In health care, the time between study findings and implementation at the bedside is frequently extended. To become a clinical reality, AI-driven breakthroughs must be submitted to and authorized by regulators (different in Europe than in the United States, for example) as "AI/machine learning-based software as a medical device." Hundreds of such requests have already been approved. There's a debate on the quality of the process, but that's a discussion for another time. However, the critical part is that clinicians must adopt the technologies, integrate them into their clinical processes, connect them to electronic health records and other systems, and have health insurance fund them.

‍

Contrast that situation with the current potential for healthcare administrative AI systems. They do have to be integrated with administrative workflows and systems. Still, while the security compliance must also follow the highest standards, cloud- and API-based AI systems are cheaper and make the process much easier.

‍

The case for administrative AI is also considerably stronger in terms of economic value. For example, David Cutler, a Harvard economist, has proposed a series of changes to administrative processes that he argues could save $50 billion in costs and "result in greater satisfaction for both patients and providers."

‍

Opportunities for Revenue Cycle AI

‍

Coding medical treatments for reimbursement and record-keeping is challenging for humans, with over 55,000 codes in the latest version of the International Classification of Diseases. Several companies have already implemented coding assistance AI systems that translate clinical notes into codes, but for now, they still require review by human coders. However, by integrating the past reimbursement data, it's possible to enhance the work of those human coders with recommendations and analyses that were previously impossible to do by a human alone.

‍

Another challenge for humans in the revenue cycle is the estimation of medical bills before treatment. Complex billing and payment arrangements make analysis difficult. Suppose those estimations were partially automatized from previous experience abroad. A similar service in the United States has reached 70% of the estimates created without human intervention and led to a 60% to 100% improvement in point-of-service collections.

‍

Providers and payers also engage in a lot of back-and-forth about payments and when they will be made, which is another excellent area for applying automation. An AI could check claims statuses with insurance companies' accounts payable departments to see whether reimbursements have been made.

‍

While no one likes to admit it, there is a lot of FWA — fraud, waste, and abuse — in the healthcare payment system. For claims analysis, this domain lends itself well to the application of machine learning (and, in some situations, rule-based expert systems, an earlier generation of AI). According to insurers, the return on investment for such systems is among the highest of all AI investments: one large health insurance we spoke with saves a billion dollars per year by avoiding FWA with AI.

‍

Conclusion

‍

We're not saying that health care providers and payers should give up on the clinical applications of AI. The challenges and cycle times for developing and implementing those advances mean that many organizations will also want to consider administrative AI. Suppose that type of AI can substantially reduce the cost of care. In that case, it could be as valuable to the health care system overall — and many patients individually — as any clinical breakthrough.

‍