Orthopaedic Biostatistics

Why Consult with a Biostatistician?

Meeting during the design phase of your project is strongly recommended to:
• Refine research question(s)
• Optimize the study design and outcomes for the question(s) being asked
• Identify validated outcome instruments, or plan instrument development and validation
• Ensure the planned data collection is amenable to analysis
• Identify the number of patients/samples needed

In addition to performing statistical analyses, the biostatistician will aid in:
• Interpretation
• Tabular and visual presentation of results for abstracts/manuscripts/presentations
• Drafting statistical methods sections and power/sample size statements
• Refining results sections to ensure proper formatting of descriptive statistics, p-values, etc.

Funding agencies and peer reviewers are increasingly scrutinizing statistical methods, and many now require or strongly prefer to see a statistician collaborator or co-investigator included on submissions.

To Initiate a New Project with Biostatistics

To initiate a new project with us, please complete the following intake form:

Intake Form

Plesase contact us for general questions, requests, or other information:

Timeline for Consultation

Note: These are estimates for typical projects. For projects requiring more complex/involved analyses or multiple iterations of analyses, turn-around time may be longer

Note: Power & sample size calculations must be completed BEFORE a budget is submitted to RMG (which should be done at least 1 month before the grant is due)

What to Bring When Meeting With Biostatistics

Analyses:
• Research question(s)
• Data set(s) [see data formatting below]
• Relevant literature

Power / Sample Size Calculations:
• Research question(s)
• Primary Outcome(s)
• Typically, mean & standard deviation or the proportion observed for your primary outcome(s) previously in similar cohort(s)/sample(s) from the literature
• Minimum clinically important difference (if available) or minimum desired difference to detect between groups
• Feasible sample size to enroll within a reasonable time frame

IRB & Sending Data Files

Your statistician will need to be added to the project IRB

Data should be de-identified if possible
• If sending data containing PHI via email, use the SECURE: label in the subject line
• Data can also be shared using Stanford Medicine Box
• What constitutes PHI: https://en.wikipedia.org/wiki/Protected_health_information

Data Formatting

Data Formatting
• Data sets should be in Excel or .csv format
• Data should typically be in one large table with each row containing a patient/sample, and each column containing a
variable/outcome
• Avoid special characters in variable names, make names concise but intuitive
• Formatting should be standardized (Example: “no”, “NO”, and “No” are treated as 3 different responses by statistical software, so ensure that capitalization and spelling are consistent)
• In a separate tab, include a data dictionary that contains information on the variables such as units, definitions, and/or calculations used
• Copy and paste as values any data derived from excel formulas after they are calculated to avoid errors creeping in if data is edited/moved; formulas should be listed in the data dictionary
• A good article on dataset formatting (Broman and Woo 2018): https://www.tandfonline.com/doi/full/10.1080/00031305.2017.1375989