It’s hard to imagine a time when claims were submitted on paper. In 2003, Medicare required electronic submission of all claims via an electronic data interchange (EDI) for auto-adjudication. This means claims are processed, paid, and have a status update without human contact. Plans should aim for an auto-adjudication rate in the high 90’s, but some systems still struggle with this benchmark.

There are a number of system errors that can impact auto-adjudication rates including faulty matching logic, incorrect claims information, and missing provider contracts to name a few. Information maintenance is critical to a successful auto-adjudication process, otherwise it front-end rejects or receives a pending status. In these cases, a manual correction is required and not preferable, as human resources are costly. In order to stay competitive, plans have to invest in a strong adjudication engine that can handle high volume.

Another area of investment that behooves plans to prioritize is the configurability of their claims processing technology. A high level of configurability allows for faster response to change such as: CMS regulatory updates, changes to the health plan’s plan types, benefit plans, contracts, provider contracts, capitation-related contracts, payor information, bank information, group information, member configuration, authorizations. There are a number of variables that must be accounted for and having a system that is highly configurable takes the edge off of changing with the times.

Host: This is Episode 7, Innovations in claims and payment models. We’re welcoming guest Kirsten Lynch. Kirsten is a claims technology expert and has been in the industry for over ten years. How are you today, Kirsten?  

Kirsten: Great, I’m excited to talk claims with you today.

Host: Same. Kirsten, to get started, would you give us a historical overview of claims and an idea of where the industry is moving in the future?  

Kirsten: Yeah, so prior to technology, providers would submit claims on paper. These forms were known as the HCFA-1500 form for professional claims and the UB-92 Form for institutional claims. Institutional claims are also known as Inpatient and Outpatient claims. The UB-92 was discontinued in 2007 and replaced by the UB-04 Form.  

In 2003, Medicare required all claims to be sent electronically. This resulted in a large shift away from the manual forms towards the use of ANSI X12 transactions to load claims into systems that were capable of auto-adjudicating them. This shift changed everything. The obvious benefit of this technology adoption is that there’s a significant reduction in errors.  

More recently, claims processing at its core hasn’t changed a whole lot, but what has changed is the way providers are paid. The current fee-for-service model is phasing out, you know, where providers get paid per claim via electronic fund transfer, EFT. The problem with this per-claim model is that there is no real accountability for the providers to provide quality healthcare to the members.  

What we’re seeing now is this movement towards capitation models, so providers are accepting part of the risk. Instead of getting paid per claim, providers are getting paid on a per member per month basis. What this means is that providers are getting reimbursed a set amount depending on the number of members they’re serving per month and the quality of their services. The monthly members are monitored by the enrollment files sent from CMS each month, and the claims records indicate the service usage. CMS sets the rates for services, and providers agree on those capitated rates prior to entering the agreement. There are a number of different types of capitation models but this is the most simple model.  

Host: That gives us a good background—thank you. So earlier, you were telling us about the transition to technology and automation in claims processing. Obviously, this has clear benefits for reducing error and maximizing efficiency. Are there any unique problems that have occurred with the adoption of claims technology?  

Kirsten: Yeah, technology really transformed the way health plans do business in a number of positive ways, but it’s a complex area, so it’s not without its problems.  

Most health plans use a claims processor, and they send their claims files to the vendor through an electronic data interchange (EDI). The various types of EDI ANSI X12 files are identified by number. So, a standard claims file is an 837. This could come to the claims processing vendor from the health plan, the provider’s office, or from a member portal. The idea behind a high auto-adjudication rate is that when the file hits the claims system, you want that claims file to be able to process through the system, adjudicate, which is actually a claims status for us— which means successfully processed so that it can be paid without a human touching it. You don’t want it to pend, and then a person has to manually correct it. You want it to flow nicely through the system, right into the check run and payout, or right into capitated status so it will calculate into capitation. Everybody wants to aim for a high adjudication rate, typically in the high 90s. This is an area where some systems can struggle.  

Host: What are some ways that auto-adjudication can be disrupted?  

Kirsten: So there are a number of things that impact auto-adjudication rates. Faulty matching logic is a big source of problems. Most claims systems have some kind of matching logic that matches data from the claims to data in the system. You have to make sure you can match the fields on the file to the data in the claims database. Then, the information on the claim has to be matched to the adjudication functions themselves. Here’s an example of what I mean by that. One of the big processes in adjudication has to do with provider data. For a professional claim, think about going to your PCP just for an annual exam. An institutional claim would be a hospital visit for an operation. For adjudicating a professional claim, there could be multiple types of professionals on that claim. If I went to my PCP, Dr. Smith, at the University of Cincinnati, and then there’s my rendering professional, which might also be Dr. Smith if I just so happen to be going to see my PCP for services, and there could also be a referring provider on there as well. So you see, it can get very complex when there are a number of professionals and providers on the same claim. Part of what adjudication will do is take the information that comes in on the 837 and the information that comes in from the provider load file and match it all up in order to adjudicate the claim. It’s important that everything is matched correctly because providers have contracts in the system that explain how the provider should price and get paid.  

