Every few days, we will be publishing layman’s abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.
Rennert, L, Xie, SX. Bias induced by ignoring double truncation inherent in autopsy‐confirmed survival studies of neurodegenerative diseases. Statistics in Medicine. 2019; 38: 3599– 3613. doi: 10.1002/sim.8185
Autopsy-confirmed studies refer to study samples which only contain individuals who received an autopsy. Despite the wide use of autopsy-confirmed studies of neurodegenerative diseases, such as Alzheimer’s disease (AD) and frontotemporal lobar degeneration (FTLD), practitioners continue to ignore the selection bias that is inherent in these studies. When this occurs, results from standard statistical procedures such as the Kaplan-Meier estimator of the survival distribution or the Cox proportional hazard’s model will not yield results that are representative of the true population under study. This article describes the inherent double truncation problem in autopsy-confirmed survival studies of neurodegenerative diseases and how to properly analyze data from these types of studies, with an emphasis on estimation of the survival distribution.
A survival study in which participant’s enrollment occurs after the onset of the disease is subject to left truncation. This is a well-known cause of selection bias, and occurs in survival studies when individuals experience the event of interest (e.g., death) before they are recruited into the study. This paper describes an additional source of selection bias in these autopsy-confirmed studies that are a result of right truncation, which occurs because only subjects who receive an autopsy are included in the sample. This simultaneous presence of left and right truncation, also known as double truncation, is inherent in autopsy-confirmed studies of neurodegenerative disease.
The main focus of this paper is to inform about the inherent double truncation in these studies and demonstrate how to properly estimate and compare survival distribution functions in this setting. The authors do so by discussing several procedures for both estimating the survival time distribution and comparing survival time distributions for doubly truncated data, and applying these procedures in a case study of autopsy-confirmed AD and FTLD. The case study is supported by extensive simulation studies. Furthermore, these simulation studies explore the conditions needed for these survival distribution estimators to yield valid results.