Hospitals are looking to invest in new technologies and work-on innovations that will improve the care patients receive.
To learn more about how hospitals are adopting new technologies such as artificial intelligence, telehealth and predictive algorithms, Becker's reached out to Phoenix Children's Chief Innovation Officer David Higginson.
Question: How does the chief innovation officer role differ from other tech-focused roles and C-suite positions?
David Higginson: The CIO role has a strong focus externally on what's going on in other industries like retail or banking, where they may be introducing innovations applicable to health care. At the same time, the CIO's job is to innovate and find solutions to problems, whether they are within the system or with patient experience. I sit on a patient family advisory committee, where I spend a lot of time listening to patients discussing what they are looking for from a family perspective, and this gives us a very important perspective that complements the big-picture view of the C-suite.
My role here is often trying to influence people to try something new. I spend time identifying champions in the organization, supporting them and launching that first version of a solution with the best possible chance of success.
Q: How do you work to get organizational buy-in around AI projects and investments?
DH: I'm very focused on operations. I do sometimes enjoy geeking out over how the math works, but my focus is primarily on how this technology improves operations or makes a difference in patient care. It's more about addressing a problem we're trying to solve and innovating a solution to it. Roughly a third of the things I work on are not that high tech, and we try to use existing technologies to innovate solutions.
Q: Where do you see predictive AI algorithms in the next five years?
DH: I see the use of AI as providing more of an assist, suggestion or prompt to a caregiver versus replacing the caregiver. At Phoenix Children's, we've had success developing and implementing an algorithm that can predict which patients may be at highest risk for malnutrition, and we prompt the physician to note this. If we could look at all 5,000 diseases and then predict which patients are most likely to have which diseases, we can put that information in front of a doctor so they can respond to it. I think predictive algorithms will continue to be successful if they find their way into that workflow as a helpful prompt.
Q: What role can AI play in helping health systems deal with the staffing crisis?
DH: I don't think AI is likely to help with staffing because the inputs are weak. I do think there may be an opportunity to apply it secondarily in terms of identifying when staff may be most likely to leave an organization. There could be an opportunity for action on something like periodically re-educating on maximizing health benefits to try to mitigate turnover, for example. AI may help in retention, but in terms of day-to-day staffing, I'm skeptical and don't think the precision is there yet.
Q: How do you ensure that you're properly training an AI algorithm?
DH: The important thing is to stay operationally focused. We have to address whether real life matches the predictions. When we get to a point where we think it's accurate enough to implement, it must by that time be as good as a human or better. That's our benchmark at Phoenix Children's, and we're continually monitoring to make sure that "as good as a human" performance is maintained.
It should be noted that COVID-19 was a huge disruptor, and so many AI models had to be discarded after the pandemic. We lost the baseline. Models can drift over time, and we just have to keep rebuilding them. We are learning all the time, and it's a brave new world.
Q: What medtech innovations are you most excited about right now?
DH: Bringing telehealth to an inpatient room is something I'm always interested in, especially as we get ready to open two new hospitals in the next 18 months. Physicians don't have to leave to travel to another facility to be there in person and can be much more efficient by joining virtually. It can accommodate an interpreter, a social worker or a family member. It can also be very useful for helping with rounding. I'm really excited about using a technology that birthed itself in COVID to solve a problem that we thought was unsolvable.