Rebecca Welling (rebecca.welling@providence.org), Senior Director Coding Compliance, Providence Health Plan, Portland, OR.
Welcome to the world of risk adjustment and Medicare Advantage!
The Centers for Medicare & Medicaid Services (CMS) developed risk adjustment as an alternative program to traditional Medicare to offer expanded benefits, such as dental, fitness, chiropractic, and eye exams, to encourage healthy patient behaviors and treat seniors at a lower cost than traditional Medicare. Risk adjustment uses private insurance companies to administer the Medicare benefit by allowing patients to select a Medicare Advantage plan offered through many large payers such as UnitedHealthcare, Aetna, Humana, Blue Cross Blue Shield, Providence Health Plan, and many others. In 2021, there are more than 3,550 Medicare Advantage plan choices available nationwide.[1]
Currently, 39% of all Medicare beneficiaries, or around 24 million people, receive their benefits through a Medicare Advantage plan, with that number expected to surpass 51% by 2030.[2] CMS uses severity-of-illness measures such as age, gender, Medicaid eligibility, and chronic conditions reported through International Classification of Diseases (ICD-10) codes on insurance claims to estimate the risk of the patient.[3] Risk is defined as the measurable or predictable chance of loss, injury, or death to which a patient is subject before receiving some healthcare intervention.
Risk adjustment allows CMS to compare patient’s illness severity, quality of care received, and payer performance across organizations, practitioners, and even communities that have a high Medicare Advantage penetration rate. Every participating health plan receives funds annually that correspond directly to the disease burden of the patients they insure. Each patient is given a risk adjustment factor (RAF) score. This score is additive and takes each contributing condition into account.
Here is an example to illustrate the funding:
Ruby Mae lives in Sacramento, California, in her own home with a caregiver. She is 94 years old and is eligible for Medicaid. Ruby Mae has type 2 diabetes mellitus with peripheral neuropathy, angina, and dementia.
The CMS benchmark for Sacramento is approximately $901.77* per month. Ruby Mae’s demographic RAF score is 0.689; in addition, CMS adds 0.287 for her diabetes, 0.37 for her angina, and 0.131 for her dementia. Thus, her total risk adjustment score is 1.477. The average RAF score for a Medicare patient is 1.0, thus Ruby Mae most likely will have higher medical costs than an average Medicare patient.
The formulary for the payment looks like this:
Geographical benchmark*(demographic score + diabetes, angina, and dementia adjustments) = monthly capitation rate
$901.77*(.689+.287+.370+.131)=$1,331.91
This amount represents dollars paid to the health plan per month to provide medical services to Ruby Mae.
One more important caveat. The RAF score is retrospective, which means insurance claims received in 2019 set the payment rate for reimbursement received in 2020. Each January 1, the RAF calculation is set back to zero, essentially implying that if the insurance company does not receive a claim for Ruby Mae’s diabetes, angina, or dementia, then those conditions no longer exist. The additional RAF score represented by those chronic conditions will be absent in the payment calculation during the next payment year.
Furthermore, these conditions must be reported on a face-to-face visit claim with a qualified practitioner. Reports from nurses, medical assistants, or lab- or radiology-related visits would not count toward risk adjustment consideration.[4]
Risk adjustment is gaining speed in the insurance industry. This methodology is now the funding mechanism for individual and small group plans offered through the Affordable Care Act. Many state Medicaid plans have also adopted a modified version of risk adjustment to administer their own funding pools.
So, what could possibly go wrong?
Let us start first with the health plans’ stake in ensuring that chronic diagnostic (ICD-10–coded) conditions are captured year over year. If accurate ICD-10 codes are missing from claim submissions, the health plan will take a direct hit to revenue the following year, often with losses in the millions or hundreds of millions of dollars. In spite of the lack of revenue coming in, health plans are still obligated to pay claims, cover prescriptions, and pay doctors and hospitals through relative value unit and diagnosis-related group reimbursement. Consequently, health plans can easily find themselves in a negative revenue situation.
There can be several reasons for diagnosis code losses along the course of a year, such as:
-
Inattention to diagnostic coding, resulting in poor ICD-10 code capture. Practitioners receive payment on the current procedural terminology code reported. There is no financial incentive in accurately coding ICD-10 codes. These codes validate medical necessity but carry no significant impact to a provider and are often overlooked or missed completely.
-
Nonspecific coding that does not validate the higher level of care required. Think “difficulty breathing” versus “chronic obstructive pulmonary disease,” or “renal insufficiency” versus “chronic kidney disease stage 3.” Each one of these lower specificity conditions directly affects revenue received for each patient if the lower specificity code is selected.
-
Underreporting of codes on claim submissions. For example, if a practitioner treats five separate conditions such as hypertension, diabetes, chronic kidney disease, chronic asthma, and osteoarthritis during one visit but only reports hypertension on the claim, the remaining conditions are lost from risk adjustment consideration.
-
Truncated claims sent through clearinghouses that remove all diagnosis codes beyond four, thus wiping away valid ICD-10 codes from risk capture.
