The healthcare industry utilizes artificial intelligence (AI) to drive innovation across various use cases, including precision medicine, imaging, clinical documentation, and patient communication. However, many healthcare leaders are not using AI to its full potential.
They largely overlook AI’s ability to advance operations and improve margins when implemented as a part of a revenue cycle management (RCM) strategy. In particular, medical organizations of all sizes can use AI-driven medical coding to reduce denials, eliminate backlogs, and accelerate payments.
My radiology practice’s story is an excellent example of how implementing the right AI medical coding solution can completely transform an RCM operations, improve your bottom line, and overcome staffing challenges.
When AI Should Be Part of a Revenue Cycle Management Strategy
Late in 2020, my practice conducted a coding audit which identified concerning deficiencies in our coding accuracy, such as under-coding, inaccurate codes, and missing codes. Due to these results, the need for outsourced medical coding help was clear.
I was aware of the struggles many encounter when outsourcing their coding to offshore sites, primarily issues with consistency, accuracy, and lack of control. To combat these challenges, I looked to AI medical coding instead.
My background is in IT, so using AI for our needs appealed to me. Better yet, interfaces already existed between the AI coding solution and our practice management system, so much of the “heavy lifting” would all be handled behind the scenes and not become another task for my staff or me to juggle.
The Challenges AI-Driven Coding Can Solve to Improve the Bottom Line
Our large number of coding inaccuracies meant we were receiving more denied claims, which only exacerbated our existing coding backlog. We also discovered that, in some cases, we were non-compliant. All of these issues created lost revenue.
The industry standard for coding turnaround time is 24 to 48 hours for office services and 72 to 120 hours for inpatient procedures; our average time between the service date and billing date was ten days. Such a long charge lag negatively affected our payment velocity and created added stress on team members.
Now, with the help of AI-driven coding, we complete almost all of our coding the day after a service occurs. And with higher levels of coding accuracy, we receive fewer denied claims.
How to Leverage AI-Driven Coding to Improve RCM and Mitigate Staffing Issues
If we did not have this solution in place, our coding backlog and inaccuracies would have continued to grow due to team members not having the time to learn and improve. After implementing an AI coding solution, our staff had the breathing room needed to attend training and increase their coding skills.
The AI solution itself was also an excellent tool for learning. There were various instances where my team realized they were not coding correctly based on the completed coding we received back from the AI vendor.
AI-driven coding also helped us navigate staffing challenges. Late in 2021, two employees announced their retirement plans for the coming year. This left me in a difficult situation, especially since we were all working remotely, which made hiring new team members even more challenging. I quickly realized that I did not need to replace the employees.
Instead, we expanded our use of the solution to perform nearly all of our coding instead of roughly half of it. With this significant change in coding workload, our coders could take on the responsibilities of the retiring team members, and we avoided having to hire replacements. With this approach, our staffing needs shifted from 2.5 full-time equivalent (FTE) coders to approximately 0.5 FTE coders, saving us both time and money.
Using this solution also means our coding keeps moving even when team members take time off. And amid the ongoing “Great Resignation,” ensuring your team members can step away from work to recharge is essential to boosting morale and, ultimately, retaining staff.
What to Look for in an AI Platform
Choosing the right AI platform to tackle your organization’s unique operations can feel challenging. However, if you know what to look for, the selection process becomes much easier.
After going through this process myself and having in-depth conversations with the AI-coding platform partner, here are the top three criteria I’d suggest searching for when choosing an AI-driven medical coding vendor.
- A solution that integrates with your practice management system.
Without a tight integration between your chosen solution and your practice management system, you may lose time to operational challenges. While you may gain some time from the coding side of your operations, someone will still need to manage files and any problems that stem from a lack of integration.
- A vendor that initiates regular communication before, during, and after going live.
The ability to communicate regularly with your vendor is crucial to success—especially at the beginning. Implementing this technology and ensuring the best results possible is an iterative process. In the first few months after implementing this solution, the AI vendor led weekly meetings and established a robust method for identifying and addressing problems.
- Transparency regarding what the AI platform can and can’t provide.
Make sure you have realistic expectations from the get-go, especially if you have never worked with AI technology or outsourced coding. Knowing what a vendor can and can’t do will help you avoid unpleasant surprises after you have already invested time and money into a specific technology.
My only regret throughout this entire process is not partnering with the AI vendor sooner. The beauty of AI technology is that it works for medical organizations of all shapes and sizes, meaning anyone can harness its power to make positive changes across their operations. Healthcare leaders looking to boost their coding accuracy and quality, maintain a healthy revenue stream, and overcome staffing challenges should consider using AI-powered solutions.