What is the difference between one-stage and two-stage IPD meta-analysis?

Two-stage analysis summarizes each study individually then combines results using meta-analysis, while one-stage analysis models all participant data simultaneously using mixed-effects, which is generally better for investigating interactions.

Which R packages are used for these statistical models?

The skill provides patterns for metafor, lme4, survival, coxme, multinma, and mice, covering the most common requirements for biostatistical synthesis.

How does this skill handle missing data in multi-study settings?

It implements multilevel multiple imputation using the mice and mitml packages to account for clustering within studies, ensuring that imputations respect the study-level structure.

Can I combine individual patient data with published aggregate data?

Yes, the skill includes specialized patterns for using the multinma package to synthesize IPD with aggregate-level data (AgD) in a single integrated model.

Individual Participant Data (IPD) Meta-Analysis in R

Name: Individual Participant Data (IPD) Meta-Analysis in R
Author: choxos

bychoxos

0•

数据科学与机器学习

Facilitates complex individual participant data (IPD) meta-analyses using tidy R workflows and advanced statistical modeling.

This skill provides a comprehensive framework for conducting Individual Participant Data (IPD) meta-analyses within the R environment, specifically tailored for biostatistical research. It bridges the gap between raw patient-level data and synthesized evidence by offering standardized patterns for one-stage and two-stage approaches, handling longitudinal and survival outcomes, and performing network meta-analyses (NMA). By implementing best practices for treatment-covariate interactions and missing data management, it ensures robust and reproducible statistical synthesis for clinical trials and medical research.

主要功能

01Multilevel missing data handling via multiple imputation

02Advanced treatment-covariate interaction analysis

030 GitHub stars

04One-stage and two-stage meta-analysis implementation patterns

05Integration of IPD with aggregate-level data (AgD)

06Mixed-effects modeling for linear, binary, and survival outcomes

使用场景

01Investigating treatment effect modification by individual-level characteristics like age or biomarkers

02Synthesizing patient-level clinical trial data across multiple study sites

03Conducting network meta-analyses to compare multiple treatments using a mix of IPD and literature results

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add choxos/biostatagent ipd-meta-analysis

For use in Claude.ai and ChatGPT

主要功能

01Multilevel missing data handling via multiple imputation

02Advanced treatment-covariate interaction analysis

030 GitHub stars

04One-stage and two-stage meta-analysis implementation patterns

05Integration of IPD with aggregate-level data (AgD)

06Mixed-effects modeling for linear, binary, and survival outcomes

使用场景

01Investigating treatment effect modification by individual-level characteristics like age or biomarkers

02Synthesizing patient-level clinical trial data across multiple study sites

03Conducting network meta-analyses to compare multiple treatments using a mix of IPD and literature results