Nov 2014

Medical devices generate a ton of data

For several years I worked for a software-based clinical solutions company, and became intimately familiar with the medical device industry, and the large volumes of data created by medical devices. My role involved many different things, mainly developing and promoting medical device integration (hospital IT infrastructure), and enterprise patient monitoring (nurse's stations and alarm management). Here are a few thoughts on the growing trend and awareness of medical device data and analytics.

Is innovation possible?
First of all, the regulations and quality system requirements imposed on makers of FDA guided medical devices makes it very difficult to innovate. Every 510(k) submission has to have predicate devices and fit into one category or another. Software based products are given the same scrutiny as hardware devices. The rising trend of wearable devices (such as FitBit) are placing people's health in their own hands. Can this trend influence the medical device market? With Covid, this process may have been overhauled, but still, the incredible amount of documentation required is still a huge barrier to innovation, one that many companies do not have the resources to do.

Data patterns
One thing I learned about medical device data is the limited commonality (or lack of shared data) between medical devices. A whole business was conceived to connect and extract data from devices that were never designed to be networked. Hospitals employ a variety of different (proprietary) devices to treat and monitor patients for various conditions, but sadly these devices never communicate with each other. Device data should be accessible to a wide range of output devices, and this data should feed the hospital's EMR so that patterns emerge to guide patient treatments that prevent adverse outcomes.

Data Usability
Medical devices monitor patients for specific conditions, and are designed to alarm at any sign of deviation from set/predetermined limits. This alarming concept stems from basic liability issues, however it has brought about a huge problem in healthcare known as alarm fatigue. Devices need to be more contextually sophisticated and be able to input and output usable data for actual human beings to interpret, as well as have the ability to interact with other devices for patient (and clinician) sensitivity and context.