Healthcare’s IOT’s data explosion
When the first wireless devices were introduced to enable remote monitoring of patients'
health conditions, it was nothing short of a revolution. As well as freeing up hospital beds for patients with more serious conditions, it meant the monitor wearer could more or less go about their usual day, such as in the case of blood pressure monitors, which although clunky in their early days have now become far more discreet.
The COVID-19 pandemic highlighted another huge benefit: infection prevention, by minimising hospital visits and contact.
The network of physical devices that contains the sensors and software that make this remote monitoring possible, is the Internet of Things (IoT). It includes wearable devices like Fitbits and Apple Watches.
"I happen to be a runner, and I’ve used monitors for two decades", explains Nicole Fagen, Director of Product Management & Strategy for AIOps and Automation at Broadcom. "They constantly improve what they're able to track. Initially it was just heart rate, and then it was how many steps, then altitude, then it was how fast I was running – and then it could map my whole run. So every step of the way, monitors collect more data that give doctors more insights that they are able to leverage.
"This is happening across the healthcare industry; a lot of these devices now send data to doctors in real time. Years ago, I wore a heart monitor for medical reasons for a while. I had to bring the device back to my doctors and they would download the data from the device so they could do their assessment. Now it’s all done remotely."
Analysing data accurately
But while the devices are becoming ever more sophisticated, Fagen says there is a way to go in terms of analysing the data efficiently.
"These tools are literally creating billions of pieces of data every day, which should be nothing but good news for medical professionals who rely on information to keep individuals and communities healthy. Unfortunately, the infrastructure needed to collect, store, and analyse this data is often years or decades behind the wearables that create and transmit it", she says. To really make best use of it, she says there needs to be an integrated system to collect it. "Because there are so many different devices on the market today, it is difficult for hospital IT departments to get a holistic evaluation of patients who track their daily runs with a Fitbit, their heart rate with an Apple Watch, their glucose levels with a DexCom, and their food intake with the MyFitnessPal app," she says.
Computer power reads the data humans can’t
All of the information that these tools generate usually exists in private silos that are challenging for third parties to access. Even when doctors can get the relevant data, they have no practical way to fuse all of the information into a usable form."
Fagen adds that the more information we have, the harder it seems to find what really matters. "Last year humanity produced 2.5 quintillion bytes of data per day", she says. "That's 2.5 followed by 18 zeros. Only a subset of that is for medical purposes, but a glaring fact remains – no human is ever going to be able to process that. The only way we're going to be able to do that is through massive computer power."
However collecting information is just a start, as centralised systems must be able to analyse the data and flag situations when necessary. "For example, someone who runs five miles a day may be considered “healthy,” but if that person has high blood pressure, it might not be advisable for them to exercise as much until they can get their blood pressure regulated", Fagen explains.
"This is one of the major flaws of the IoT revolution: the lack of communication between devices can mask serious problems."
Last but not least is ensuring data is secure. This can be particularly challenging for hospitals, which can have anywhere between 10,000 and 50,000 devices amassing data. A 2021 survey by Philips found that almost half of all surveyed IT staff in hospitals felt that IoT security staffing is “insufficient,” following recent ransomware attacks and data breaches.
However Fagen says that while the challenge is significant, it is not impossible to overcome. "Most healthcare systems in the US already have the right tools in place to enable collection of mass amounts of data in a way that supports security and privacy. That’s because most hospitals, not to mention 90% of major insurance companies, rely on mainframe computers to handle just about every aspect of their operations. Mainframes have been around for a long time, continuously evolving and innovating and they have proven their worth time and time again."
Mainframe computers are the large machines used industrially to process huge amounts of data, named after the "main frame" that used to house the central processing unit. While not as fast as supercomputers, large parts of the healthcare sector depend on mainframes. "The latest generation of these powerful computers can handle nearly 15bn transactions a day, which is more than enough to handle the amount of healthcare data being generated on remote devices", Fagen says. "They are also the most secure platform in use today, which allows healthcare institutions to remain compliant with all relevant data security regulations."
This level of computing power can then enable advanced data analysis through artificial intelligence (AI). "Hospitals can add higher-level algorithms into the mix", Fagen explains. "This unleashes AI to help them find relevant insights buried in the data - insights that might otherwise go unnoticed, and that could make a real difference in a person’s life and health. "By capitalising on hybrid IT environments, healthcare providers can connect insights from the wearable devices back to individual patients’ records, enabling them to take another leap forward in proactive care."