This article is the second in a three-part series on artificial intelligence (AI) in the health information management (HIM) revenue cycle field. Part 1 focused on what AI is and how it can benefit the healthcare revenue cycle, and Part 3 will focus in-depth on AI use cases present in healthcare today.
In the first part of this three-part artificial intelligence (AI) series, we covered “What Is AI, and How Can It Benefit the Healthcare Revenue Cycle?” In the next part of our three-part series, we will take a look at the considerations for evaluating and planning for AI technology in healthcare revenue cycle. Now that we have an understanding of what AI is, we can look at our current revenue cycle pain points to assess where AI can help ease the burden placed on our workforce, create efficiencies, improve quality, and accelerate cash flow.
According to an article in RevCycleIntelligence, “AI “intelligence” can effectively address the most pressing revenue cycle management issues, such as prior authorizations, claim status checks, and out-of-pocket cost estimates, all while getting the information that needs human intervention to the right person at the right time.”
There is a place for AI at any point in the revenue cycle management process in healthcare. The steps it takes to get reimbursed for a claim can be daunting and require an exorbitant number of touches without the support of technology. AI can often add a much-needed additional layer of quality and efficiency. The steps associated with the revenue cycle start even before the patient is seen and carry through to payment. Those steps typically include preregistration, preauthorization, registration, charge capture, clinical documentation integrity, coding, claim submission, remittance processing, insurance follow-up, and patient collections.
Most AI implementations in the revenue cycle are driven by a desire to change and accelerate workflow in an effort to decrease cycle time and improve efficiency. Too often, the sense of urgency driving the pace at which an AI implementation proceeds is driven by revenue leakage. This urgency can compel leadership to proceed without fully understanding current process capabilities, where improvements can be made, and how to measure improvement to create value proposition.
Deployment of a Kaizen A3 methodology can be a powerful first step in any AI project to assist stakeholders in gaining insight into the perceived existing workflow, current process capability, process variation, and workarounds. The methodology is also invaluable in identifying the critical to quality variables and key performance indicators that will be measured as part of vendor evaluation and post implementation success. An example of a PDSA A3 Template is illustrated in Figure 1. Sample templates for Kaizen tools can be found at the Lean Enterprise Institute.
Figure 1: Example PDSA A3 Template
After potential area(s) of improvement have been identified utilizing a Kaizen A3, you will need to start doing your homework. Speak to your current vendors and see if there is a solution that they offer that will meet your needs. Cast a wide net beyond current vendors and look at other technology that may sit on top of (bolt on) or work with (embedded) current revenue cycle solutions. It’s imperative to talk to organizations who use the types of AI your organization is interested in implementing. When speaking to colleagues in the industry, be sure to talk to organizations similar to yours as well as those who may have some different service offerings. Getting as much insight as possible independent from the vendors you are speaking with is invaluable.
It is also important to read product reviews and use tools offered by companies such as KLAS Research. KLAS is a way for healthcare professionals and payers to share their experience with the solutions they use. This data is vetted, analyzed, and shared. Request demos from any vendor you are considering with specific use cases that meet your organization’s needs. Do not settle for a slideshow presentation of features and functionality when making your AI decision. Be cautious when vendors share with you that they can potentially do something in the future that their product does not currently do. An AI vendor needs to be innovative and evolve their products in tandem with your organization’s needs. If the AI does not meet your needs, don’t settle and keep shopping. Since AI is continuously evolving and new vendors are stepping into the revenue cycle space, be very cautious of vaporware. According to Merriam-Webster, vaporware is “a computer-related product that has been widely advertised but has not and may never become available.”
Attending vendor exhibits at industry conferences such as the AHIMA Annual Conference can be a wonderful time to meet with vendors and explore new technology. Not only is it a good time to connect and network with vendors, but it gives you an opportunity to meet with peers in the industry who may already be using the solutions you are looking at or are on the same journey. It is likely that your current solutions have user groups that may offer their own annual conferences. If you are getting ready to take your revenue cycle to the next level with AI, it is strongly encouraged you utilize these platforms as a time to explore your options.
Budget is always a key consideration when evaluating any new technology solution. At the beginning of your journey, you may not think that you can afford to implement a new solution. This is where vendor discussions and close work calculating your return on investment will help guide your decisions. When calculating your return on investment, be mindful of your current cost and change to future expenses. Hand in hand with your return on investment, you should be considering what key performance indicators (KPIs) you are looking to change by implementing AI. Are you looking for a decrease in accounts receivable days, an increase in productivity, or an increase in accuracy? Do you track this KPI now, or is this something new you will need to measure success?
Organizational constraints are also a key consideration when evaluating AI. Don’t engage with your information technology (IT) team too late in the process. You will want their understanding and engagement on the workload it will take for them to help implement and support the product. They will be able to assist you in determining what interfaces you may be able to leverage, duplicate, or need to have built from scratch. You will need to take into account new hardware considerations in your budget process as well as the labor associated with any work that needs to be done by your IT counterparts.
Another consideration in AI planning is your workforce. Do not forget about the end users and how they will be interacting with the AI. Consider different learning styles and your current culture around continuous learning and self-improvement at your organization. According to an article by Carnegie Mellon University, “AI technologies are evolving so quickly that any specific requirements might soon be overcome by advancement. For that reason, organizations looking to adopt AI need to grow a culture of learning.” A culture of learning is key to success as you embark on your AI journey. Get early buy-in from your staff by sharing your evaluation process and how it will enhance their work experience when implemented.
The information outlined above can be used in any healthcare setting (acute, physician practice, clinic, etc.) when considering and evaluating AI in the healthcare revenue cycle. It is clear that AI in the revenue cycle is here to stay. It is the duty of health information management professionals to keep our pulse on the technology and keep evolving as the industry evolves. In the third part of our AI series, we will focus on use cases where AI is successfully being used in the revenue cycle today.
Kimberly Seery is the associate director of coding and data quality at ChristianaCare.
Michelle Wieczorek is the vice president of coding and clinical integrity at CSI Companies.
Take the CE Quiz