The Statistical Analysis and Research Programming (STARP) core aims to provide the Department of Surgery with high quality statistical and programming expertise, with an emphasis on collaborative, ethical, and reproducible research to advance healthcare innovation.
What We Do
Study Design and Grant Application
We assist with study design and grant preparation through the selection of outcome variables, statistical analysis planning, power analysis and sample size calculation, and random assignment of treatment conditions. It is strongly recommended that investigators include STARP at the beginning or an early stage of their studies to ensure optimal study design and a statistically appropriate data analysis plan.
Database Management and Preparation
We work closely with the investigators in Wisconsin Surgical Outcomes Research Program (WiSOR) on research projects that utilize large national health outcomes databases, as well as UW Health’s HealthLink database. We provide database management, maintenance, patient cohort extraction, and analytic data set preparation. Please see the WiSOR page of the Department of Surgery intranet for a complete list of outcomes databases available.
Statistical Analysis and Results Interpretation
We perform research hypothesis testing and data analysis using a variety of statistical methods. The choice of statistical method depends on the study design, type of outcome measure, and research question. The statistical methods we can apply include (but are not limited to):
• Analysis of Variance (ANOVA)
• Linear Regression
• Logistic Regression
• Generalized Linear Model
• Survival Analysis (Kaplan-Meier, log-rank test, Cox proportional hazards models)
• Nonparametric Statistics
• Meta-Analysis
• Longitudinal Data Analysis
• Multilevel Modeling
• Structural Equation Modeling
• Data Mining
• Machine Learning
• Model Selection and Variable Selection
We also interpret the analysis results and translate statistical significance to clinical implications.
Evaluation of Measurement
We evaluate psychometric properties of measurement instruments including survey questionnaires. We can assess reliability, validity, and factor structure of the instruments, and perform item and test analysis.
Manuscript and Presentation Assistance
We edit and provide feedback for manuscripts, abstracts, and presentations. We can also prepare a written description of statistical methods and generate summary tables and figures. We request the opportunity to review any publications which use work provided by STARP.
Consultation and Training
We provide consultation and training on study design, data creation and management (for both in house and national databases), statistical analysis methods, and statistical software programs (including SAS®, R, SPSS, SQL, and Stata). We also periodically hold lectures on statistical methods for residents, trainees, and other interested members of the department.
Publication and Authorship
STARP team members also welcome and appreciate the opportunity to receive co-authorship according to the guidelines of the International Committee of Medical Journal Editors (ICMJE). For a detailed overview of STARP’s authorship policy please refer to our STARP Guidelines page.