How Payers Can Gain the Full Benefits of Medical AI

Introduction

Health care insurers are increasingly dependent on clinical data for making business and coverage decisions. However, payers are frequently unable to access and analyze unstructured clinical data from provider organizations due to technological barriers that limit data extraction and data sharing. Without the ability to access and make clinical data usable, payers put themselves at a competitive disadvantage and underserve their members. 

Fortunately, a new generation of medical-grade AI technologies is beginning to transform health care. Able to accurately extract and understand unstructured data at scale, which comprises 80% of all data found in electronic health records (EHRs), medical AI offers payer organizations the promise of better member outcomes, reduced costs, and increased retention. 

However, payers must take several steps before and after implementation to make sure they and their members are getting the most out of this transformative technology. Below are three things payers should do to gain the full benefits of medical AI.

1. Identify use cases and goals

Since medical AI can be deployed by payers to accomplish several useful clinical and business goals, it is important that payers first determine what they hope to accomplish using the technology. Otherwise, lack of clearly defined objectives may result in a lack of focus and intent, which lowers return on investment. There are often several candidate use cases of keen interest to payers. 

Many payers may prioritize improving risk adjustment. With its ability to understand unstructured data, medical AI allows health insurers to obtain a clear picture of a member’s risk profile based on historical and recent data, including medical history, current medical condition, lifestyle factors, social determinants of health, test results, and medications. This information can then be used to calculate what to pay a provider based on predictive analytics that forecast how frequently a specific member will access health care services. 

For other payers, meeting regulations such as the Center for Medicare and Medicaid Services (CMS) final rule for streamlining prior authorization (PA) is a top priority. Due to take effect in 2026, the rule advances requirements around interoperability, turnaround times, communication with providers, and sharing of performance data, which is intended to help alleviate common PA concerns. Medical AI facilitates access to the clinical data needed to help payers further automate PA processes, accelerate decision-making, and provide greater transparency. Critically, medical AI data can then be “human-verified” by clinical experts to ensure the information is accurate.

Others may focus on using medical AI for reducing waste and fraud. Medical fraudsters employ techniques like intentionally submitting claims for services that were not provided, submitting codes for more expensive services than were rendered to a patient, and billing multiple times for the same service. Medical AI can analyze claims data at scale to identify potential fraud (and unintentional coding errors) in ways that rules-based and manual review processes cannot.

Some may want to use medical AI for increasing member engagement, which frequently is undermined by siloed data, limited channels, and poor outreach. Medical AI can smooth the member journey by 1) better understanding member needs through sentiment analysis of members’ texts and call logs and 2) interacting with members using conversational language to keep them actively engaged in their care.  

2. Create a unified clinical data strategy

It is vital to understand that payers who rely on different point solutions for each step in data’s journey from EHRs to actionable insights may lose data along the way. Thus, an end-to-end solution is a critical component of ensuring comprehensive, high-quality data. 

In an era of increased M&A activity and vertical integration, health care organizations typically are faced with terabytes of clinical data in a variety of formats, across multiple siloes and business units. Implementing medical AI offers an opportunity for payers to eliminate silos and build a single clinical data strategy across the organization. Indeed, payers won’t get the most from their medical AI initiatives until they eliminate these functional data silos. 

An end-to-end platform would include pre-processing capabilities – such as splitting file-based documents – along with the abilities to understand and interpret complex medical terminologies in context, structure data, and provide fully auditable output via a user-friendly interface. This helps payers reduce the overall cost, accelerate the program and ensure the data is served up in a way for the organization to act and monetize the outcomes. 

Getting multiple teams and vendors on the same page with a unified solution architecture and a clear plan to deliver these capabilities in a way that maximizes speed to value is no small undertaking. But only by creating a unified and robust data pipeline with internal players, technology, and vendors in sync can payers activate their business initiatives, applications, analytics and teams with useable, structured data.

3. It’s a Long Journey, Stay Committed to AI

AI is evolving and payers need to set themselves up to leverage this technology effectively. Getting the most out of medical AI, however, is virtually impossible without the right tools and right strategy. As payers continue on their AI journey, they should:

• Establish clearly defined clinical and business objectives for AI, measure against target metrics.
• Build a team focused on piloting medical AI and NLP technologies and establishing a data pipeline that produces accurate, complete, and standardized clinical data.
• Document lessons learned in implementing AI and evangelize across the organization.
• Setup a multi-dimensional governance team that includes technologists and business leaders who are tuned in to ROI.
• Integrate vendor partners into the decision-making and leverage strong partners to stay on top of new uses and new advancements in technology.
• Keep pace with regulatory and ethical issues and establish organizational policies and controls.
• Leverage AI in a way that augments or facilitates team outcomes and doesn’t replace them, integrating AI into the experience and designing human-verified workflows. 

© 2024 HMP Global. All Rights Reserved.
Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of First Report Managed Care or HMP Global, their employees, and affiliates.