New publication in the American Journal of Kidney Diseases
At a glance
An open-access paper, co-authored by Drs Gasparini and Crowther, on prediction models for assessing the suitability of deceased-donor kidneys.
We are excited to share our latest work on prediction models for assessing deceased-donor kidneys, which was recently published open access in the American Journal of Kidney Diseases. This paper was co-authored by Dr Alessandro Gasparini, then our Principal Biostatistician, and our CEO and Director of Statistical Methodology Dr Michael Crowther.
The Kidney Donor Profile Index (KDPI) is routinely used in clinical practice for kidney allocation, despite modest predictive accuracy and calibration issues. This study of 75,867 adult kidney transplant recipients found that incorporating recipient characteristics substantially enhanced discrimination and calibration of the risk predictions, while machine-learning approaches and longitudinal laboratory donor data did not. This approach of including recipients' data could therefore be explored by transplant policymakers and professional societies committed to improving the kidney transplant allocation system.