MIT AI algorithms aim to revolutionize organ transplant system
MIT AI algorithms aim to revolutionize organ transplant system Bill Siwicki
The U.S. organ transplant system is partnering with the Massachusetts Institute of Technology on an AI-driven algorithm framework, which launched March 9. This framework, known as continuous distribution, aims to make organ transplants more equitable.
Continuous distribution, which considers all patient factors simultaneously, is designed to be more effective than previous organ allocation processes but also increases transparency through a weighted score that is unique to every organ candidate.
Ultimately, the idea is to provide the sickest patients with even faster access to life-saving organs. The framework is projected to cut waitlist deaths in half through timely reactions and updates to allocation that can take months if not years under the current process.
To get the full story on this new process and the role artificial intelligence plays, we interviewed two top players in the project.
James Alcorn is the senior policy strategist with the United Network for Organ Sharing (UNOS), the non-profit leading the U.S. organ donation and transplantation system. In his role, he is responsible for the development of the continuous distribution of organs and setting strategy for how the country determines how to allocate donated organs for transplant in accordance with federal laws and regulations, engaging with medical professionals, patients, academics, institutions, government entities and the public to meet this need.
Nikos Trichakis is the Zenon Zannetos career development professor and associate professor of operations management at the MIT Sloan School of Management and a co-author of the research paper, "Applying Analytics to Design Lung Transplant Policy." He is one of the MIT Sloan researchers who assisted UNOS in developing the continuous distribution framework by applying machine learning and mathematical optimization to make allocation policies more efficient, more equitable and more inclusive.
Q. Please describe the state of the organ transplant system today? Further, what are the faults you're trying to solve?
James: The U.S. has a complex organ donation and transplant system that saves tens of thousands of lives every year. In 2022, 42,888 transplants were performed nationwide – the most ever done in one year. September 9 also marked the one millionth organ transplant in the U.S. since the first one – a kidney transplant – was performed in 1954.
This is all possible thanks to the gifts of generous donors, their courageous families, constant medical and scientific breakthroughs, and the ongoing expansion of intricate organ allocation policies.
Despite these achievements, over 100,000 patients across the country are still waiting for an organ transplant. We are dedicated to pursuing new ways to make the donation and transplant process safer, faster and more equitable as we continue to address the nationwide organ shortage.
The number of patients in need has always outweighed the supply of organs that are viable for transplant, but with more scientific breakthroughs and careful policy design – in this case combining the two – we can make the system even more efficient and ultimately save more lives.
"We hope all patients will have a fairer chance of receiving a transplant, regardless of their sex, race, blood group, unique physiological characteristics or where they reside."
Nikos Trichakis, MIT
Nikos: The previous lung allocation system was a tier-based system: Candidates were placed in different priority tiers, and organs were offered within each tier first, before moving to the next tier. Patients who were at the boundary of a tier could have vastly different chances of getting a transplant, depending on which side of the boundary they were on.
For example, there were geographic tiers, and whether patients resided within or outside a tier (even by one mile) had a big impact on their success of receiving a transplant. Although the use of boundaries and tiers yielded operational benefits, it also introduced the possibility of inequitable access to lungs for certain patients.
Q. Please describe continuous distribution. What is it? How does it work?
James: Continuous distribution is a new allocation framework to match donated organs with registered transplant patients. It is not only a new way of allocating organs in a much more equitable and efficient manner, but it also presents a new way of developing the policies by which we decide who gets the next available organ.
Nikos: In essence, continuous distribution is a point system in which points are awarded for pre-determined attributes. Points then are added up, and the higher the cumulative score, the higher the priority is.
For organ allocation, examples of attributes that earn you points are medical urgency, proximity of donor hospital to transplant hospital, and waiting time, among others. By adding points across multiple attributes, it means that continuous distribution accounts for all these attributes simultaneously. Therefore, it becomes unlikely that a single attribute will drive priority.
