trial-sequential-analysis

Trial sequential analysis (TSA) is a method used in cumulative meta-analyses to control type I and II errors by applying sequential monitoring boundaries. You should use it to determine if accumulated evidence is conclusive and to estimate the Required Information Size.

Calculate the Required Information Size (RIS) and information fraction using your cumulative meta-analysis data. This Skill processes binary and continuous outcomes to assess whether the accumulated evidence is sufficient.

Monitoring boundaries in TSA define thresholds for benefit, harm, and futility. Crossing these boundaries during cumulative analysis indicates that the evidence is conclusive, helping you decide whether to stop or continue future trials.

Yes, you can perform TSA and generate plots in R using the RTSA package or TSA software. This Skill supports R-based workflows to calculate information size, apply heterogeneity adjustments, and visualize monitoring boundaries.

Heterogeneity adjustment modifies the Required Information Size calculation to account for variance between studies. This adjustment ensures that monitoring boundaries and information fraction estimates remain reliable despite diversity in your cumulative meta-analysis data.

TSA is not recommended if your cumulative meta-analysis lacks sufficient data to estimate variance or if heterogeneity is excessively high. It requires binary or continuous outcome data to properly calculate information size and apply monitoring boundaries.