VTCSOM Mentorship & Collaboration
VTCSOM provides medical students with one-on-one mentorship from a CBHDS collaborative biostatistician to support their Research Domain projects.
Interior photos of Virginia Tech Carilion School of Medicine and Research Institute
CBHDS Zoom Drop-in Hours
(Open to VTCSOM)
VTCSOM Specific Hours
Tuesdays & Thursdays:
2pm-4pm
Public Hours
Mondays: 10am-2pm
Wednesdays: 3pm-7pm
M1 students should initially contact CBHDS, ideally by the end of January of their first year at the latest to prepare for their prospectus. Sign up for the Zoom drop-in hours above or email biostats@vt.edu with a brief description of your project.
M1-M4 students should meet with their CBHDS co-mentor regularly as their project progresses.
Collaborators need sufficient lead-time to discuss the project, contribute to the project's proposed methods and aims, and conduct any power/sample size calculations.
Research Domain projects are student driven. Students are responsible for contacting mentors as needed throughout the course of the multi-year project.
Students should...
Consult statisticians early in a project. This can prevent the collection of unusable or un-interpretable data, and can avoid fatal flaws in the ability to execute the research project.
Involve primary mentor in the early planning of the study and continually throughout the course of the three-year project.
Work with assigned CBHDS co-mentor within a team science model, through all phases of the project.
Perform their own data cleaning, generate their own coding, run their own data analyses, and include their primary mentor in the decision-making process.
Understand and defend the final choice of statistical methodology.
Interpret analysis results appropriately and translate technical findings into lay language.
Discuss authorship with their CBHDS co- and primary mentors. Do not add anyone as an author without their discussion and approval. Collaborators need the option to decline or approve authorship of final manuscripts.
Be accessible to students outside of scheduled class and clinic times.
Train, mentor, and educate students in the areas of study design, data management, analysis, and interpretation.
Work with students and their primary mentors to finalize study goals and aims.
Mentor students through the task of creating a statistical analysis plan (SAP).
CBHDS co-mentors will not actively carry out programming and analysis tasks on students’ behalf*.
*Exception for power calculations, which will be outlined with the student and primary mentor, but carried out by CBHDS.
CBHDS provides VTCSOM research faculty with a monthly summary of all SOM student interactions, including both Research Domain-related and non-Research Domain interactions.
CBHDS co-mentors will advise in the following areas as needed:
Operationalizing measures
Data collection instrument selection and development
Choice of study design type
Power and sample size estimation
Randomization schemes
Inclusion and exclusion criteria
Tailored survey/questionnaire design
REDCap database design & building (VT)*
Create processes for data quality checks
*Projects through Carilion must work with the HART team for REDCap builds.
Understand underlying assumptions
Identify potential control variables
Prepare a comprehensive analysis plan
R or SAS programming to support data cleaning and wrangling, preliminary analyses, data visualizations, primary analyses, assumption checking, sensitivity analyses, etc.
Types of analyses we support include (but are not limited to):
- Biostatistical Analyses: Descriptive statistics, regression models, etc.
- Next Generation Sequencing (NGS) Analyses: Differential expression analysis, pathway analysis, epigenomic profiling, etc.
- Machine learning and deep learning applications and techniques: Medical image analysis, Large Language Model (LLM) refinement, etc.
Students should be able to demostrate a clear and accurate understanding and interpretation of methods, results, and conclusions.
CBHDS co-mentors will mentor students as they prepare for project presentations and will attend all presentations related to the Research Domain projects.
VTCSOM students are expected to conduct their own programming to complete their project, with guidance from CBHDS. This includes independent work to gain and apply these skills, and help from CBHDS mentors when needed. Practicing and applying your skills outside of CBHDS mentor meetings ensures time spent with mentors is productive.
We discourage the use of excel and encourage the use of R (or other statistical programming software).
Developing and retaining syntax/code ensures that all steps involved in data cleaning, wrangling, analysis, and visualization are documented and preserved to ensure research is transparent, reproducible, and robust.
Programming also reduces the risk of human error and more easily facilitates identifying and correcting any errors or mistakes.
Any and all checks and changes made to a dataset should be documented such that someone could reproduce exactly what was done if handed the same raw dataset, documentation, and syntax/code.
Comments should be included thoughout the syntax to describe all programing steps.
Organization is key. Take the time to prepare folders and employ your organizational skills to manage files and code throughout your pogramming.
R Programming Resources: When getting started with R, we recommend the following. Most of programming is copy/pasting and tweaking code as needed for a project's specific data and aims. These are useful to reference when working with your data.
Biostatistics Resources: For a refresher on biostatistical and epidemiogical concepts, we recommened the following:
Students and their primary mentors (if possible) should reach out to meet with CBHDS at the outset of the study to describe the project and potential data sources.
The initial interaction with the paired co-mentor will orient CBHDS to the project and orient the student to CBHDS and their work style.
To initiate the process, students should sign up for the Zoom drop-in hours above or email biostats@vt.edu with a brief description of your project.
Three things will take place in the initial meeting(s) with the student and primary advisor:
- Outline study aims.
- Identify general statistical support needed.
- Assign a CBHDS co-mentor to the project.
Students should sign up for drop-in hours (see above), preferably with their paired CBHDS co-mentor.
Students should log-in to their vt.edu account to sign up for Zoom drop-in meetings. These are limited to 30 minutes per person per day when demand is high.
If students need to meet outside of drop-in times, CBHDS co-mentors and students will coordinate their schedules.
The student and CBHDS co-mentor will discuss their preferred scheduling method once paired. In-person appointments are also an option. Students should email the CBHDS co-mentor if in-person meetings are preferred.
Examples of the types of questions appropriate for drop-in consulting hours include debugging code or interpretation of output. Decision making regarding statistical methods may require a longer dedicated meeting as it is typically a more in-depth discussion/process.
CBHDS follows authorship guidelines from the International Committee of Medical Journal Editors (ICMJE)
Students should ask and discuss authorship with their CBHDS co-mentor and primary mentor.
Some projects rely heavily on CBHDS support and co-authorship is appropriate.
For projects that minimally involve CBHDS, consider acknowledging CBHDS and any specific collaborators.
Do not add anyone as an author without their discussion and approval. Collaborators need the option to decline or approve authorship of final manuscripts.
CBHDS will support medical students in non-Research Domain projects. This support can take place through...
- short-term drop-in consulting hours, or
- collaborative relationship within a team science model that spans from study design through dissemination.
Funding is required for projects involving CBHDS collaboration on non-Research Domain projects beyond the publicly available drop-in hours.
CBHDS includes experienced Master's and PhD level data scientists, biostatisticians, epidemiologists, and bioinformaticians.