How Payers Can Utilize Artificial Intelligence

Sam Leo, PharmD, director of specialty clinical programs at Magellan Rx, discusses how artificial intelligence (AI) is being used within managed care and he addresses issues payers may face in the future.  

What is AI and how can it be used in health care? Are many health systems currently aware of AI?

At its core, AI is using computer systems to perform a myriad of tasks that normally require human intelligence. We are seeing developments and innovations in both machine learning and natural language process technologies that will have health care applications in decision making support, speech and facial recognition, predictive analytics, forecasting and budget impact modeling, care management tools, customer service, drug discovery, and many more. 

Most health system stakeholders are aware of the concept of AI and its potential impacts, however compared to many other business sectors, and due to some inherent challenges around security, privacy, and data structure, the health care industry tends to lag behind other fast adopters. The initial challenge will be building the foundation within our data systems to put us in the position to leverage solutions. Health care data, although plentiful, is still very disjointed and unstructured. 

What AI is currently available for payers?

As of 2016, there were more than 100 startups with a focus on transforming health care with AI, so the available choices are abundant. Payers consequently are challenged to sift through the numerous options to find the right opportunities for them. First and foremost, payers should have realistic expectations and not get caught up in the hype. Once a payer believes they have a strong analytic base and organized data collection processes, my advice would be to start small in a pilot setting and define what success will look like whether it’s to improve patient care outcomes, improve efficiency, improve the patient experience, or other goals. We then need to ensure results are measurable and gather lessons learned to determine if there is a viable path to scaling the technology in a cost-effective manner. 

In the near term, health systems will need to be prepared to evaluate and make formulary and coverage decisions on new health care technologies that leverage AI. Improved modeling can predict impact of new drugs, formulary placement, and specific criteria on utilization and outcomes in a specific population. Furthermore, we will likely see more adoption of diagnostic AI tools. Health plans will need to consider and evaluate how to encourage and reimburse new predictive diagnostic tools, as well as, determine the value in partnering and incentivizing providers who are early adopters. 

Are payers currently utilizing AI? Do you believe more payers will adopt the use of AI in the near future?

While I wouldn’t yet say AI is widespread among payers, there are plenty of examples of organizations exploring pilot concepts and rolling out new AI technologies in certain targeted areas within customer service and care management. Magellan Rx specifically has developed multiple predictive analytic tools that leverage machine learning to proactively identify at-risk patients. We have also recently partnered with AICure to explore the impact of a digital virtual assistant mobile application that uses facial and pill recognition technology to ensure patients are taking medications as prescribed. From a payer perspective, we are interested in the impact on health outcomes and whether patients are willing and able to use it. We’ve been more than happy with the results seen to date and are excited to discuss the pilot and potential next steps at the upcoming AMCP meeting. 

I do think for a variety of reasons we will see more AI adoption in the health care system as a whole in the near future. First, most patients as consumers are welcoming and more willing than ever to accept the use of more digital tools in health care. Second, it will likely become a competitive advantage to the organizations that adapt and grow in the appropriate areas: to improve the patient experience and to help us target and focus on preventative care. 

How does the use of AI impact costs? 

This is obviously very dependent on the type of AI technology we are discussing. In brief, AI can be a tool that helps us use our resources more efficiently to identify high-risk patients and proactively intervene to prevent a negative and costly outcome before it occurs. This of course will have to be balanced with the cost of rolling out such technologies in the first place. 

How will the use of AI impact competition among payers? Will they start to aim to have the lowest reduced costs? 

AI can support efforts to improve patient experience and provide the most effective care to prevent negative outcomes, improve budget and cost impact of formulary and benefit structures, and more effectively identify cases of fraud and abuse. These and others will certainly create market competition among payers.  

Will payers face issues adopting the use of AI? How will they address these issues? 

Organizing and structuring multiple data streams upfront is one of the biggest initial issues. And even after that, there are many other challenges to address, including patient access and inequity (in other words are all patients able to use AI-based care management tools equally), unknown regulatory considerations, impact on health data security, and ensuring any technology is validated and accurate. There is no quick or easy way to address these issues, but instead all need to be carefully considered and evaluated before any vast incorporation of new technology or processes. 

Can  you share what you are most excited about and what you hope the attendees take away from your session?

I am most excited that a topic like this is coming to the forefront in a forum like AMCP. Payers continually are challenged with a changing and adapting market and the introduction of more digital and AI health care tools will be another one of many challenges we face as we aim to ensure cost-effective and appropriate use of medications. I hope payers attending the session will be able to walk away with a better understanding of what questions they should ask and what challenges they should anticipate when evaluating AI-based health technologies. I also hope they can walk away encouraged that some of these tools have the potential to impact patient care and experience in a hugely positive way. 

—Julie Gould