Converting Risk Into Revenue

An AAPC Services risk adjustment case study

The value of risk adjustment is drawing more and more attention at the organization and provider levels. What used to be primarily beneficial to health plans has been trickling down to participating providers in Medicare Advantage (MA), Medicaid plans, and other value-based contracts. This introduces new questions around auditing and compliance:

Is your documentation sufficient to support the risk adjustment hierarchical condition categories (HCCs) your organization is submitting?
Are your providers and coding staff capturing the correct laterality and specificity supported by the documentation?

Sometimes specificity is the make or break between the ability to report an HCC or a higher-paying HCC and not being able to report one at all.

Challenges Ahead

Annually, the Office of Inspector General (OIG) updates its Top Management and Performance Challenges Facing HHS, which targets key issues and vulnerabilities within the Department of Health and Human Services (HHS). These challenges change as new issues arise. According to the OIG, in 2021, Management Challenge 3 is “Delivering Value, Quality, and Improved Outcomes in Medicare and Medicaid.” Included in the challenge is the reason why accurate reporting of HCCs is not only a health plan issue anymore:

Both Medicare (FFS, Part C, and Part D) and Medicaid have proven susceptible to fraud, waste, and abuse. FY 2020 estimates of improper payments range from 6.3 percent (for Medicare FFS) to 21.4 percent (for Medicaid) of total expenditures. Improper payments for Medicare, Medicaid, and CHIP totaled approximately $134.2 billion. These programs are on the Government Accountability Office’s (GAO’s) list of high-risk government programs.

How is the OIG standing up to this challenge, and what are the Centers for Medicare & Medicaid Services (CMS) and HHS doing to ensure accurate reporting of risk adjustment HCCs?

CMS has been auditing Medicare Advantage health plans since 2013, implementing fail-safes to recoup overpayments from plans. This recoupment of overpayments does not only affect health plans but also contracted participating providers and their organizations through Medicare’s Quality Payment Program or the Medicare Shared Savings Program.

Case Study

Background: Between June and December 2021, AAPC Services performed risk adjustment audits using the V24 Medicare Advantage model and dates of service from large physician health organizations and a large hospital system. These organizations employ internal training departments that include risk adjustment training for providers and staff.

Audit: All ICD-10-CM diagnosis codes were reviewed for accuracy and supporting documentation. There were two main goals for these audits:

Identify documentation deficiencies that did not support reported HCC conditions; and
Identify additional HCCs supported in the documentation but not reported by the organization.

These additional HCCs were labeled as missed opportunities and supplied to the respective clients for further data review.

Let’s first review ICD-10-CM accuracy for all codes associated with the claim (both non-HCC and HCC codes).

ICD-10-CM Accuracy

Results:

80% Correct
12% Incorrect
8% Added

As shown in Chart 1, AAPC Services found an overall 80 percent of the ICD-10-CM diagnosis codes were reported accurately within the providers’ medical record documentation; 12 percent of incorrectly coded ICD-10-CM diagnosis codes were due to insufficient documentation to support the reported ICD-10-CM diagnoses; and 8 percent of the diagnoses were supported but not captured in the medical record, identifying these as the missed opportunities.

HCC Accuracy

Results:

62% Correct
14% Incorrect
19% Added
5% Revised

As shown in Chart 2, AAPC Services found that 62 percent of the reported HCCs were supported within the documentation in accordance with CMS Risk Adjustment Guidelines for the Medicare Advantage V24 model. There were still 14 percent of HCCs being reported that did not provide adequate documentation to support the reported HCC and, therefore, the diagnosis was deleted.

As shown in Chart 3, the most common HCCs deleted were:

HCC108 Vascular Disease
HCC12 Breast, Prostate, and Other Cancers and Tumors
HCC19 Diabetes Without Complication
HCC8 Metastatic Cancer and Acute Leukemia

Conditions such as aortic aneurysms (HCC108) require additional documentation to support the reported HCC code; conditions should not be captured from the provider’s statement alone. Best practice is to notate in the documentation diagnostic imaging of the aneurysm size, location, and overall status. If there is mention of a repair in the surgical history, this brings up doubt and a fair question as to whether the aneurysm is still current. If you aren’t sure as to the status of a condition, caution should be enforced.

