Anytime a brand new technological advancement makes its way into an industry, there generally is a temptation to anoint that shiny recent toy as an anecdote to all of an industry’s ills. AI in healthcare is an awesome example. Because the technology has continued to advance, it has been adopted to be used cases in drug development, care coordination, and reimbursement, to call a number of. There are an awesome variety of legitimate use cases for AI in healthcare, where the technology is way and away higher than any currently available alternative.
Nevertheless, AI—because it stands today—excels only at certain tasks, like understanding large swaths of knowledge and making judgements based on well-defined rules. Other situations, particularly where added context is crucial for making the suitable decision, usually are not well-suited for AI. Let’s explore some examples.
Denying Claims and Care
Whether or not it’s for a claim or care, denials are complex decisions, and too necessary to be handled by AI by itself. When denying a claim or care, there’s an obvious moral imperative to accomplish that with the utmost caution, and based on AI’s capabilities today, that necessitates human input.
Beyond the morality element, health plans put themselves in danger after they rely too heavily on AI to make denial decisions. Plans can, and are, facing lawsuits, for using AI improperly to disclaim claims, with litigation accusing plans of not meeting the minimum requirements for physician review because AI was used as an alternative.
Counting on Past Decisions
Trusting AI to make decisions based solely on the way it made a previous decision has an obvious flaw: one fallacious decision from the past will survive to influence others. Plus, because policy rules that inform AI are sometimes distributed across systems or imperfectly codified by humans, AI systems can find yourself adopting, after which perpetuating, an inexact understanding of those policies. To avoid this, organizations have to create a single source of policy truth, in order that AI can reference and learn from a reliable dataset.
Constructing on Legacy Systems
As a comparatively recent technology, AI brings a way of possibility, and plenty of health plan data science teams are anxious to tap into that possibility quickly by leveraging AI tools already built into existing enterprise platforms. The difficulty is that healthcare claims processes are extremely complex, and enterprise platforms often don’t understand the intricacies. Slapping AI on top of those legacy platforms as a one-size-fits-all solution (one that doesn’t account for all of the assorted aspects impacting claim adjudication) finally ends up causing confusion and inaccuracy, relatively than creating more efficient processes.
Leaning on Old Data
One in every of the most important advantages of AI is that it gets increasingly higher at orchestrating tasks because it learns, but that learning can only happen if there’s a consistent feedback loop that helps AI understand what its done fallacious in order that it might probably adjust accordingly. That feedback must not only be constant, it have to be based on clean, accurate data. In any case, AI is barely nearly as good as the info it learns from.
When AI in Healthcare IS Useful
The usage of AI in a sector where the outputs are as consequential as healthcare actually requires caution, but that doesn’t mean there usually are not use cases where AI is smart.
For one, there isn’t any shortage of knowledge in healthcare (consider that that one person’s medical record might be hundreds of pages), and the patterns inside that data can tell us lots about diagnosing disease, adjudicating claims accurately, and more. That is where AI excels, in search of patterns and suggesting actions based on those patterns that human reviewers can run with.
One other area where AI excels is in cataloging and ingesting policies and rules that govern how claims are paid. Generative AI (GenAI) may be used to remodel this policy content from various formats into machine-readable code that may be applied consistently across patient claims. GenAI may also be used to summarize information and display it in an easy-to-read format for a human to review.
The important thing thread through all of those use cases is that AI is getting used as a co-pilot for humans who oversee it, not running the show by itself. So long as organizations can keep that concept in mind as they implement AI, they might be ready to succeed during this era during which healthcare is being transformed by AI.
