The Blood Service has received a grant of €40,000 from the European Blood Alliance (EBA) to study the possibilities of adjusting the time between blood donations individually.
To be initiated this year, the research project will seek to utilise machine-learning to create a model or a profile based on the blood counts and genetics of a donor for determining more personalised time intervals between blood donations.
At the moment, the minimum time between two donations is determined by gender: women can donate every three months, men every two months. The research projects by the Blood Service indicate that for some donors, this could be too often for the haemoglobin levels to restore. It is unfortunate for donors to be refused due to a low haemoglobin level when they come in to donate blood.
There are already mathematical models to predict the rejections based on the age and gender of the donor. The research project by the Blood Service will make use of the genetic data collected of the donors into the Blood Service Biobank as well as the results of the FIN Donor 10 000 iron study.
“This is the first time that genetic information is combined with blood counts and iron markers when the health of blood donors and the preconditions for donating are studied,” comments the leader of the research project, Docent Mikko Arvas from the Blood Service.
Background materials of the research project also contain the haemoglobin level and blood group data from the more than six million donations given in Finland in the 2000s. The goal is to utilise artificial intelligence to produce an algorithm that can provide information about the most optimal minimum interval between donations for a single donor. The research project would also estimate the cost effects of individualised donation times and potential effects on the entire blood supply chain.
The research project will be initiated in the summer of 2019, and it will develop a model that will be tested in co-operation with the University of Cambridge.
“If the model works, it will be published so that blood services in other countries can further test and develop it,” Arvas comments.
This is not the first time that the Blood Service explores the potential of machine learning and genomic data; they were previously used in a study to predict the recurrence of leukemia after a stem cell transplantation. The findings of this study were published earlier this year in the Leukemia journal.