories using SQL
•Report Design & Integration: creates automated and non-automated reporting objects for data solutions
•Data Archiving: storage of paper and electronic data sets from studies that are no longer being analyzed

Personnel

Primary statistical consultation and collaborations are undertaken by Robert McMahon, Ph.D. Julie Grim-Haines, M.A., is Assistant Director for Data Management, with primary responsibility for database management system design and implementation and direction of SMH-IRC data management activities. Eric Barr serves as primary data entry and quality assurance across the center. Ming Zhao, MPh.(ORP), Amie Elliott, M.A.(SRD), and Yang Yu, M.A.(TRU) serve as data managers and statistical consultants within their various programs. Under the direction of Dr. McMahon and Ms. Grim-Haines, they produce basic and inferential statistical analyses, graphics, and carry out data management operations for their programs. Robert Munson, M.A., Assistant Director for Network and Computer Services, has primary responsibility for network administration and computer services, assisted by Tony Knight, computer technician.

Printer friendly Email this page