A0398
Title: Optimal treatment allocation in the presence of competing risks and clustering
Authors: Erica Moodie - McGill University (Canada) [presenting]
Misha Dolmatov - McGill University (Canada)
Dipankar Bandyopadhyay - Virginia Commonwealth University (United States)
Abstract: The precision medicine framework has been used to discover tailored treatment strategies in a variety of settings and has largely been derived within a causal framework due to the need for large (and thus typically observational) datasets. Existing methods are extended to address the pressing question of how to optimally allocate kidneys for transplantation when the pool of deceased donors includes individuals who were living with hepatitis C virus (HCV). The proposed approach accounts for the non-random allocation of kidneys from people with or without HCV, multiple (competing risks) endpoints, and clustering of data due to side effects. The proposed method is applied to data from the US National Organ Procurement and Transplant Network registry from 1994 to 2014.
