Pitfalls of Administrative Censoring in Survival Models with Time-Indexed Inputs

2026-07-11

AI Summary

This study identifies and systematically investigates temporal information leakage in survival analysis caused by administrative censoring, wherein models inadvertently exploit time stamps in input features to infer record dates, thereby conflating follow-up availability with event risk and inflating predictive performance. We formally define this leakage phenomenon for the first time, distinguishing it from classical censoring mechanisms and genuine time-varying risks, and propose corresponding detection methods and data construction guidelines. Through simulation studies and validation on a real-world mammography screening cohort—evaluated using both fixed-time window AUC and Harrell’s C-index—we demonstrate that this bias substantially distorts model assessment. To mitigate the issue, we advocate a design principle: “n-year prediction requires at least n years of potential follow-up,” which effectively prevents such leakage.

Abstract

Survival models can model time-to-event outcomes using partially observed data. They are widely used in clinical prediction, including cancer risk, disease progression, treatment response, and mortality. Recent models often rely on rich inputs collected at a specific clinical encounter, such as medical images, laboratory tests, electronic health record snapshots, or sensor measurements. In large retrospective datasets, these inputs are usually collected over many calendar years. As a result, they may contain clues about when they were acquired through changes in devices, protocols, documentation, patient mix, or clinical practice. This creates a potential failure mode when outcomes are observed only up to a fixed study end date. More recent records necessarily have less potential follow-up than older records. A model that can infer the record date from the input may therefore learn to predict how much follow-up was available rather than the patient's true risk of experiencing the event. We call this failure mode administrative-cutoff leakage. In this paper, we characterize when this leakage can occur, distinguish it from classical informative censoring and genuine temporal changes in risk, and propose practical ways to detect it. In simulations, we show that administrative-cutoff leakage can inflate fixed-horizon AUC and can also affect Harrell's C-index under realistic follow-up patterns. We then demonstrate the same behavior in a real mammography cohort. These results motivate a simple design principle for survival prediction: for an n-year prediction task, the dataset should provide at least n years of potential follow-up after the latest input date. Otherwise, the models may be subject to bias induced by administrative-cutoff leakage.

Problem

• administrative censoring
• survival models
• time-to-event
• follow-up bias
• administrative-cutoff leakage

Innovation

• administrative-cutoff leakage
• survival analysis
• time-indexed inputs
• informative censoring
• follow-up bias