TABLE OF CONTENTS
Guest Speaker
Introduction
Transcript
Accurate coding in healthcare is critical, especially with the recent changes brought about by the RADV extrapolation rule. This regulation allows the Office of the Inspector General (OIG) to apply audit findings from a sample across an entire Medicare Advantage plan, significantly increasing the financial risks associated with coding errors. For example, a $500,000 penalty for mistakes in coding can now balloon to millions, making it essential for health plans to prioritize coding accuracy and compliance.
To navigate these challenges, the OIG has developed helpful resources that guide health plans in identifying high-risk diagnosis codes and highlighting common mistakes to avoid. Organizations can leverage this information to enhance their coding practices and protect themselves against potential audits.
Tune in to learn more about effective strategies and resources to improve coding accuracy and compliance in your organization.
Guest Speaker
Kristi Reyes is the Director of Risk Adjustment Coding Operations. She leads a team of medical coders from various aspects of risk adjustment—including retrospective, prospective, CDI, QA, risk mitigation, and RADV. Kristi holds certifications in Certified Professional and Outpatient Coding, Risk Adjustment Coding, and Auditing.
Host: Today we’re talking with Kristi Reyes about her Lessons Learned From the OIG. Kristi is the Director of Risk Adjustment Coding Operations. She leads a team of medical coders from various aspects of risk adjustment—including retrospective, prospective, CDI, QA, risk mitigation, and RADV. Kristi holds certifications in Certified Professional and Outpatient Coding, Risk Adjustment Coding, and Auditing. Welcome Kristi.
Kristi: I’m excited to be here!
Host: Let’s level-set. The Office of the Inspector General, the OIG, conducts compliance audits on Medicare Advantage plans to ensure CMS submissions for risk adjustment programs are supported in the medical records.
Kristi: Yeah, so they publish these audit findings which is super helpful for health plans to review to make sure their coding guidelines align with what the OIG audits are looking at. Side note, some of my coworkers and I were able to attend a coding conference where the OIG was a featured speaker. We actually got to ask them questions.
Host: Were you geeking out?
Kristi: Yes, we were. We were posting on LinkedIn, “We are getting a photo with the OIG!”
Host: “laughs..”
Kristi: A coder’s version of exciting. (Laughs) But anyway, we got to ask our questions to the OIG. This is the government agency that oversees who determines what passes audit and what doesn’t.
Host: Let’s talk about why coding accuracy is more important now than ever.
Kristi: I think you’re hinting at the Data Validation Final Rule, RADV for Medicare Advantage plans.
Host: Exactly.
Kristi: Okay. Prior to the extrapolation rule, if the OIG ran a sample of 200 charts and found 200 chart errors, the plan only had to pay for the 200 errors. Now, with the RADV extrapolation rule, moving forward in 2018 dates of service and beyond, if a 200-chart sample contains 200 errors, they are going to extrapolate those 200 errors across the entire health plan population. Something that might have previously been a $500,000 refund due to CMS, is now going to be extrapolated across the entire health plan membership and could be in the millions of dollars.
As a specific example, the OIG recently did a review of high risk codes submitted by a health plan and found 100% of colon cancer codes were in error. Since that audit was for 2016 dates of service, they mandated the health plan pay them back for those codes. With the RADV final rule, if this had occurred in a 2018 date of service, all colon cancer codes would be extrapolated across the entire member population.
Host: So, what’s the plan? How are health plans doubling down on accuracy?
Kristi: Plans are looking to the OIG for guidance. The OIG has an online toolkit called, Toolkit to Help Decrease Improper Payments and Medicare Advantage Through the Identification of High-Risk Diagnosis Codes. It’s a 48-page PDF about how to identify high-risk diagnosis codes that are at risk of being miscoded. They are providing the actual codes they are auditing. They are trying to help us help ourselves by showing us the problem areas that they see. And they are providing information on why these specific diagnosis codes are at high-risk. The point is to ensure proper payments. The OIG provides details on the different scenarios that commonly lead to miscoding. And, they provide the actual computer codes, the sequel programming language, to query CMS’s system. They are giving health plans the recipe to their audits so plans can improve their coding accuracy and proactively identify compliance issues. This is amazing information for anyone involved in coding for Medicare Advantage.