Another example of deleted codes is for cancer. Providers and coders have continually struggled with coding cancer accurately. As with the aneurysms above, cancer coding requires additional documentation to support its reporting. A provider documenting that a patient has cancer and is being seen by their oncologist is no longer sufficient.

Risk adjustment coders and auditors should understand the definition of active treatment; this is current ongoing medication treatment such as Lupron or Tamoxifen; current ongoing chemotherapy sessions where the provider mentions beginning treatment or states the cycle the patient has next; or toxicology visits post-chemo treatment for evaluation of symptoms and side effects of treatment. Patient refusal of treatment is also a form of active treatment; it does not mean their cancer is no longer current and reportable — they will still need palliative, hospice, and other resources to support their continued care.

Rounding up the rest of the 49 percent of deleted codes includes HCC75, HCC9, HCC96, HCC188, HCC134, and HCC136.

Only 5 percent of the ICD-10-CM codes resulted in a change to a different HCC code. Chart 4 shows the HCCs that experienced the most changes.

HCC19 received the highest number of changes (46 percent). This could be due to several different challenges we’ll discuss later. “No HCCs” came in a high second (36 percent). These were non-HCC conditions listed such as G62.9 Polyneuropathy, unspecified or N18.9 Chronic kidney disease, unspecified, where further specificity was found in the documentation to support a reported HCC. In this case, the polyneuropathy was further specified as “due to chemotherapy,” allowing the auditor to report G62.0 Drug-induced polyneuropathy (HCC75). The unspecified chronic kidney disease was further specified to include the stage, and the auditor was allowed to report N18.4 Chronic kidney disease, stage 4 (severe) (HCC137).

On the flip side of the reported changes, the results show an over-reporting of HCCs. HCC18 accounted for 5 percent of the changes made during the review. Lack of documentation for the reported diabetic manifestation code resulted in a downcoding of the ICD-10-CM diabetic code to E11.9 Type 2 diabetes mellitus without complications, decreasing the HCC value. It is always best practice to have supporting documentation for both the etiology condition and the manifestation to ensure proper reporting of all HCC combinations.

Another category that AAPC Services saw changes in was HCC100. This HCC covers the ICD-10-CM category I63 Cerebral infarction and was in the news not too long ago. In an OIG audit report, 582 enrollees were reviewed for accuracy and 580 of them were found to have been reported incorrectly. This amounted to an overpayment of $14.4 million to Medicare Advantage organizations. That audit found providers were reporting active stroke codes when patients were being seen for follow-up post-stroke, or providers were seeing patients who were thought to be undergoing an acute stroke, but a confirmed diagnosis was not obtainable, and only symptoms should have been reported.

A significant 19 percent of the HCCs supported reporting additional diagnoses that were not captured by the organization:

HCC18 Diabetes With Complication
HCC111 Chronic Obstructive Pulmonary Disease
HCC19 Diabetes Without Complication
HCC85 Congestive Heart Failure
HCC59 Major Depressive, Bipolar, and Paranoid Disorders
HCC96 Specified Heart Arrhythmias
HCC22 Morbid Obesity

As shown in Chart 5, results varied as to why and where within the documentation additional HCCs were supported. HCC18 was the highest at 11 percent. After review of the documentation, it was found that patients were being seen for conditions other than their diabetes; however, the diabetic conditions were addressed and supported in the documentation, resulting in the addition of the HCC. Coders should be mindful of the fee-for-service mentality when coding for risk adjustment: All conditions addressed by the provider affecting patient care and supported on that date of service should be captured.

HCC111 came in a close second with 10 percent of the total adds. Code J44.9 Chronic obstructive lung disease, unspecified was captured 26 total times within the reviews. As with diabetes, patients were seen regarding a wide range of issues such as upper respiratory infections, colds, flu, COVID-19, etc., where the coding was driven by what conditions supported billing services, and the chronic obstructive pulmonary disease was missed, even though the provider addressed the condition and it had an impact on the treatment and care of the patient.

