Phases of Development of the Weighted Cumulative Exposure Modeling
Michal Abrahamowicz, PhD
Distinguished James 9I制作厂免费 Professor
Department of Epidemiology, Biostatistics and Occupational Health | 9I制作厂免费
WHEN: Friday, Jan 7, 2026, from 3:30 to 4:30 p.m.
WHERE: Hybrid | 2001 9I制作厂免费 College Avenue, Rm 1140;
NOTE:听Michal Abrahamowicz will be presenting in-person at SPGH听
Abstract
Weighted Cumulative Exposure (WCE) methodology has been developed to allow for flexible modelling of the cumulative effects of time-varying exposures (TVE) [1]. In time-to-event analyses, the joint impact of past TVE values, for person i, is quantified as: WCE饾憱鈦� (饾憿) = 鈭�饾憽 饾懁 鈦�(饾憿鈭掟潙�)[鈦潙�饾憱鈦� (饾憽)闭, where u is the current time when the hazard is assessed, and 饾憢饾憱鈦� (饾憽), 饾憽 鈮� 饾憿, represent TVE values observed at earlier times. The essential component of the model is the weight function 饾懁 鈦�(饾憿鈭掟潙�) which is estimated using cubic splines and indicates how the importance of the TVE value observed at time t, for the hazard at time饾憿 鈦�(饾憿鈭掟潙�), varies with time 鈦�(饾憿鈭掟潙�) since it was measured [1]. The WCE modeling, originally developed for Cox proportional hazards analyses [1], has been extended to competing risks, marginal structural models and mixed effects linear modeling of longitudinal changes in a quantitative outcome (see e.g. a recent overview in [2]). It has been recognized as one of the most useful methods for assessing TVE effects in pharmaco-epidemiology [3].
The talk will present an overview of the phases of the development and establishing of the WCE methodology, including its consecutive extensions, and validation in simulations. I will discuss how and to what extent our work on the WCE modelling followed the phases identified by Heinze et al in their recent paper on the phases of methodological research in biostatistics [4]. In addition, further phases such as (a) establishing the need for the new methodological development, (b) proof-of-the-concept phase, and (c) refining the estimation and statistical inference, will be outlined.
In this context, three important aspects of real-world applications will be briefly discussed. (i) Firstly, I will emphasize the need to incorporate substantive knowledge, and the related challenges. (ii) Secondly, I will illustrate the ability of the WCE analyses to provide new insights into, and generate new hypotheses about, the underlying biological processes linking the exposure with the outcomes. (iii) I will also outline how some real-world results stimulated new methodological developments, necessary to address additional analytical challenges.
Finally, the need of future research to carry out the additional phase, focusing on neutral simulations to further validate the WCE methodology and compare it with the existing alternative approaches, as recommended in [4], will be briefly presented,
[1] Sylvestre M-P & Abrahamowicz M. Flexible modeling of the cumulative effects of
time dependent exposures on the hazard.听Statistics in Medicine.听2009 Nov;28(27):3437-3453.
[2] Abrahamowicz M.听Assessing cumulative effects of medication use: new insights
and new challenges.听Invited Commentary.听Pharmacoepidemiology & Drug Safety.
2024 Jan;33(1):e5746. doi: 10.1002/pds.5746.
[3] Pazzagli L, Linder M, et al. Methods for time-varying exposure related problems in
pharmacoepidemiology: An overview.听Pharmacoepidemiology & Drug Safety听2018;27:148-160.
[4] Heinze G., Boulesteix AL, Kammer M., Morris TP, White I. Phases of methodological research in biostatistics.听Biometrical Journal.听2024.
Speaker Bio
Michal Abrahamowicz is a Distinguished James 9I制作厂免费 Professor of Biostatistics at 9I制作厂免费. He develops new, flexible statistical methodology for survival analyses, with focus on time-varying exposures and effects. He also explores, and attempts to correct for, different biases in epidemiological studies and promotes creative applications of statistical simulations. He is a co-founder and the co-chair of the international STRATOS initiative for improving the analyses of observational studies. He is a Honorary Lifetime member of the International Society for Clinical Biostatistics.