Second, the ability to adjust the number of points awarded per attribute offers tremendous flexibility to policymakers to control the level of priority to be given to each patient group. By carefully selecting the number of points awarded per attribute, the optimization framework we developed at MIT was able to produce potential policies that balance the priority that is given to different patient groups as much as possible, thereby making the system more equitable.
Q. How do you intend continuous distribution to resolve the aforementioned faults in the organ transplant system?
Nikos: The lack of boundaries means that the transplant system can be more fair, more efficient and more flexible for the entire patient population.
James: Historically, organ allocation requires us to categorize transplant candidates based on numerous characteristics such as age, disease severity, blood type and more. These categories were ordered, and organs were offered sequentially, which in addition to other challenges posed limitations for “edge cases” – patients with similar characteristics that place them on opposite sides of particular categorizations.
Continuous distribution allows us to consider all these variables simultaneously and combine them into one unique, patient-centric score that can be assigned across all registered transplant candidates.
What really makes this an innovative framework as opposed to a regular policy change is that continuous distribution is designed as a cross-organ system to be adjustable in a way that reflects the priorities of the transplant community and advances in clinical science.
It changes the interface by which clinicians interact with the transplant system as well as how future allocation policies are developed and adopted. The framework leverages artificial intelligence to support decision-making among patients and donation and transplant professionals.
In determining allocation, there are millions of potential scenarios that can be adopted and under the previous processes it would take years to draft, model, discuss and vote on the best way forward. We can’t take that long when lives are on the line.
Using AI, the continuous distribution framework can instead evaluate these millions of policy scenarios and their outcomes within mere days, allowing committees to quickly narrow down the solutions that will save the most lives and speed up the policy development process.
Q. Today, now that your system has launched, who is using continuous distribution and how has it been working?
James: Continuous distribution officially launched for lung allocation on March 9. The ultimate goal is to establish a single allocation framework for all organs, creating a new standard not only for lungs, but also kidneys, pancreases, hearts, livers and other types of organ transplants in the future.
We’re closely monitoring lung allocation and will release in-depth reports evaluating its 3-month, 6-month and 1-year performance. Modeling results to date have shown that with continuous distribution, we are going to save more lives in the long run, including reducing deaths on the waitlist, and that we can expect additional benefits as we transition other organs, as well.
Specifically, we have statistical modeling predicting the new framework will reduce the number of patients who pass away while waiting for a lung transplant as well as make transplants possible for even more patients, particularly those with the greatest medical need.
A recent study by the Cleveland Clinic and the Scientific Registry of Transplant Recipients (SRTR) determined that across all modeling scenarios, “the system led to improved overall measures of equity” when compared to the previous system.
Continuous distribution not only provides a new way for medical professionals to interact with the transplant system, it also increases transparency by offering new insights for patients. Candidates can now go online and use the lung Composite Allocation Score (lung CAS) calculator to see how their various medical characteristics will be weighed to determine their priority on the waitlist.
Q. With the implementation of continuous distribution, what do you hope the organ transplant system looks like in five years?
Nikos: We hope all patients will have a fairer chance of receiving a transplant, regardless of their sex, race, blood group, unique physiological characteristics or where they reside. Additionally, we hope the new matching system will increase overall patient welfare – in particular, fewer patients will die on the waitlist and patients who receive transplants will live longer.
James: In five years, we expect to see an organ donation and transplant system that saves more lives than ever before, and continuous distribution will be an essential part of that. By then, we will be well on our way to transitioning all solid organs to the continuous distribution framework with the goal of completing all within a decade.
We also have a long-term plan to roll out additional features and enhancements in future continuous distribution versions that will further benefit the transplant community and, most important, the patients we serve.
Finally, we also will have a system that is more transparent and interactive for patients, allowing them to make more informed choices about their healthcare.
Follow Bill's HIT coverage on LinkedIn: Bill Siwicki
Email the writer: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.