GRAFT - Generic Rating of Allograft Function Post Transplant
Theses
Overview
Overview
abstract
Background
Research on the optimal management of deceased organ donors poses unique challenges including the fact that one deceased donor may provide up to 8 organs for transplantation. Measuring the post-transplant function of these organs – good or bad – represents an attractive way of deciding whether treatment of deceased donors is working well, or not so well. Function, however, is organ-specific. Therefore, to conduct the most efficient and informative research on deceased donor management, we need an outcome measure that works well in all organs. The new outcome measure is called Generic Rating of Allograft Function post-Transplantation (GRAFT).
Methods
In this thesis, I highlight the methods for developing the cardiac-specific version of the GRAFT instrument. The same methods, however, have and will be applied to other organ-specific versions. The work comprised various study designs and developed novel research tools, all of which have advanced the development of the GRAFT instrument. At first, we developed a simple conceptualization for the instrument. Through regular consultation with research methodologists, biostatisticians and clinical experts, we refined the fundamental conceptualization and then refined the generic instrument, itself. One key concept is that GRAFT ratings should correlate with one-year graft function. To maximize its utility, I developed a heart--specific guide for applying GRAFT in future studies, and other organ-specific guides are underway. Specifically, we developed these guides by identifying the most robust predictors of one-year graft function through the conduct of organ-specific systematic reviews and meta-analyses of prognostic factors. The evidence from these reviews, in consultation with a focus group of organ-specific transplant physicians, lead to refinements of our guides. We subsequently conducted a mixed-methods user testing to assess reliability and usability of the organ-specific guides. In appraising the evidence informing the guides, we developed GRADE guidance and a novel absolute risk calculator to assess our certainty in the body of evidence on prognostic factors informing our guides.
Results
We developed a 6-point generic rating instrument for classification of graft function to be applied post-transplant across all major solid organs. We designed GRAFT to be applied at the time of discharge, 1-month post-transplant, or at the time of death (whichever occurs first). We classify function as 1) normal, 2A) impaired but likely to gain normal function, 2B) impaired and unlikely to gain normal function, 3A) severely impaired but likely to gain some function, 3B) severely impaired and unlikely to gain some function, and 4) irreversible graft failure. Clinical expert collaborators for each organ type confirmed face validity of the GRAFT instrument.
For all organs, we identified a number of prognostic factors that can guide users in classifying organ function post-transplant. In consultation with clinical experts, we determined that the most important factor is graft function as measured by left ventricular ejection fraction (LVEF) or right atrial pressure (RAP). Due to limitations with the quality and quantity of the evidence, however, the heart transplant experts did not rely on the results of their organ group’s systematic review. In turn, we conducted a retrospective cohort study to calculate the best estimate of association between LVEF, RAP, and overall mortality post heart transplant.
For the cardiac version of GRAFT, user testing demonstrated high reliability (Kappa of 0.87, 95% CI 0.62 – 1.00) and acceptable usability (system usability score of 75, inter-quartile range of 72.5 – 80).
In the process, we developed and published GRADE guidance for assessing certainty in the body of evidence addressing prognostic factors and devised a calculator to transform relative effect of each prognostic factor to absolute risks (http://hiru.mcmaster.ca/AbsoluteRiskCalculator/).
Conclusion
In this thesis, I advanced the development of an innovative generic instrument for the classification of graft function specifically for the purpose of application in clinical trials of deceased donor interventions. This work is ongoing, but very advanced for heart-specific components, for which I have ensured face validity, and demonstrated reliability and usability. The GRAFT instrument may better facilitate the conduct of future research to improve care of deceased organ donors with a view to improving quality and quantity of organs for transplantation.
