On the heels of a great event and presentation along with Rick Gaylord, our healthcare solution specialist, at the 2018 CEAI Conference, I want to continue the conversation about the far-reaching impacts of machine data and artificial intelligence for healthcare technology.
Machine data from IoT connected devices is growing at 50 times the growth rate of traditional business data. By 2025, more than 42 percent of the world’s data will be machine-generated.
There are massive amounts of machine data, hidden from normal view, in unstructured, complex and messy formats. With the right specialized tools, this data can be cleaned up, organized and is ideal for interpretation through machine learning and predictive analytics. Machine data can be transformed into insights that can drive automation, optimize utilization, diagnose and prevent equipment issues, and save organizations millions of dollars in increased efficiencies and decreased equipment failures.
Machine data has the power to transform major industries, including oil and gas, power, aviation, rail, and of course, healthcare. In fact, efficiency gains from IoT and machine data – including automated diagnostics, remote patient monitoring, and performance monitoring – could have a $63 billion business impact in healthcare. Some of the medical machines that generate the amount of data include MRI machines, CT scanners, x-ray machines, defibrillators, ultrasound equipment, ventilators, patient monitoring systems, anesthesia machines, and many more.
Key Use Cases of Machine Data
Out of the vast medical devices that generate machine data, a few of the top use cases include:
Machine Utilization – Machine data can help organizations understand the number of procedures per machine, per facility, by manufacturer type, to optimize equipment utilization and budgets.
MRI Machine Health – Analyzing machine data can alert operators of key triggers and send proactive maintenance alerts.
CT Scanner Health – Predictive analytics allows for real-time maintenance to avoid equipment and system failures and downtime.
Environmental Sensors – Monitoring environmental variables can indicate when key triggers might lead to equipment failure and send proactive alerts to avoid downtime.
Operator Usage & Analytics – Machine analytics offers insight into the behavior of equipment operators and can help identify and address gaps in training and balance load.
For more on Machine Log Data Use Cases: Data Doesn’t Lie: 5 Ways Hospitals Can Use Machine Log Data
Machine Data, Artificial Intelligence, and Machine Learning
Consider the massive amount of log data that even a single system produces. More than 50,000 events are logged each day by each system, with more than 2,500 different types of warning and error events. Machine learning (ML) and artificial intelligence (AI) allows us to process and interpret machine log data that might otherwise be too complex or simply too time-consuming for the human mind to analyze effectively.
With ML and AI-enabled predictive maintenance, organizations can avoid unplanned reactive maintenance and downtime, plan preventative maintenance during non-business hours, and even set up custom alerts and rules about equipment maintenance notifications. The possibilities are virtually endless.
The business impact of predictive analytics is quite impressive. On average, an expensive imaging machine like an MRI or CT scanner will face an issue eight to ten times per year and will be down six to eight hours each time – equating to about 62 hours of downtime, per machine, per year.
A facility with five MRI machines and five CT scanners that uses predictive analytics and maintenance can perform 500 additional procedures per year and earn $3 million in additional revenues over 3 years. This is at an assumed operating schedule of 10 hours per day, six days per week, and one procedure per hour at $2,000 per procedure.
More on business impact: Machine Learning and Predictive Maintenance Maximizes Healthcare ROI
Who Owns the Data?
With all of the machine data out there, one of the most pressing questions circulating in clinical engineering and healthcare technology communities is this – who owns it? Is it the Original Equipment Manufacturers (OEMs) who make the equipment, Independent Sales Organizations (ISOs) who re-sell the equipment, or healthcare providers who operate it?
Among other reasons, OEMs often believe they own the data because they own the software that generates the data. ISOs believe they need access to the machine data to better service customers. Providers believe that they have a right to the data because they paid for the machines. They are all valid points, and we will continue to expand on them in upcoming initiatives and content on data ownership. Stay tuned!