System processes data at a nationwide level and provides doctors in hospital transplant centres with a ranked list of recommended recipients for the 15 to 20 organs donated each day
The health service’s use of algorithms to support the allocation of organs to patients is saving lives and ensuring a “scarce resource” is used to best effect, the NHS has claimed.
A newly published transparency document reveals that NHS Blood and Transplant has deployed algorithms that, for each donated organ,“ produces a ranked list of patients that are suitable to receive the organ on offer…. with the highest ranked patient deemed most suitable for that particular organ on offer”.
The system covers all 7,400 people around the country currently on the waiting list for an organ. This means that allocation decisions to be taken at a national level, rather than the previous process in which “organs were offered to the transplant centre that was closest to the donor hospital to transplant in to a patient of the centre’s choice”, according to the algorithmic record, which adds that “this process was neither equitable or transparent”.
The rules that dictate the tool’s operation – and, ultimately, its conclusions – were designed by clinical research and statistics experts.
Similarly, human “clinicians in hospital trust transplant centres make the final decision as to whether to accept the offer of an organ for the intended patient”.
The record adds: “If clinicians treating the patient feel [this is] not the best match, it will then be offered to the next highest-ranked patient.”
The document says outlines that there are no more than about 15-20 organs become available each day – donations that potentially could meet the need of about a quarter of one percent of those on the waiting list at any given time.
“To achieve the best use of this scarce resource, development of organ allocation algorithms is required,” the record says. “An algorithm is required to apply these rules in a speedily, standard and repeatable way with limited error.”
Related content
- NHS plans £180m framework for AI diagnostics and predictive analytics
- NHS study probes whether algorithms can detect Covid from cough recordings
- NHS Blood and Transplant plots £40m omni-channel contact centre
These benefits were reinforced by Professor Derek Manas, medical director for organ and tissue donation and transplantation at NHS Blood and Transplant. In response to enquiries from PublicTechnology, Manas said that the algorithmic system was both more effective and fairer than the procedures it has replaced.
“Due to the shortage of suitable donor organs, some patients sadly die before receiving a transplant. To save more lives, we must make the best use of available organs as effectively and fairly as possible,” he said. “NHS Blood and Transplant, together with the wider transplant community, has developed new policies for specific organs that use evidence-based, human-designed algorithms to support allocation decisions.”
Manas added: “More organs now go to the people who would benefit most, wherever they are in the country, based on the best available evidence and with fairness for people from all backgrounds. This reduces regional variation and improves equality and waiting list outcomes. Each algorithm produces a ranked list of suitable patients for the organs available. Clinicians in transplant centres make the final decision on whether to accept the offer for a specific patient. All systems are regularly reviewed by specialist advisory groups and patients can request further information if they have any questions. People die every day waiting for a transplant, but more lives could be saved if more people chose to donate. We urgently need more people to support donation by joining the NHS Organ Donor Register and sharing their decision with their families.”
The algorithm used to support kidney donations is supported by computer science experts from Glasgow University, the transparency record indicates, while the overall system’s operation is supported by software from IBM and Oracle.
But “the developed algorithm is embedded within NHSBT IT systems and so no data access to [an] external supplier is granted”, the document adds.
