This post hoc analysis focuses solely on a Trial Sequential Analysis (TSA) of the results comparing Quadratus Lumborum Block (QLB) versus intrathecal morphine. TSA is a statistical method used to determine the required sample size to reliably confirm the effectiveness of an intervention, ensuring that no further trials are necessary. By applying TSA to the current data, we aim to assess whether the existing evidence is sufficient to draw a definitive conclusion on the efficacy of QLB in comparison to intrathecal morphine, thus avoiding the need for additional studies.

The analysis performed using TSA Copenhagen is based on the Sequential Trial Analysis (TSA) model, specifically using the sequential boundaries by Lan and DeMets, as described in the 1983 Biometrika article “Discrete sequential boundaries for clinical trials” (Lan KKG, DeMets DL, 1983). TSA allows for the adjustment of sequential boundaries and determines if a trial can be stopped early, either for efficacy or futility, based on accumulated data throughout the clinical trial.

Regarding the importance of performing this type of analysis, here are several key points to highlight:

Sample Size Calculation in Sequential Trials: The “theoretical” or maximum sample size needed in TSA is calculated under the assumption that the trial may be stopped at any point during the sequential stages if an efficacy or futility boundary is crossed. This makes the calculated sample size more efficient than a traditional fixed design, where the sample size is determined without considering the possibility of early stopping.

Heterogeneity Correction in Meta-Analysis: TSA accounts for variability or heterogeneity between studies in a meta-analysis, which is crucial for obtaining accurate estimates of treatment efficacy. The heterogeneity correction improves the robustness of the results, adjusting sample size calculations and the probability that the treatment is effective. This is especially important when combining studies with different populations, methods, or conditions.

O’Brien-Fleming Spending Function: In TSA Copenhagen, the O’Brien-Fleming alpha (Type I error) spending function is applied across the sequential stages. This function adjusts the critical alpha value at each stage to control Type I error while performing interim analysis on the accumulated data. In this way, sequential analysis can be conducted without compromising the trial’s validity over time.

Importance of TSA Version 0.9.5.10 Beta: TSA Trial Sequential Analysis Viewer Version 0.9.5.10 Beta is a powerful tool that facilitates the analysis of sequential trials, enabling the implementation of these advanced methods to estimate efficacy within the context of a meta-analysis. With this version, researchers can integrate and visualize the data more accurately, making adjustments as the trial progresses.

Based on the established parameters, including an alpha level of 0.05 and a beta of 0.20 (80% power), the analysis using Trial Sequential Analysis (TSA) indicates that a total of 1,475 patients will be required for the upcoming studies. This sample size is necessary to ensure that the sequential boundaries are adequately maintained, allowing for early termination of the trials if required for efficacy or futility.

library(png)
img <- readPNG("/Users/fabricio/Desktop/TSA/Adjusted Boundaries Print.png")
grid::grid.raster(img)