An Inside Look at the Implementation of AI in Healthcare Mid-Revenue Cycle

The use of Artificial intelligence (AI) is expanding and inevitable in health systems. Recently UASI hosted a webinar in collaboration with BKD CPAs and Advisors titled “Artificial Intelligence in the Mid-Revenue Cycle.” The webinar featured an overview of AI and applications in the Revenue Cycle led by BKD and a roundtable panel discussion hosted by Mary Stanfill, Vice President of Consulting Services at UASI. Panelists included Mikki Clancy, Chief Digital Officer at Premier Health and James Ford, Vice President of IT and Digital Technology at Physicians Health Plan of Northern Indiana (PHP).

Excerpts from our panel discussion:

Mary Stanfill: What has your experience been in terms of stakeholder alignment and reaching agreement to investigate AI for a particular use? Who are the stakeholders within the healthcare organization that need to be involved and what has that process looked like?

Mikki Clancy: Before implementing a digital strategy at Premier Health it was a free-for-all. After three years of planning, the Chief Digital Officer position was created and that is the position I hold. I have experience as COO and CIO which are the exact divisions that must be involved in the process. My position reports to the CEO. There is also a Transformation Officer that organizes prioritization within the organization based on the value proposition for each case.

Mary Stanfill: What are the reasons for exploring AI enabled applications within your organization?

James Ford: Our main focus has been fostering a healthier population. To do that, we need more sophisticated analytics. Identifying our sick patient population to enable care management that will prevent further decline saves a lot of money and improves outcomes. Our AI journey is in its very early stages, but I’m excited about the data analytics we’ve seen so far.

Mary Stanfill: There are a lot of different functions that can be enabled by AI and a lot of different AI-applications available out there. How did you identify and prioritize which functions to explore enabling with AI?

Mikki Clancy: We looked at business functions, especially during COVID, that had staffing shortages or large amounts of manual work. Revenue Cycle was high on that list as was operating room block scheduling and staff scheduling.  These are areas that we knew we could make an immediate impact. We tried to use our in-house knowledge base as much as possible but outsourced to vendors when needed.

Mary Stanfill: Can you describe how your organization approached vendor selection? Did you define specific selection criteria? Do you have any specific lessons learned or advice for our audience in this respect?

James Ford: Although our team can build many cool things, sometimes vendors offer more. Training your own team is very time consuming. It took us about 18 months to research vendors before deciding on three. We brought the stakeholders in to meet with our top vendors and from there we decided what vendor could best meet our needs.

Mikki Clancy: Our vendor selection process is standard. Particularly with AI, many vendors are startups though. Which means you will experience growing pains with them. It is important to ask vendors how many clients they have and what their growth plan is. You don’t want your own project derailed because of their growth challenges.

Mary Stanfill: How do you know if it’s worth it? What benefits do you expect to get from the AI-enabled application? Can you share some of the benefits that you anticipate receiving from the applications you are planning to implement (or have implemented)?

Mikki Clancy: We use an agile methodology to identify objectives. We dedicate four weeks at the start of each project to a ‘sprint’ and we use that time to identify the benefits. We use some assumptions and apply percentages. Many benefits are quality and insight related instead of monetary. I also have a scorecard that reports the value that we are bringing. We bring the stakeholders together to review each use case.

Mary Stanfill: Can you provide examples of data that you can get from AI that you never had before?

James Ford: Predictions are key insights we were not able to glean before using AI. We expect AI will help identify patients with undiagnosed conditions or chronic conditions with undiagnosed co-morbidities, etc.

Mary Stanfill: How are you measuring success of an AI implementation?

James Ford: In the future, it should reduce the cost of care, help members become healthier, and help them get the right kind of care at the right time–including prevention and ensure follow ups, etc. This will reduce expenses, enable performance improvements, and help staff and nurses do what they are really good at and get them the info they need at the right time.

Mary Stanfill: How did you improve your outcome(s) using AI?

James Ford: We’re getting the right data at the right time. This should help patients receive the right care at the right time, and prevent future expenses as it relates to fostering a healthy population.

To listen to the full BKD and UASI webinar, including an overview on AI led by BKD CPAs and Advisors and the full panel discussion led by UASI, click here.