Accounting for Selection Bias in Transplant Benefit and Waitlist Urgency Models
PresenterFlash Talk Presenter
Erin Schnellinger is an Epidemiology PhD Candidate at the University of Pennsylvania Perelman School of Medicine. Prior to her current position, she completed her Master’s in Biostatistics at Harvard T.H. Chan School of Public Health. Her methodological research interests include determining when and how to update clinical prediction models to ensure that their performance is maintained over time. Her applied research interests focus on mitigating biases in organ allocation models, such as the lung allocation score, to ensure that donor organs are allocated to transplant candidates in an equitable manner. If you are interested in collaborating with Erin, please feel free to contact her via email (email@example.com), phone (330-858-3846), or Twitter (@ESchnellinger).
The lung allocation system in the U.S. prioritizes lung transplant candidates based on estimated pre- and post-transplant survival. However, these models do not account for the fact that individuals who receive a transplant must survive on the waitlist long enough to receive the offered donor organ. Individuals who meet these criteria can differ from those who do not, resulting in survivor bias and inaccurate predictions. We propose a weighted estimation strategy to account for survivor bias in the pre- and post-transplant models used to calculate Lung Allocation Scores (LAS), the current basis for prioritizing lung transplant candidates in the U.S. We then created a modified LAS using these weights, and compared its performance to that of the existing LAS via time-dependent receiver operating characteristic (ROC) curves, calibration curves, and Bland-Altman plots. Overall, the modified LAS exhibited better discrimination and calibration than the existing LAS, and led to changes in patient prioritization. Our approach to addressing survivor bias is intuitive and can be applied to any organ allocation system that prioritizes patients based on estimated post-transplant survival. This work is especially relevant to current debate about methods to ensure more equitable distribution of organs.
KeywordsSurvivor bias, inverse probability weighting, organ transplant
Commenting is now closed.
To understand health and disease today, we need new thinking and novel science —the kind we create when multiple disciplines work together from the ground up. That is why this department has put forward a bold vision in population-health science: a single academic home for biostatistics, epidemiology and informatics.