Preventing Brain Injuries with Predictive Analytics
From the National Football League’s various attempts to crack down on concussions and dodge various concussion-related lawsuits to the recent death of former British Prime Minister Margaret Thatcher from a stroke, the treatment and prevention of traumatic brain injuries (TBIs) has caught the public attention of late.
Coincidentally, IBM and Excel Medical announced a partnership with the UCLA Department of Neurosurgery recently in an attempt to tackle TBIs with predictive analytics. “In today’s world we manage patients in the intensive care unit by identifying the crises that occurred in the last 24 hours and reacting to them,” said Dr. Neil Martin, Professor and Chair of the UCLA Department of Neurosurgery on the state of brain treatment today. “In tomorrow’s world, we want to manage patients proactively.”
Healthcare has been an opportunistic use case for big data, as analytics have, among other things, helped inpatient optimization in hospitals as well as detecting fraudulent claims that grind government programs like Medicare to a halt.
However, few areas convey the potential power of such predictive techniques better than brain injuries. For example, the difference between detecting a stroke within an hour and detecting one within 24 hours could mean the difference between recovering within a few weeks to not recovering at all.
According to a video below that was produced by IBM, approximately 1.7 million traumatic brain injuries occur in the United States every year.
“In the past it has not been possible to integrate all the data from thousands of data points, streaming from all the monitors so we can forecast this kind of a crisis,” said Martin on the struggle to detect critical brain injuries before they happen. “Now working with Excel Medical and IBM, we can collect information in real time on the fly so that we can generate timely alarms that allow us to see a crisis coming and treat it proactively.”
Such real time information could help bridge the gap between reactive and proactive medicine mentioned by Martin. For example, sensors are adept at telling monitors when an abnormal cranial pressure buildup is happening. This pressure is often a result of swelling and brain hemorrhaging that can prove disabling or fatal to the patient if left untreated.
Detecting such dangerous levels of pressure means more than simply applying and reading sensors. It also means collecting and analyzing a large enough data set such that individual baselines for tolerable pressure variance are established.
Detecting and treating TBIs is a critical goal for modern medicine and one that will likely be advanced by big data analytics initiatives such as this one.