BIOSTATISTICS UNIT ACTIVITIES

 

The Unit activities are broadly related to problem specification, study design, data collection and quality control, data processing, statistical modeling and analysis, and interpretation of statistical results.  The following describes more fully the individual components of the Unit support services.

 

Problem Specification:

 

The Biostatistics staff is experienced in consulting with investigators, assisting in the  identification of research parameters and goals. Below is a sample of the services provided:

 

§  Formulate statistical and mathematical models for genetic, epidemiologic and other biomedical phenomena under study

§  Specify the hypotheses to be tested and parameters to be estimated

§  Review and interpret past experiments and literature reports

§  Conduct meta-analyses when literature based reviews require quantitative summarization

 

Study Design and Methodology:

 

§  Assist investigators in the preparation of grant and contract proposals

§  Determine the population to be studied and the necessary background information to be gathered

§  Specify the dependent and independent variables to be measured

§  Identify nuisance variables and methods to control for them statistically

§  Choose appropriate study designs (longitudinal, cross-sectional, case-control,  cross-over, balanced incomplete blocks, population based and family based genetic studies, affected sib pairs with simplex or multiplex ascertainment, DNA microarray designs, etc.)

§  Calculate sample sizes required to achieve sufficient statistical power for testing hypotheses

§  Specify methods of assigning experimental units (i.e. patients, animals, lab techs.) to treatments including blinding of participants where appropriate 

§  Develop a protocol and manual of operations to minimize errors in data collection and avoid operational changes which might compromise interpretation of study results

 

 

Data Collection and Monitoring:

 

Project Start-Up

§  Advise and/or train the investigator’s data collection staff 

§  Carry out pilot trials in order to test or refine questionnaire phrasing and response categories, and detect unanticipated problems

§  Conduct focus groups to obtain preliminary data

§  Assist in setting-up uniformity trials to test methodology

 

 

     Data Collection

 

§  Design, develop and implement procedures for computer scanned data forms as well as web-based data collection

§  Administer mail, telephone, and/or personal interview surveys of defined populations

§  Customize computer-assisted telephone interviewing (CATI) system for telephone surveys

§  Develop customized programs for the Interactive Voice Recognition (IVR) system and test them prior to delivery 

§  Subject registration and randomization (stratified block allocation, Simon and Pocock’s minimization, and Raghavarao’s distance function)

§  Provide technical support for the development of contract specifications for external vendors and data collection oversight of larger scale data collection efforts

 

Project Monitoring

 

§  Institute quality control audits to ensure the accuracy of data

§  Review proposed protocol changes to consider possible effects on subsequent analysis and interpretation of study results

§  Participate in Data Safety Monitoring Committees and generate periodic progress reports

 

Data Entry

 

§  Code and enter data using double entry with verification

§  Create and archive data files for analysis

§  Design and implement database systems using MS Access for both local and remote data entry

 

Data Processing and Software Development:

 

§  Create code books describing data file format

§  Summarize original data for preliminary analysis and development of additional hypotheses of potential interest

§  Produce histograms, scatter plots, and other graphical displays where appropriate

§  Prepare data files for use with a variety of specialized software programs (e.g. conversion between various file formats, determination of affected family based controls (AFBACs), transmitted and non-transmitted alleles, etc.)

§  Develop, test and implement specialized software programs for data acquisition, processing, statistical modeling, and simulation studies

§  Investigate new biostatistical software packages that may enhance the analysis of a given study

 

Statistical Analysis:

 

§  Perform analysis using commercial statistical software such as BMDP, SAS, SPSS, SPLUS, SUDAAN, PC-NonLin, EQS, AMOS, EGRET, HLM, SOLAS, NQUERY, GENEHUNTER, SAGE, LINKAGE, and CART, as well as special-purpose statistical programs developed by the Biostatistics  Staff.

§  Develop hypothesis-based models

§  Refine model formulations and investigate alternative models

§  Perform computer intensive analyses and simulations to assess model robustness

 

Statistical Interpretation:

 

§  Prepare and present written reports summarizing the results of analyses including tables, charts, and graphs

§  Assisting in the preparation of manuscripts, reports, and presentation materials

 

Authorship:

 

Over the years, the Unit has attempted to foster a collaborative atmosphere within the University community.  The contribution of each researcher needs to be evaluated as a manuscript is prepared.  Consideration for authorship should be based on the accepted criteria for most journals that generally cite both study design and statistical analysis as intellectual input sufficient for authorship.  While it is impossible to define every situation in advance, consideration should be given and it should be noted that reimbursement for time does not preclude or replace authorship.