MIT Sloan School of Management researchers help redesign national lung transplant allocation system using data analytics

New system's goal: Reduce overall deaths and increase access for patients waiting for lung transplants

CAMBRIDGE, Mass., March 28, 2023 /PRNewswire/ -- United Network for Organ Sharing (UNOS) has rolled out a new organ allocation system designed to be more equitable and effective for patients in need of lung transplants.

As of March 9, 2023, all donated lungs across the US are allocated based on a new policy, developed with the assistance of researchers from MIT Sloan School of Management. The policy considers a range of factors to determine which patients in need of donor lungs will be prioritized to get them. The policy change is predicted to reduce overall deaths and increase access for many patients waiting for a lung transplant.

Despite the record numbers of transplants performed in recent years, there are not enough organs of all types for patients who need them. More than 100,000 patients are currently on the national waiting list, while each year, approximately 65,000 new candidates are registered. In 2022, more than 42,000 received transplants, of which around 2,600 were lung transplants.

Working with UNOS, MIT Sloan's Dimitris Bertsimas, Boeing Leaders for Global Operations Professor of Management and professor of operations research, and Nikolaos Trichakis, associate professor of operations research, modeled a points-based framework called continuous distribution (CD) based on artificial intelligence and machine learning.

The new system will be more equitable and patient-focused, said Marie Budev, D.O., M.P.H., who serves as chair of the Organ Procurement and Transplantation Network (OPTN) Lung Transplantation Committee. UNOS operates the OPTN under contract with the US Department of Health and Human Services.

"We are excited about the ability of the new system to yield gains across all fronts," Trichakis said. "The new system is projected to be more equitable across different blood groups, height ranges, sexes, races, age groups, and geographic areas."

"The work underscores the power of data, artificial intelligence and optimization to achieve equitable and efficient allocations of donated lungs,'' Bertsimas said. "We are hopeful that our work will extend to all other organ allocation systems."

Their underlying work captured in their study, Applying Analytics to Design Lung Transplant Allocation Policy, which they co-wrote with MIT Sloan doctoral student Theodore Papalexopoulos and research and policy analysts from UNOS, will be published in the September-October 2023 edition of the INFORMS Journal on Advanced Analytics.

Under the previous system, candidates were placed in priority tiers. Patients whose criteria placed them at the boundary between tiers–just inside or outside a geographic tier, for instance–could have vastly different chances of getting a transplant.

The MIT Sloan researchers focused on the interplay of fairness and efficiency in resource allocation. In 2011, Trichakis, who holds a PhD from MIT in operations research, earned his thesis centered around issues of fairness, with the optimal design of organ allocation policies being one of the applications considered. In 2019, soon after Trichakis presented his and Bertsimas' findings to UNOS, the nonprofit organization that oversees the nation's transplant system, an MIT team, including Papalexopoulos, Bertsimas and Trichakis, started collaborating with UNOS on a new system.

"Our previous research was directly applicable to the problem that UNOS was trying to solve–the design of new, efficient and fair organ allocation policies," Trichakis said.

Studies have documented geographic and gender disparities for liver and kidney allocation, among others. The mathematical model developed by Bertsimas and Trichakis showed that a CD points system based on medical urgency, proximity to the donor hospital, age, and several other factors would result in fewer disparities related to age, gender and blood type, fewer deaths of patients awaiting transplants, and fewer post-transplant deaths.

"Our model utilized artificial intelligence and machine learning to analyze tradeoffs among patient outcomes and to come up with an optimal system design," Bertsimas said.

Data for candidates seeking a deceased-donor transplant are maintained by UNOS in its role as the OPTN, the nation's organ transplant system. When a new organ donor's information is entered into the database, the system produces a computerized ranking of candidates.

Under continuous distribution, lung waiting list candidates receive points, up to a maximum of 100. The first 50 points are related to two equally weighted criteria: how quickly a patient needs a transplant and the candidate's likelihood of surviving up to five years with the transplanted lung. A potential recipient would get 20 points for being under 18 when they register for an organ transplant, and five points if they had previously donated an organ.

Geographic points would be determined by the proximity of the candidate's hospital to the hospital where the organ is available. Data is gathered on potential recipients' physical stature–lungs need to fit in the rib cage.

Candidates with harder-to-match blood types will now receive a boost in points over those with more common blood types.  Another factor considered is an immunological profile that can limit the number of donors who could be a match. Previously, those with certain antibody profiles and those who were taller or shorter than average tended to get fewer transplants.

"The CD-based system prioritizes patients by considering multiple patient factors at the same time," Trichakis said. "This means that the transplant system can be more fair, more efficient and more flexible for the entire patient population."

Following the success of the initial MIT-UNOS collaboration, the MIT Sloan researchers have extended their work: more students across MIT Sloan are working with larger teams within UNOS to develop new systems for kidneys and pancreases. "A new protocol that we helped develop for kidneys will soon be presented to OPTN policymakers for consideration," Bertsimas said.

For further information, contact:

SOURCE MIT Sloan School of Management