Each week, 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.

The article featured today is from Statistics in Medicine, with the full article now available to read here.

Matsouaka, RA, Atem, FD. Regression with a right‐censored predictor using inverse probability weighting methods. Statistics in Medicine. 2020; 4001-4015. https://doi.org/10.1002/sim.8704

Longitudinal study is where individuals are followed over a certain period. In these types of studies, recorded data of key variables might be incomplete due to drop-out, no follow-up, or early termination of the study before the advent of the event of interest. This paper focused primarily on the implementation of a regression model with a randomly censored predictor, a terminology adopted to define variables with partial information. It examines the use of inverse probability weighting methods in a generalized linear model (GLM) in which the predictor of interest is right censored (or above an estimated value) to adjust for the censoring. Three different weighting schemes were considered to improve the performance of complete-case analysis and to prevent selection bias: the inverse censoring probability weights, the Kaplan Meier weights, and the Cox proportional hazards weights. Monte Carlo simulation study, which is used to understand the impact of uncertainty in estimating models, was performed to evaluate and compare the empirical properties of the different weighting estimation methods. These weighting estimation methods were illustrated by being applied to the Framingham heart study data, estimating the relationship between onset age of a clinically diagnosed cardiovascular event and low-density lipoprotein (LDL) among cigarette smokers.