The lay abstract featured today (for Survival Analysis Without Sharing of Individual Patient Data by Using a Gaussian Copula by Federico Bonofiglio) is from Pharmaceutical Statistics with the full article now available to read here.

Bonofiglio, F. (2024), Survival Analysis Without Sharing of Individual Patient Data by Using a Gaussian Copula. Pharmaceutical Statistics. https://doi.org/10.1002/pst.2415

Abstract

Common survival analyses such as Cox regression and Kaplan-Meier (KM) estimations are often needed in clinical research, especially post-approval and Health Technology Assessment, and this requires access to individual patient data (IPD). However, IPD cannot always be shared due to privacy or proprietary restrictions, which complicates the making of such analyses. We might use published summary data to approximate such analyses, e.g. via meta-analytic techniques. In this paper, however, we propose a different paradigm: we use non-disclosive aggregates such as IPD marginal moments and a correlation matrix, shared upon request, to build pseudo IPD. Then, Cox and KM estimations are run on the pseudodata as if it was the original IPD. The aggregates are collected and input as parameters to a Gaussian copula (GC) that generates the pseudodata.  While the used IPD aggregates are very simple indeed, we show that the method, a bit unexpectedly, can reproduce a range of common survival analysis rather well. The method has its own limitations and yet it shows that sharing more IPD aggregates than is currently practised could facilitate “second purpose”-research and relax concerns regarding IPD access.