-
Poor documentation practices from providers that preclude higher-acuity code assignment. In this case, the physician may prescribe Furosemide to treat congestive heart disease but never actually includes an assessment or plan for the congestive heart failure. During chart review, there is no condition documented to validate the reporting of the code.
-
Long-standing conditions displayed on problem lists without subsequent attention or plans (e.g., respiratory dependence, below-the-knee amputations, hemiparesis, or chronic hepatitis).
Understandably, health plans have taken proactive approaches to solving the missing diagnosis codes problem over the years. These include various approaches, such as:
-
ICD-10 diagnostic coding education that includes guidelines for most chronic conditions coding. This can include pocket guides, cheat sheets, or specific conditions such as diabetes or congestive heart failure.
-
Retrospective coding audits to capture conditions that were documented in the previous year’s chart notes but never made it on a claim submitted to the health plan.
-
In-home risk adjustment–trained nurse practitioners and physician visits to ensure chronic conditions capture and documentation is complete.
-
Supplemental claims files sent from the provider groups and clinics that identify all diagnosis codes that may not have made it to a claim.
-
Risk-sharing contracts with providers to encourage active and robust ICD-10 coding.
-
Submission of diagnostic codes from patient assessment forms, yearly physicals, or annual wellness visits.
However, what happens when these proactive approaches cross the line from ethical practices to “gaming” the system—driving up the RAF score for patients that may not be as sick as the reported ICD-10 codes indicate? Remember, we have millions or hundreds of millions of dollars at stake here.
Consider the answers to the following list of questions:
-
How closely is CMS going to monitor whether Ruby Mae has full-scale diabetes or only elevated blood sugar? What are the guidelines developed by the health plan? Do those guidelines encourage overcoding or upcoding?
-
If a health plan performs retrospective reviews to capture missing diagnosis codes, is it looking both ways? Is it adding supported diagnosis codes and deleting diagnosis codes when support is not found in the medical record documentation? This “looking both ways” is now showing up on work plans at agencies such as the U.S. Department of Justice and Office of Inspector General.
-
Was that vendor in-home visit necessary for Ruby Mae, and did it provide continuity of care for her medical needs? Alternatively, was it simply a means to capture missing codes with no further follow-up or care?
-
Do the supplemental claims files submitted from the provider groups include labs and radiology claims that are not eligible for risk adjustment? Is the health plan removing those claims prior to submission to CMS? Are the ICD-10 codes included audited to ensure there is supporting documentation or treatment in the medical record to validate these extra codes?
-
Do the risk-sharing contracts reward clinics/providers for generating higher patient RAF scores rather than accurate scores? Is there an auditing program within your organization to ensure accuracy of the codes submitted on your claims? Furthermore, is there feedback to the providers when problem-coding issues arise?
-
Do the codes submitted from a yearly visit include all the documentation necessary to validate the capture of codes submitted? Is there reliable and consistent coding feedback given to the treating physician?
CMS has instituted yearly risk adjustment data validation (RADV) audits to address these concerns.[5] These audits take a subset of medical records that represent claims submitted by a Medicare Advantage–participating health plan and review them for accuracy. If the records do not validate the conditions reported, then CMS will assign an error rate to that subsample of members and then extrapolate against the remainder of the population patient panel. This error rate correlates to a “clawback” for all monies received to that health plan.[6]
RADV audits occur yearly for CMS but do not affect all plans every year. However, expect to see these audits ramp up in coming months and years as CMS targets unethical risk adjustment practices. Furthermore, the Office of Inspector General has also included RADV audits in its work plans for 2021.[7]
If this is not enough of a concern, then consider the risk of qui tam or whistleblower complaints leveraged by your own employees or a participating provider to CMS for any questionable risk adjustment practices. These complaints fall under the False Claims Act, and one visit to the U.S. Department of Justice news release web page exemplifies what could come next. As an added bonus, you will see your health plan’s name splattered across the news and internet with corresponding descriptions of “fraud, abuse, and ongoing investigations.” To further illustrate my point, I invite you to take a moment and search for some recent examples of risk adjustment whistleblower lawsuits.[8]
Summary
Medicare Advantage insurance benefits our senior population by offering better coverage and programs not available through traditional Medicare. All payers have a vested interest in ensuring the program is around for years to come. It is up to each of us to ensure all risk adjustment practices are compliant and centered on doing the right things for our health plans, communities, and, most importantly, our members.
Takeaways
-
Medicare Advantage is a highly popular and effective alternative to traditional Medicare.
-
Accurate documentation and coding are critical to Medicare Advantage success.
-
Proactive solutions, such as physician education, retrospective coding audits, prospective outreach, and alternative submissions, can help ensure a compliant risk adjustment program.
-
Compliance safeguards must be in place to ensure ethical practices do not drift into unethical upcoding/gaming of Medicare Advantage programs.
-
The Centers for Medicare & Medicaid Services will audit payers through the Risk Adjustment Data Validation Audits.