Methods in regression analysis in surgical oncology research-best practice guidelines.

Area of Science:

• Surgical Oncology
• Biostatistics
• Medical Research Methodology

Background:

• Highlights the importance of regression modeling in surgical oncology.
• Addresses common challenges and pitfalls in applying regression techniques.
• Emphasizes the need for best practice guidelines in research.

Purpose of the Study:

• To provide practical strategies and demonstrate best practices for linear and logistic regression.
• To identify predictive factors influencing patient outcomes in breast reconstruction.
• To illustrate potential pitfalls in regression modeling within surgical oncology.

Main Methods:

• Utilized linear and logistic regression models on a cohort of 1986 patients undergoing tissue expander breast reconstruction (2019-2021).
• Assessed factors affecting BREAST-Q Physical Well-Being of the Chest (PWB-C) scores at 2 weeks using linear regression.
• Evaluated predictors of overall complications and malrotation with logistic regression, including model fit and performance assessment.

Main Results:

• Linear regression identified age, single marital status, and prepectoral pocket dissection as significant predictors of PWB-C scores.
• Logistic regression indicated that BMI, age, bilateral reconstruction, and prepectoral dissection were associated with an increased likelihood of complications.
• Specific statistical significance (p-values) and effect sizes (β or OR with 95% CI) were reported for all identified factors.

Conclusions:

• Provides clear directives for the effective use of regression techniques in surgical oncology.
• Recommends researchers use clinical judgment for variable selection and interpret model results with clinical plausibility.
• Stresses the importance of confirming appropriate model fitting for reliable regression analysis in surgical research.