Incorporating Missingness in a Framework for Generating Realistic Synthetic Randomized Controlled Trial Data

TL;DR

This paper develops a framework for generating realistic synthetic RCT data that accounts for missing data mechanisms, demonstrating improved performance over traditional methods by modeling missingness explicitly.

Contribution

It introduces methods to generate synthetic missing values in RCT data, comparing strategies like complete case, inverse probability weighting, and multiple imputation.

Findings

Incorporating missingness models improves synthetic data realism.

Methods accounting for missingness outperform complete case approaches.

Synthetic data closely mimics real data distribution.

Abstract

The current literature regarding generation of complex, realistic synthetic tabular data, particularly for randomized controlled trials (RCTs), often ignores missing data. However, missing data are common in RCT data and often are not Missing Completely At Random. We bridge the gap of determining how best to generate realistic synthetic data while also accounting for the missingness mechanism. We demonstrate how to generate synthetic missing values while ensuring that synthetic data mimic the targeted real data distribution. We propose and empirically compare several data generation frameworks utilizing various strategies for handling missing data (complete case, inverse probability weighting, and multiple imputation) by quantifying generation performance through a range of metrics. Focusing on the Missing At Random setting, we find that incorporating additional models to account for the missingness always outperformed a complete case approach.