Going back to the example of matching logic and how that works. The claim comes in and says the University of Cincinnati is my billing provider, and that matches the University of Cincinnati’s billing contract. Dr. Smith is my rendering, and it’s mapped to Dr. Smith’s billing contract.  

Sometimes, the claim information is incorrect. You can have claims front-end reject so it never gets to the system, or claims can pend to where it needs a manual approval to move forward. For example, the logic finds your billing provider but not your rendering provider, so the claim is rejected. These are the types of errors that require a person for manual correction.  

Another type of error that would require manual correction is if the provider contract isn’t on file. The updated information would have to be manually added, and then the claim would have to be re-processed.  

The important thing for plans is to make sure that they have a strong adjudication system that can handle high volume. One of the novel solutions is to offer adjudication scaling solutions, which can be anything from multiple adjudication engines that offer the ability to process more claims in a given time span. Or, if the plan is having difficulty getting through the process, maybe the matching logic needs to be updated or a review of the pends that are happening to determine if a system update is needed.  

Host: That’s a lot of coordinating. It’s hard to imagine a time when automation wasn’t driving this process. And we haven’t even talked about the impact that regulatory changes have on the claims process.  

Kirsten: Yes, a lot of the changes within claims are driven by regulatory changes, and that occurs multiple times a year. An example of a really impactful change in the past couple of years is a regulation lovingly called MIPS. It’s the Merit-Based Incentive Payment System. It’s actually pretty cool! So what it does is it allows certain professionals to take part in specified quality metrics throughout the year and report back to CMS. If, at the end of the reporting period, they met the metrics goals, what would happen in the claims system is they earn a percentage bonus payment on the claim. On the other hand, if a provider participated and didn’t meet the metric goals for quality standards, they could be deducted a percentage up to a certain percentage. For example, the 2023 MIPS payment adjustments vary between -9 % and +2.33%. So there’s risk and reward. This was a major change for claims applications.  

Host: That’s interesting. You mentioned that major changes occurred because of this. What kind of changes?  

Kirsten: To implement this and update the technology for the MIPS calculation, we had to look at claims with only non-participating professionals, or professionals that do not participate in the health plan’s network, that used services on the Medicare physician fee schedule. There were some services that weren’t applicable to MIPS, so we had to also create some conditions that excluded, for example, anesthesia services. The CMS regulation allowed for the calculation both with or without member cost shares, which are things like copay or deductible—so we had to update the system both ways. Then the payments could be either real time or as a monthly settlement. These are just a few examples of the CMS regulatory options that had to be taken into account. This required a highly configurable system to keep pace here because the program parameters were so expansive and complex. And really, I would say this isn’t just specific to MIPS; one of the major areas of concern in the industry as a whole right now is having claims processing technology that’s highly configurable. The reason this is a big deal is because the industry evolves at a rapid pace, and plans have to be able to adapt and pivot quickly with those changes to keep up.  

Host: Okay, Let’s talk more about configurability. What are some ways that today’s claims technologies need to be configurable?  

Kirsten: Yeah, there’s configurations for regulatory changes, changes to the health plan, adding plan types, benefits, benefit plans, contracts, provider contracts. Let’s see what else is on the menu? Capitation-related contracts, payor information, bank information, group information, member configuration, and authorizations. You know, a high level of configuration just allows for faster responses without having to change system logic.  

Of course, you’ll always need some manual input for review. You know we mentioned earlier if something isn’t in the system, you’ll have to add it. High dollar amounts need manual approvals. Another CMS-related regulatory item is called the preclusion list. CMS says you can’t process claims for the providers on this list. Typically, people will end up on this list for fraud, waste, and abuse. First, they’ll go through a pending status phase, where it needs manual approval to move forward with the claim. It’s basically a watch list. Pending can occur at contract levels, professional levels, product levels, and product network levels. Pends can be added for different situations in which a manual review is preferable. So it’s important to have automation because it’s too much information for humans to churn through accurately, but you can’t replace the human element completely. You’ll always have scenarios that require discernment.  

Host: Thank you, Kirsten. You’ve given us a really deep dive into what’s going on behind the scenes in the claims world. I know talking with you has really enhanced my own appreciation for the complexity of the process. It was great talking with you today.  

Kirsten: Thanks for having me. Always happy to talk claims with you!

Host: Thanks to everyone joining in today. Don’t forget to follow the podcast, and if you liked this episode, share it on LinkedIn with your colleagues.