Rounding up the bottom of the highest reported additional HCCs were HCC85 and HCC96. In 12 of the cases reviewed, unspecified atrial fibrillation was found to be addressed and supported by the documentation. In some instances, echocardiograms were performed and atrial fibrillation was noted and addressed by the provider; however, reporting of the condition as an additional diagnosis was not followed through by the coding staff or provider.

HCC Documentation Challenges Persist

The industry is faced with several documentation challenges surrounding risk adjustment and HCCs. Providers and coders do not fully understand ICD-10-CM guidelines for combination coding such as the ICD-10-CM guideline at Section I.A.15 and the use of “with” in the Alphabetic Index. This guideline allows for linkage of two conditions without the direct provider linkage stated in the documentation. If both conditions are addressed on the date of service and are supported by monitor, evaluate, assess, or treat (MEAT), then a combination can be made.

For example, when a patient is being followed by their provider for diabetes and neuropathy and the provider addresses these conditions separately in the assessment, the outcome is E11.9 (HCC19) and G62.9. However, with the use of the Section I.A.15 guideline, the correct coding is E11.40 Type 2 diabetes mellitus with diabetic neuropathy, unspecified, which results in higher-valued HCC18. Without the extensive knowledge of the ICD-10-CM guidelines, organizations chance not reporting addressed and treated HCCs, or they risk reporting lower-valued HCCs even though the efforts support reporting higher HCCs.

Another challenge found is unspecified versus higher-specified codes. This has become not only an HCC issue, but health plans are now denying services when unspecified codes are being reported. As an industry, we are seeing an increase in payers denying claims based on unspecified codes.

Here’s an example of an unspecified HCC code: When “depression” is reported in a medical record, this diagnosis will index to F32.9 Major depressive disorder, unspecified, which does not support an HCC value. If documentation includes a higher level of specificity, then a more specific condition can be reported, such as F33.0 Recurrent depressive disorder, current episode mild, and HCC59 is supported. Unspecified code selection can be caused by a few things such as the electronic medical record (EMR) search engines and/or pick lists being incomplete or not allowing providers to easily select the most appropriate code.

Coders face a larger challenge with problem lists and past medical histories. Lack of updates and attention to these sections result in over-reporting HCCs and an increase in provider queries to update status of the patient’s chronic conditions. Annual cleanup of problem lists and histories ensures the most current, up-to-date reporting of a patient’s chronic conditions.

Revenue Impact and Solutions

Each HCC reported has an average estimated value of $2,500. HCCs that require intense monitoring and additional support are at a higher estimated value than lower-valued, less intense HCCs. If we take into consideration the 14 percent of deleted HCCs, this roughly accounts for 211 individual HCC conditions. This is a loss of about $527,500 and poses a more serious risk of over-reporting patient diagnoses, which can open organizations up for government and payer audits. On the other side, our findings show 19 percent of conditions were added, or 275 HCCs. This resulted in missed opportunity of $687,500. Based on these results, the total net impact would be $160,000 in missed revenue opportunities.

An increased risk for organizations is a review by CMS and contracted carriers, which will impact organization staffing and overall function and could create a compliance risk leading to legal action. There has been a significant increase in legal action and cases in the news recently around health plans providing Medicare Advantage coverage. We cannot assume that CMS and the OIG will stop at the health plans. All provider organizations need to be prepared and ensure their ICD 10-CM documentation and coding will pass scrutiny.

The OIG’s Challenge 3 is a continued effort and is not going away anytime soon. Protect your organization by reviewing internal policies, performing risk adjustment audits, and hiring an outside audit vendor such as AAPC Services to conduct your audits and trainings. There are also technology services that provide options for increased review of HCC recapture and missed opportunities. Wherever you are in your risk adjustment journey, know there are solutions and ideas out there specifically for you.