AI in healthcare—Time for action, not barriers

Artificial intelligence has made its mark in many different industries. So why does it seem that AI is struggling in healthcare?

I think part of it is cultural—a reticence to change, and fear of legal issues and personal impacts. I also think that there is a barrier based upon specific use cases, which should not stop broad AI use today.

Why is there unease over AI in our industry? What really is different from what we have now—except that the evolution of better support for the delivery of healthcare is changing faster?

• Workflow optimization efforts are not new.
• Clinical decision support is not new.
• Diagnostic support tools are not new.
• Research data results are not from “unbiased patient sets”—yet we have applied data like Framingham in practice for years.

Let’s face it. Despite the best data sets and machine learning, AI will never have 100-percent completely “unbiased” algorithms or be 100 -percent accurate. Neither will humans. New interns don’t have documented or consistent “algorithms” for how they work—at least that is what I am sure my attendings said about me. Fortunately, I was trainable.

While being under continuous pressure to lower cost and improve outcomes, providers are part of the barrier to AI adoptions. However, the largest part of healthcare cost is not due to provider actions. Many of the inefficiencies of current health IT systems may well be mitigated by AI.

Burned out by sub-optimal IT, administrative burden and time pressure, physicians should lead the charge about changing the healthcare operations and model of care. We should also be embracing the use of these resources to support health, and not simply looking at how the tools may transform our jobs or work processes. AI application and more data availability can support transitioning to a proactive approach from our current reactive system.

We need our culture to change. All of the components of the AI ecosystem (machine learning, natural language processes, deep learning, decision support and others) can provide the paradigm change to move toward the Triple Aim of healthcare. It will involve a change in how we think about care, train physicians, educate patients and how the system approached administrative task.

Within the workflows of physicians, practices, hospitals and insurers, artificial intelligence brings similar opportunity to augment—not replace—the value of employees. Just like the industrial revolution the type, number, and skills of employees will change.

We need to be more proactive and innovative. Here are the key actions you should be taking:

• Educate yourself on how AI has transformed other industries.
• Examine how that can be applied to utilization management, billing, logistics, and other less glamorous back office needs.
• Champion the use of tools for more efficient use of our skills, and the skills of all member of the healthcare team.