-- JessicaHeft - 14 May 2018

Jessica Heft, ObGyn/Urogyn

Background:
* Retrospective cohort comparing two surgical approaches (open vs. laparoscopic). Will be looking at perioperative complications and outcomes.
* Stage of project (select one): Design
* Data collection method (select one): Data are exported in electronic format
* Data management system (select one): REDCap
* Expected outcome (check all that apply): VICTR Biostatistics voucher
* Investigator experience (select one): Resident or fellow
*Discussion & Action Items:*
* Perfect confounding between surgeon and surgical technique.
* Jessica will coordinate with Thomas Stewart to develop a statistical analysis plan for submission of an application for VICTR voucher.

---

21 June 2018

E. Sample Size Justification and Statistical Analysis Plan

This is a retrospective cohort study to describe the rates and degree of surgical complications among the two surgical approaches. The sample size is fixed by the number of patients identified in the EHR.

A limitation of this study is that abdominal sacrocolpopexy with rectus fascia autograft is only performed by one surgeon, which means the study will not be able to conclude that similarities or differences in event rates are a result of the surgical approach or the operating surgeon.

However, the study will account for patient level characteristics by collecting factors related to disease severity (LIST HERE) and other known risk factors (LIST HERE). For binary endpoints, such as SSI, the covariate-adjusted odds-ratio of the outcome between surgical approaches will be estimated with a logistic regression model. For ordinal outcomes, like a complication severity score, the covariate-adjusted difference in means will be estimated with a linear regression model (or, if warranted, a cumulative probability ordinal regression model). The complexity of the regression models will be limited by the effective sample size, specifically no fewer than 10 effective samples per parameter. Data reduction of the covariates using techniques like principle components will be used as needed. Missing data will be handled using multiple imputation.

In order to conclude equivalence between surgical approaches for a specific outcome, the analysis will use 2nd generation p-values as follows: For each outcome, a range of event rates differences or mean differences will define equivalence. If the estimated difference for the outcome and its confidence interval fall within the range of equivalence, the analysis will conclude that the two surgical procedures are equivalent. If the confidence interval only partially falls within the range of equivalence, the analysis will be inconclusive. LIST A RANGE FOR EACH OF THE OUTCOMES