During an initial meeting, we will help clarify your research question(s) and suggest enhancements to further your research. Collaboratively, we will define the study design including potential sources of bias, and determine the scope of analyses and resources needed. Often these meetings result in introductions to other collaborators both within and outside the institution that may be helpful in pursuing external grant funding.
Additionally, you may be asked to review the literature for pertinent statistics for conducting power analyses. Our collaborations are most effective to investigators if begun during the planning phase; this process will help focus your protocol, expedite IRB approval, guide your data management, and avoid unnecessary complexities in statistical analyses.
Once the study design is established and information for the power analysis is retrieved, at a subsequent meeting we will provide sample sizes tailored to your research question(s). We will finalize the statistical methodology section of your protocol that includes the power analysis and analysis plan, and provide guidance for the next steps of your project. For clinical trials that require data and safety monitoring boards (DSMBs) for interim monitoring, we will help to identify experts to serve and suggest stopping criteria for safety and efficacy.
If you plan to collect primary data, we will help you devise data collection forms, review content, and provide instruction in data entry in Access or Excel, or alternatively a web-based data entry system. We will collaborate to define checks for validity, consistency, and duplicate records, and establish a procedure to document data to be queried and an audit trail of changes. Collectively, we will target dates for final data entry and final database revision. Our staff will document security checks and a data management plan.
If you are planning to conduct a randomized clinical trial, we will prepare a randomization scheme and instructions for use. If you are conducting secondary analysis of primary data using government or other databases, we will facilitate electronic transfer to a compatible format for analysis, review the database(s) for validity and consistency, and define a codebook that includes a narrative of the content and history of the database and explanations of variables and definitions of codes. Finally, once we provide results, we will help you to best present the data visually using state of the art graphics software.
The biostatistician will prepare the analytic strategy and statistical programming of the database that will include descriptive statistics, and measures of reliability, association and difference. Statistical methods depend on the research question(s), study design, level of measurement and other characteristics of outcomes.
In the process of analyzing the data, questions may arise regarding further validity checks of the data, violations of statistical assumptions, unexpected distributions of missing data, and need for sensitivity and/or subgroup analyses. The investigator should be prepared for further questions and explanations! Once we derive the study results, we will annotate/highlight the computer output with descriptions of what is significant, corresponding p-values and statistical methods, and would be happy to go through results with you after you have had a chance to review.
We will analyze your pilot data for the grant application, and provide the statistical methods and summary of results. With regard to government grant applications involving training and mentoring, we will draft a section on innovative training approach, teaching platforms and mentoring experience.
For research studies supported by the grant, we will formulate the study design and include design features that attempt to minimize bias and maximize security, integrity and validity of the data. As part of every grant application, we will draft an evaluation plan to include appropriate metrics and intervals of assessment. Depending on the scope of the grant, we will collaborate on defining our role as Key Personnel (Co-Investigator, Co-Principal Investigator or other), and percent of effort.
The best way to avoid reviewers’ comments regarding the statistics is to collaborate on the research project from its inception (see above). If we had not collaborated on the study design, statistical analysis, reporting and/or interpretation of results, collaboratively we work to allocate resources to review all aspects of the study as well as the entire manuscript. Authorship is negotiated based on criteria established by the International Committee on Medical Journal Editors. Additionally, we have experience assuming the lead role in drafting a manuscript.
In addition to collaborations on research project which necessarily result in opportunities to learn, we can support our academic units by providing seminars and workshops tailored to the interest of attendees, and by conducting journal club sessions and mock grant reviews.