What’s Blocking Real-Time Predictive Analytics and Machine Learning in Healthcare?
Artifical intelligence is going to revolutionize healthcare, right?
We’ve all heard the promises of faster, more accurate, and personalized medical treatment facilitated by incredible computing power and machine learning, yet when walking through North American hospitals, this future seems unlikely to manifest anytime soon. Hospital IT infrastructure is mired with antiquated equipment procurement processes and technology unsuited for the coming era of artifical intelligence and the Internet of Medical Things (IoMT).
When it comes to medical device purchasing, hospital procurement officers typically only ask that a single data export criterion be met for new medical devices — that devices integrate with the hospital EMR systems.
The problem is that single-stream connectivity with low-resolution numeric data may be adequate for the EMR, but is wholly unsuited for data-intense algorithm-based processing which is at the heart of all machine learning and predictive analytics systems that use real-time medical device data.
Rarely will hospital procurement teams ask deeper questions about medical device data export capabilities. Questions about data quality, available parameters, patient data binding, and time syncing/stamping are left unaddressed. Clinicians and researchers are saddled for many years with biomedical equipment ill-fitted to meet their current and future analysis needs.
And if you think machine learning and predictive analytics are still ten years down the road, you’d be surprised at the number of companies and projects that are already analyzing real-time medical device data. Take a look at companies like Etiometry and CleMetric who are already providing innovative algorithmic-based analysis with device data.
I’ve put together a short questionnaire that can help you evaluate your potential medical device purchase for suitability concerning real-time data analytics and machine learning.
Additionally, it is important that current and future hospital IT infrastructure be assessed during new equipment purchasing decisions. Even if biomedical devices meet the guidelines outlined in the questionnaire, the IT infrastructure may not be able to support or take advantage of this functionality.
I’m always available to discuss medical device connectivity, integration, and interoperability. You’ll find me on Twitter at @TrueProcess_JM
Interview: The Value of Medical Device Connectivity: Sitting Down with Doug Frede
Download our interview with True Process President Doug Frede to get an insider’s view of the business value of medical device connectivity, including:
internal vs. external value propositions, the technologies pushing connectivity forward, and the four basic challenges currently preventing widespread connectivity adoption.