Lund JL, Garcia de Albeniz X, Maringe C, Gaber C, Wyss R, Salas M, Ghazarian A. Cloning, censoring, and weighting – oh my! Applications of target trial emulation for evaluating cancer interventions when study eligibility, intervention initiation, and start of follow-up are not aligned. Presentation to be given at the 2022 ICPE Conference; August 27, 2022. Copenhagen, Denmark.


BACKGROUND: The growing availability of real-world data (RWD) provide more opportunities to evaluate the effectiveness and safety of interventions to prevent and treat cancer in clinical practice. However, the validity of study findings using RWD can be compromised by systematic biases which can often be avoided through principled study design. The alignment of study eligibility, intervention initiation, and the start of follow-up is one critical design feature necessary to avoid immortal-time and other related biases. Yet, this alignment can be challenging when interventions are not observed at the point of study eligibility, but rather over follow-up.

OBJECTIVES: This CER and Cancer SIG-sponsored session will review the target trial framework as a guide to robust observational study design and focus on the design and application of the clone-censor-weight (CCW) method for aligning study eligibility, intervention initiation, and start of follow-up. In this method: (1) clones (or copies) of individuals are created and assigned to each intervention strategy, (2) clones are then followed and analytically censored when they deviate from the assigned strategy, and (3) the uncensored population is reweighted using inverse probability weighting to account for potential selection bias induced by analytic censoring. This symposium is targeted towards pharmacoepidemiologists interested in causal inference methods or cancer outcomes research.

DESCRIPTION: We will review the target trial framework and introduce the core components of the CCW method (Jennifer Lund, 10min). Next, three speakers will present an application of target trial emulation using the CCW method to evaluate the effectiveness of an intervention at distinct points along the cancer care continuum. Each speaker will highlight one design or implementation consideration critical for the success of the CCW method in their study (18min each). Confirmed speakers and topics are listed below.

Xabier García-Albéniz: Breast cancer screening continuation versus stopping for women aged 70-84 years old in the US.
Camille Maringe: Initiation of surgery versus no surgery for non-small cell lung cancer in England.
Charles Gaber: Trimodal therapy versus definitive chemoradiation for locally advanced esophageal cancer in the US.

The symposium will close with a 25-min panel discussion moderated by CER and Cancer SIG leaders (Richard Wyss, Maribel Salas, Armen Ghazarian) covering the strengths/limitations of CCW versus other approaches (e.g., landmark method), use cases beyond cancer, and future methodologic directions. We hope this symposium will stimulate interest in the CCW method and provide resources that will support its dissemination and uptake within the pharmacoepidemiology community.

Share on: