Will AI really lower the cost of healthcare? A lesson from a hospital without beds

By:

Sep 29, 2016

Earlier this year I wrote a blog about why computers will not replace doctors in our lifetime. I argued that the current patient care processes and the liability model cannot easily be automated and taken over by a machine, and physicians will remain integral to healthcare. A recent article seems to support this view, as doctors appear to be the early adopters of supercomputers. A group of oncologists in New York City are researching the potential of IBM’s supercomputer, IBM Watson. Rather than look to Watson as a replacement, they are researching its potential to enhance the work that they do, by assisting oncologists with research and treatment recommendations.

Recent advances in supercomputing and artificial intelligence have raised hopes of delivering more precise and personalized healthcare services to people. More articles are being written on how the future of machine learning could mean discovering new treatments for complex diseases like cancer and heart diseases. However, the cost of involving a supercomputer in patient care has not been deeply discussed. Some people may logically conclude that just as personal computers have become affordable to a broad population, supercomputers that provide healthcare services will also become affordable over time. Unfortunately, this is not the case. The theory of disruptive innovation explains that an improved technology, in the absence of changes in its delivery, becomes a sustaining innovation, which increases the costs of products and services.

The history of medicine is full of examples of sustaining innovations, where increased personalization from greater precision in technology has also increased the cost of care. For example, MRIs and CT scans provide more accurate images of individuals’ anatomy and associated problems, but they are also much more expensive than traditional x-rays. Similarly, the cost of an IBM Watson license to a hospital or a group of physicians will be significantly higher than the cost of a common software license. In order to cover this higher cost of technology, physician consultations using IBM Watson—which can access thousands of medical journals in a nanosecond—will have to be reimbursed at a much higher rate than a consultation without the supercomputer. Insurance will likely comply with this request, because a supercomputer-assisted consultation will likely improve the patient’s outcome. While it is great that the quality of patient care has improved, the cost of care has also proportionally increased. A supercomputer in the absence of a fundamental shift in the objective and delivery of care will simply add more costs to the system.  

How do we then change the objective and delivery of care? The latest innovation unfolding at the Mercy Virtual Care Center just outside of St. Louis might provide a glimpse of the future we want. This hospital uses a set of existing technology to make sure that patients are effectively and efficiently taken care of in their respective homes. For example, patients are being managed via video and remote monitoring devices. Although these virtual solutions are not for everyone, they have become viable solutions for many patients.  Not only does Mercy Hospital send patients home to recover, but they are also minimizing their share of more than $40 billion in annual costs from readmissions. Since they do not expect patients to stay overnight and return with complications, they have no beds. With no inpatient population expected, Mercy Virtual Care Center is set to deliver significant savings. By transforming the way patient care is delivered, technology is leading the way to providing low-cost solutions. Along the same vein, if we are looking to lower the cost of care using IBM Watson, a whole new way of care delivery must be established.  

Not all patients will be suitable for a virtual care model. But, we need to remember that the chronic disease population is the highest and fastest growth segment, and most of these patients can be cared for via remote management. Moreover, the aging of the younger and tech-literate population allows plenty of options to use existing technology to deliver low-cost care that is superior in quality and outcome. For those who are still interested in using supercomputers to deliver healthcare, I offer one recommendation. Rather than providing solutions for oncologists and specialists dealing with complex cases, supercomputers should be providing solutions for nurses and retail clinics. By using a supercomputer in situations where a physician is not present, a nurse consultation or a retail clinic experience could be substantially enhanced without adding significant costs. In this way, the supercomputer could be positioned as an agent to help lower-cost nurses and retail clinics disrupt higher-cost physicians. Simple and low-end solutions are how new technology can disrupt the incumbent processes while reducing the costs of care.

Spencer researches disruptive innovation in the healthcare industry. He has over 15 years of professional experience working with U.S. and international healthcare enterprises, most recently as an equity research analyst covering medical technology companies.