Host: This sounds like a great resource for MA plans. Can you give some examples of the most common coding errors?
Kristi: Yes, as of November 2023, the toolkit lists the eight most common errors. The highest risk for error is the code for acute stroke in the outpatient setting. Anytime a patient is having a stroke, it should be managed in a hospital setting. It would be rare to have a patient diagnosed with a stroke in an outpatient setting unless they were getting loaded into the ambulance en route to the emergency room. OIG found that 96% of acute strokes in an outpatient setting were paid in error. In actuality, it was a history of. The difference for Medicare Advantage is that an acute stroke has a RAF, a risk adjustment factor coefficient value, that is paid to the health plan. Whereas a history of a stroke has a zero RAF score, which means zero payment to the health plan. This is why accuracy in coding is so important. Another high-risk code that OIG is targeting includes acute heart attack, or acute MI. This is very similar to acute stroke in an outpatient setting with a few different nuances. Another one is embolism blood clots, which has an error rate of 79%. This depends on the type of blood clot. If it’s a pulmonary embolism in the lung, that’s almost exclusively seen in a hospital setting. Blood clots in the legs could be outpatient, but probably would have some hospital claims to support the diagnosis.
There’s also different types of cancer: breast cancer, lung cancer, colon cancer, prostate cancer. There are very specific coding guidelines on what constitutes active cancer versus a history of cancer. The overarching guideline with cancer is that there has to be evidence of current treatment and active disease. Let me talk about breast cancer. The current treatment could include anything from chemotherapy, hormonal therapy, or surgery for tissue removal. It could be the patient is diagnosed and refusing treatment, or maybe they are referred to hospice. Something needs to indicate that the cancer is a current, active condition.
Now, let’s look at a miscoded incident of cancer. If a patient had cancer in 1998 and they went through chemo. Today, they are doing well. Of course, they will have ongoing monitoring for the recurrence of cancer with oncology. But, this is a history of, not active cancer. It’s a historical diagnosis.
For Medicare Advantage payments, there is a payment associated with active cancer under treatment as part of the risk adjustment factor that is paid to the health plan. Whereas a history of cancer, is a zero RAF score.
Host: What about errors like a coder transposing numbers in a diagnosis code? That’s bound to happen.
Kristi: Absolutely. I’ll use an example that was common with ICD-9. I’m using ICD-9 as an example because the OIG is currently behind on their audits, and the last published audit was around 2017 dates of service. During dates of service where we were still using ICD-9 diagnosis codes, ICD-9 had all numeric diagnosis codes. A common miskeyed diagnosis that was on the OIG’s radar, was for ICD-9 diabetes type II, uncomplicated. The code is 250.00. Now, lymphoma, a type of cancer, has an ICD-9 diagnosis code of 205.00. You can see the potential for this mis-keyed diagnosis. At a payment level, lymphoma has a higher reimbursement than uncomplicated diabetes. Obviously, there’s a different level of care for those conditions.
Host: What about about the current codes for ICD-10? What should plans be looking for there?
Kristi: OIG audits for the last two years have brought to light details about what constitutes a nonspecific depression versus major depressive disorder. The OIG wants to ensure that diagnosis codes are correct, and they want to prevent up-coding. If the patient has depression NOS, that is a zero dollar RAF value for Part C HCC. There is no reimbursement as a chronic condition. However, if the patient has major depressive disorder, that does have a RAF value and is associated with a reimbursement. It’s important that coders are making the distinction between these two mental health conditions. The OIG is looking for support in the medical record. Is there a prescription for an anti-depressant? Are they undergoing therapy?
