WHAT ABOUT AI & HEALTHCARE TECHNOLOGY?

You cannot avoid hearing about AI. There's extraordinary variation in public perception of what it means. That ambiguity is greatly compounded by both the variability in and frequency of its presentation. In the healthcare arena, I think it's fair to say that, in general, there is some degree of improved accuracy in this perception that has occurred in the last 5 to 10 years or so as the implications of digital transformation have accumulated, avalanched and worked their way into the lives of the involved stakeholders. It was clear that as electronic medical records (EHR) and related systems settled into the processes that make up the functional healthcare system, data capture and thus, discrete data would accumulate. The capturing of standardized data at the point of care was the foundation of how "meaningful use" would be measured and ultimately, how evolving metrics that were put in place and even mandated by federal programs, would benchmark care "quality" and comparisons across providers. Concurrently, mandated and non-mandated data would become more standardized and data would be available to "slice" and combine in different ways, leading to "knowledge discovery" and soon after, the general awareness grew that data would increase exponentially in value if it was captured, standardized, combined and visualized so as to provide "data insights". Predictably, just about everything I mentioned is a developing discipline in and of itself. After the wave of EHR incentives, certification, implementation, adoption and reporting, the wave of data for "business intelligence" came across the greater healthcare arena (the data security and privacy wave was concurrent). This was essentially the awareness of how the organized and intentional use of this data could provide "descriptive analytics", providing insights into clinical care, process, logistics, cost, etc. As you may have guessed, the potential to leverage data for assistance in the form of categorizations and/or recommendations for commonly encountered scenarios wandered into "prescriptive analytics". At the care delivery level, "clinical decision support (CDS)" was incorporated into EHR systems for this general purpose but CDS wasn't specifically making predictions; it was automating routine categorizations for the most part. There is some degree of blurriness between pure categorizations and predictions and a lot of what was marketed as "AI" was in fact, if you broke it down to its bare bones, descriptive or maybe "prescriptive". 

NOTE: I understand that many of the concepts here have been in existence, in development and applied in other industries and even in parts of the healthcare sector for quite some time prior to this more recent intimate incorporation with healthcare. I am summing up the rapid progression of healthcare IT as the integration into workflows and awareness of its concepts became mandatory. 

The ability to make predictions, often with machine learning, helped trigger a spike in awareness of AI in healthcare as standardized data accumulated and storage as well as computer processing became more economically feasible. However, its reception was guarded, especially as you got closer to those involved with care delivery. Providers were still adjusting to a somewhat forced imposition of technology on so many aspects of their work life and there was justified reluctance to acquiesce to technologic systems when it came to decision making. This was a general concern across patients as well. You would not think this based on the press AI in healthcare was receiving, but adoption and functional integration on the part of many healthcare organizations was tepid, to say the least. There was great enthusiasm for the idea of AI, largely for the potential operational efficiencies and thus economic benefits that were possible and touted in healthcare media. 

TO BE CONTINUED...