A new algorithm developed by scientists at UCLA may bring us as close to fortune telling as we'll ever get. The computation is able to predict life expectancy of heart failure patients, and can tell how long they will live if they do or do not receive a heart transplant. This will help doctors when making decisions on patients awaiting heart transplants.
The algorithm is called the Tree of Predictors and is detailed in a study published online in PLOS One. The Tree of Predictors uses 33 data points to assess patients awaiting a heart transplant, such as their age, gender and body mass index in order to understand how long a patient would live with or without a heart transplant. The algorithm also takes 14 points of data from the donor and six points related to the compatibility of the donor and the recipient. The algorithm can learn from additional information over time and in a way mirrors the human thinking process.
The algorithm was tested on 30 years of data from people who were matched with donors via the United Network for Organ Sharing. Results showed that The Tree of Predictors was more accurate in predicting outcomes than current methods and outperformed current models by 14 percent.
People require heart transplants for a number of reasons, with severe heart failure being the most common ailment, the American Heart Association report. The heart is removed from a deceased donor and then given to a recipient. About 88 percent of patients survive the first years after a transplant surgery.
3,000 people in the U.S. are on the heart transplant waiting list any given day, but there are only about 2,000 hearts available each year. As the demand for heart transplants is greater than the supply of available hearts, the hope is that this will help narrow down which patients would benefit most from this operation.
"Our work suggests that more lives could be saved with the application of this new machine-learning–based algorithm," said Mihaela van der Schaar, Chancellor's Professor of Electrical and Computer Engineering at the UCLA Samueli School of Engineering, in a statement. "It would be especially useful for determining which patients need heart transplants most urgently and which patients are good candidates for bridge therapies such as implanted mechanical-assist devices."
The potential of this algorithm is not limited to heart transplant patients though, and tests have shown that it can also accurately predict credit card fraud and the popularity of specific news topics. It's not yet clear if and when the algorithm will be implicated into the heart donation decision making process.