The number one topic in healthcare seems to be Patient Safety. Government initiatives, healthcare provider projects, medical device manufacturer innovations, new service announcements all reference Patient Safety as the primary benefit.
Patient Monitoring, one of the technologies critical to improved patient safety, also receives a lot of coverage and particularly in respect of remote, or ambulatory, patients. Checking patient condition outside of the hospital environment promises to improve care and reduce care costs for all sorts of conditions.
No wonder it gets so much attention. Monitoring in intensive care has been business as usual for years now. Remote monitoring offers massive increases in market size for both devices and communications services. The medical devices market is thought to total more the $450 billion, and expected to grow around 10% each year. Innovation in medical devices, particularly in terms of size, power consumption and communication capabilities, is good business for manufacturers, governments, healthcare providers, and of course patients.
We’re working on a different type of innovation which promises an exponential expansion in the value delivered by patient monitoring – statistical modelling.
For years now the engineering industry has understood the value of monitoring current performance and condition data.
Every commercial flight by a major airline sends a constant stream of data back to computers on the ground. Engineers and their software analyse that data to make sure the machine behaving the way it should.
We’ll all be familiar with the TV pictures from Formula One Grand Prix races. Lines of engineers watch banks of computer screens reporting readings from cars racing at 200 mph. Many more engineers back at base are doing the same.
Who knows what the military does? Somewhere in New Mexico they’re directing pilotless drones over Afghanistan, and that’s just the bit they tell us about.
Of course they aren’t investing so much effort in being told when an engine breaks. They’re using low level data to predict “when” it’s going to break, and adjusting operating parameters to make sure it doesn’t.
In medicine monitoring software alerts clinicians to the events that have happened. In engineering monitoring software calculates what is going to happen. In medicine software merely alarms when a parameter is breached. In engineering the software compares and contrasts hundreds of parameters minute by minute, making millions of calculations, identifying patterns and comparing them with models. Apparently simple measures such as fuel consumption, temperature, vibration, voltage, and many more can take on a whole new meaning when compared with theoretical models and trended in relationships with each other.
Applying this technique for deeper insight requires some accurate data, some fast computing power and some extremely clever software – not generally available to healthcare providers. Making statistical modelling work in hospital intensive care environments isn’t simple, but it is possible.
We’re doing it.
Our research project is collecting minute by minute patient parameter data and comparing analysis results with statistical models of hypotensive episodes.
We’re adding statistical modelling to standard patient monitoring and using the results to help clinicians be on hand when hypotension occurs, improving patient outcomes and reducing healthcare costs.
What’s next? Don’t know yet, but it’s bound to be interesting.
Turning monitor readings into insight as opposed to reporting is where lies the real potential of patient monitoring.