Pros, cons of Google AI predict when patients will die
Google study claims algorithm is 95% accurate to predict when a patient is going to die within 24 hours in the hospital.
Google has developed an artificial intelligence algorithm that can predict when you will die with 95 percent accuracy, according to the tech giant researchers.
The research, which tackles a range of clinical problems in hospital patients, was published recently in the journal Nature. Google applied artificial intelligence to a huge amount of data of more than 216,000 adult patients admitted to the hospital for at least 24 hours each in two medical centers.
The study tapped into data from Electronic medical records.
HOW HACKERS CAN MAKE USE OF ARTIFICIAL INTELLIGENCE AGAINST US
“We were interested in understanding whether deep learning could produce valid predictions about a wide range of clinical problems and results,” the researchers explain in the journal. “We have, therefore, chosen for the results from the most diverse domains, including an important clinical outcome (death), a standard measure for the quality of care (re-uptake), a measure of the use of resources (length of stay), and a degree of understanding of a patient’s problems (diagnosis).”
The proof-of-concept study has shown that the algorithm can accurately predict the risk of mortality, hospital readmission, prolonged hospital stay and discharge diagnosis. “In all cases, the method proved to be more accurate than previously published models,” he said.
The AI is 95 percent accurate in predicting patient mortality based on data from the University of California, San Francisco health care and 93 percent accurate using data from the University of Chicago Medicine system, according to the research.
DOOMSDAY AI MACHINES CAN LEAD TO A NUCLEAR WAR, A THINK TANK WARNS PAPER
This was significantly more accurate than the traditional predictive model, the augmented Early Warning Score, which makes use of a number of factors to help physicians determine how sick a patient is. This method was 85% accurate at the University of California, San Francisco system, and 83 percent accurate at the University of Chicago Medicine system, the study says.
The Google research comes at a time when the potential benefits and risks of the application of artificial intelligence is hotly debated. Of cybersecurity risks and so-called ‘doomsday’ machines that can wreak devastation on the technology, the potential for economic growth, and the experts are weighing the possible long-term impact of AI.
The health, which depends on a bewildering array of information, is increasingly touted as a good fit for the use of AI. FDA Commissioner Scott Gottlieb, for example, talked about the promise of AI in a speech earlier this year.
WHO IS AFRAID OF ARTIFICIAL INTELLIGENCE? IT COULD SOLVE MANY OF OUR NATION’S MOST DIFFICULT PROBLEMS
The technology, but also makes for important challenges.
Speaking on Fox and Friends Tuesday, Family Medicine Physician Dr. Mikhail Varshavski said that, while the connecting of a wide range of health-related information may be useful for patients, data privacy is important. “The thing that is worrying me is what happens with this data and who is in possession of this information?” he said. “I hope that, as a physician, that these companies make use of the data, to the benefit of the patients, not the companies themselves.”
“Machines make mistakes, and sometimes make mistakes based on erroneous data,” Varshavski added. “There needs to be oversight of what these things do.”
GOOGLE IS SET TO PUT AN END TO THE CONTROVERSIAL PROJECT MAVEN-MILITARY AI PROGRAM, REPORT SAYS
Even within Google, the application of AI has proven to be controversial. The search giant recently ended its involvement in the Project Maven, a controversial military program that uses AI to improve drone targeting. Project Maven is a source of tension in Google. In April, about 3,100 Google employees signed a letter to the CEO of the company, Sundar Pichai, and asks him to pull the tech giant out of the project.
Follow James Rogers on Twitter @jamesjrogers