What is a good Effective Sample Size (ESS) percentage?

An ESS above 70% of the original sample size is generally considered good. An ESS between 30-50% is concerning, and below 30% suggests the results are likely unreliable due to poor overlap or over-adjustment.

What is MAIC in clinical research?

MAIC stands for Matching-Adjusted Indirect Comparison, a statistical method used to compare treatment effects across different clinical trials when individual patient data (IPD) is only available for one of the studies.

When should I use anchored vs. unanchored MAIC?

Use anchored MAIC when both trials share a common comparator; it is more robust as prognostic factors cancel out. Use unanchored MAIC only as a last resort when no common comparator exists, as it requires much stronger assumptions.

Does this skill provide R code examples?

Yes, the skill includes quick reference R code utilizing the maicplus library for centering IPD, estimating weights, and running both anchored and unanchored analyses.

MAIC Methodology & Analysis Guide

Name: MAIC Methodology & Analysis Guide
Author: choxos

bychoxos

Data Science & ML

Provides expert guidance for conducting and reviewing Matching-Adjusted Indirect Comparisons (MAIC) in clinical data analysis.

About

This skill equips Claude with deep methodological knowledge for performing Matching-Adjusted Indirect Comparisons (MAIC) according to the NICE DSU TSD 18 framework. It helps researchers and data scientists navigate the complexities of adjusting for population differences when comparing treatment effects across trials with disparate data types, such as Individual Patient Data (IPD) and Aggregate Data (AgD). The skill covers critical aspects including covariate selection, interpreting Effective Sample Size (ESS), diagnosing weight distributions, and deciding between anchored and unanchored approaches, ensuring rigorous and statistically sound indirect treatment comparisons.

Key Features

Weight diagnostics and ESS interpretation thresholds
NICE DSU TSD 18 compliant methodological guidance
Decision framework for anchored vs. unanchored MAIC
R code implementation snippets for the maicplus package
Covariate selection and balancing strategies
0 GitHub stars

Use Cases

Reviewing MAIC results and weight stability for clinical submissions
Designing a statistical analysis plan for indirect treatment comparisons
Selecting and justifying effect modifiers for population adjustment

About

Key Features

Weight diagnostics and ESS interpretation thresholds
NICE DSU TSD 18 compliant methodological guidance
Decision framework for anchored vs. unanchored MAIC
R code implementation snippets for the maicplus package
Covariate selection and balancing strategies
0 GitHub stars

Use Cases

Reviewing MAIC results and weight stability for clinical submissions
Designing a statistical analysis plan for indirect treatment comparisons
Selecting and justifying effect modifiers for population adjustment