I think this type of scrutiny is going to increase as we move into the V28 model for risk adjustment. Right now, we’re using a blended model of both V28 and V24. But, starting with 2025 dates of service, we’ll be using V28 one hundred percent. Under V24, all of the major depression diagnoses risk adjust and result in a RAF score. But with V28, only part of the major depressive diagnosis codes are going to risk adjust. Unspecified will go away; the condition has to be reoccurring, moderate, or severe. I have a feeling that the OIG is going to place a heavier focus on mental health disorders in years to come.
Host: What is the process for the OIG audits? Is the process public?
Kristi: OIG sends a notice to the MA organization. They usually audit a sample size of 200 and evaluate the accuracy of the high-risk diagnoses codes. They use the ICD-10 guidelines published by the World Health Organization and the RADV Medical Record Reviewer Guidance, which is a wonderful website that explores the grey areas of what is acceptable for coding—everything from particular coding scenarios to “Is the signature on this page acceptable?” The OIG also recognizes the AHA, American Hospital Association Coding Clinic, as an accepted source of truth for the Medicare Advantage OIG audits. AHA is a group of coding SME’s that get together at the American Hospital Association and review the grey areas of code. A great example, we talked earlier about cancer. If you remember, there has to be evidence of active cancer. Well, what does this mean? This could be widely interpreted. So, what the AHA Coding Clinic does is publish answers to these types of questions. What is considered active treatment for cancer? What if the patient refuses treatment? The AHA Coding Clinic provides clarification on the grey areas of ICD-10 codes.
Host: So, how do we use these OIG audits to run a compliant coding program?
Kristi: Health plans have the OIG Toolkit. The codes identified in the toolkit as high-risk should be the ones plans prioritize and focus on. Aside from preparing for a potential government audit, this information is beyond valuable for guiding internal audits, and policy and procedure improvement. So, it’s definitely helpful for proactive improvement. Our organization has a targeted audit system in place. We re-review 100% of the conditions addressed in the OIG work plan. For example, we address acute conditions; we look at all acute strokes, all MIs, and all cancers. UST HealthProof has implemented this internal audit sampling to ensure we send these codes to our clients and/or CMS as fully compliant and supported in the medical record.
Another use case is MA organizations can establish ongoing monitoring and reporting mechanisms to track compliance. It’s imperative that organizations have access to reporting tools to identify how miscoding occurs. As an example, plans should be able to click a button and run a report on how many acute stroke diagnosis codes were keyed in by a coder in an outpatient record single date of service. Organizations can also conduct a comprehensive risk assessment against the predetermined areas to identify vulnerabilities. Organizations should run reports on their member profiles. Maybe I want to know how many of my members have had strokes in the last year. If 50% of your members have had a stroke in the last year, you might want to take a look at your policies, procedures, and what you’re coding. Because the vast majority of those will probably be a history of. You might also want to drill down to the data at the provider level. You might find your providers need some continuing education on documentation requirements. So, it’s important health plans utilize the sequel language to do a risk assessment of the member population. UST HealthProof has reporting functionality to show us at the category level, the diagnosis level, the coder level, and even when the codes were entered. I would encourage MA organizations to understand the internal parameters and structures that are in place to run those reports and monitor them.
The OIG is providing us with the step-by-step process for how to get through their audit. There are no surprises. They want health plans to do well on the audit. The overarching goal is accuracy of coding that supports a compliant medicare advantage program.
Host: Kristi, this is amazing information. Thank you for introducing the listeners to these resources!
Kristi: Yes, I’m going to get you links to AHMIA compliant query brief and the CMS Risk Adjustment Participation Guide.
Host: Awesome!
Kristi: You bet!
Host: Thanks to our listeners. If you liked this episode, follow on Apple, Spotify, and all major podcast apps.
Resources:
CMS RA Participant guide https://www.csscoperations.com/internet/csscw3_files.nsf/F/CSSCparticipant-guide-publish_052909.pdf/$FILE/participant-guide-publish_052909.pdf
https://ahima.org/media/51ufzhgl/20221212_acdis_practice-brief.pdf
https://oig.hhs.gov/oas/reports/region7/72301213.asp
RADV Medical Record Reviewer Guidance
https://www.cms.gov/files/document/medical-record-reviewer-guidance-january-2020.pdf