IBM sort of seemed like they were asking for it. From their claims to revolutionize cancer care in two years to their multi-billion $ growth projections, they worked real hard to project as healthcare’s knight in shining armor. Of course, those inside of healthcare have seen many such knights burned to a crisp by the healthcare dragon – from Google to Microsoft.
Now the Monday morning quarterbacking begins. Analysts, journalists, and otherwise opinionated people like me get to breakdown the failings of others. Most will make accurate observations. Healthcare is complicated, other vendors put up EPIC battles to prevent data sharing, hospitals are tough customers, etc.
In doing so, they will miss the most important lesson of IBM’s experience.
The greatest threat to the adoption of machine learning technology in healthcare is the lack of incentive to learn.
It is not the overly ambitious projections of CEOs, the incorrect but often made argument that clinicians are slow adopters of technology, or the tightening of budgets. If it was the job of hospital executives to figure out how to deliver care as efficiently and effectively as possible, IBM and other tech companies would have long figured out how to make money in healthcare. After all, the ability to learn from data and automate efficiency via software has led to the most productive economic era of our times.
However, for as long as we support a system that gets paid based on volume and complexity, you can be sure that any new information technology will be applied for the purposes of maximizing volume and complexity. We’re already seeing it. The few AI / machine learning companies that are succeeding in the hospital / clinic space are optimizing scheduling and maximizing reimbursement through coding (documenting complexity).
That means the full potential of machine learning to help us analyze, optimize, prophesize and customize will only be useful for the 5% to 12% of US healthcare that is delivered in true value-based models – not coincidentally, the space Cyft operates in. In the meantime, we’ll continue to use AI / ML elsewhere to revolutionize how efficiently we sell ads, books, movies, products, drill oil, run casinos, pick up waste……