As Medicare-managed care health plans (Medicare Advantage (MA) plans) expand—especially in the past five years—providers are more regularly affected by “risk adjustment.” When done properly, the risk adjustment model has great potential to enhance the quality of patient care. Unfortunately, risk adjustment is also susceptible to fraud by the proverbial “bad apple.”

Risk Adjustment Basics

Risk adjustment is a modified version of the traditional capitation system. Under traditional capitation, a managed care organization or provider group is paid a fixed amount per member per month (PMPM) to pay for all services the member requires during that period. Traditional capitation sets the PMPM rate based on demographic factors such as the member’s age, gender, and geographic location.

Risk adjustment enhances traditional capitation by adding payments for patients who are being actively treated for certain diseases and conditions known to be expensive to treat. Risk adjustment classifies patient sickness using hierarchical condition categories (HCCs), which are groups of related diagnosis codes. HCCs are similar to diagnosis-related groups (DRGs) or ambulatory payment classifications (APCs) used for hospital reimbursement, but they are based on diagnosis codes rather than procedure codes. Individual patients may fall into multiple HCCs. For each additional HCC, the MA plan is paid an extra amount.

Unlike traditional managed care, where there is a strong financial incentive to seek out only healthy members, the risk adjustment model rewards managed care organizations and provider groups to care for sick members, as well. They can see a significant financial reward if they actively manage those sick members to reduce health care costs.

Risk Adjustment Fraud: “Upcoding” DxCodes

The current design of the risk adjustment system largely relies on MA plans to police themselves. MA plans are responsible for determining which diagnosis codes its members were treated for in the prior year, using a combination of traditional claims data and other medical documentation (such as the patients’ medical charts). The plan then submits the diagnosis codes to the Centers for Medicare & Medicaid Services (CMS) to get the increased risk adjustment capitation payments.

Unethical MA plans and vendors take advantage of the system’s structure to essentially “upcode” the diagnoses they submit to CMS. They do this by submitting a risk adjustment claim to CMS for a diagnosis the member either did not have or was not treated for in the year in question. In such cases, “risk adjustment” may be offered as an explanation for why patient medical records should be changed or “supplemented” (sometimes a year or more after the patient was treated). Or the MA plan, or its vendor, may suggest that a provider call a patient in for an office visit so certain diagnosis codes can be “captured” for “risk adjustment purposes” (regardless of whether the patient actually needed any medical treatment).

CMS rules are clear that a risk adjustment claim may be submitted only if the diagnosis meets ICD-9-CM standards and there is documentation in the medical record that the member was treated face-to-face by a qualified provider in the year questioned.

Common schemes used to upcode diagnoses for risk adjustment purposes include the following:

Coding from Problem Lists: CMS rules explicitly state that a “problem list” may be used only to code a diagnosis if it is “comprehensive and show[s] evaluation and treatment for each condition that relates to an ICD-9-CM code on the date of service.” It is improper to submit risk adjustment claims for diagnoses that are merely mentioned in the member’s problem list if the diagnoses were not treated or considered by the provider during that visit.

Improper Linkages: The risk adjustment system pays MA plans a higher capitation rate when certain conditions are “linked.” For example, a patient may have both diabetes and nephropathy. CMS will pay the MA plan more if the diabetes caused the nephropathy because diabetes with renal complications is generally significantly more severe than diabetes without complications. Diabetes without complications, which falls within HCC 19, has an average value of $1,500 per year. In contrast, diabetes with renal manifestations, which falls within HCC 15, is valued at over $4,500 per year.

For an MA plan to submit a linked diagnosis code to CMS, the provider must document the linkage between the two conditions in the medical record. It is improper for an MA plan or vendor to assume the two conditions are linked.

Coding from Test Results or Prescriptions: CMS prohibits the submission of risk adjustment claims based solely on laboratory or radiology test results, drug prescriptions associated with particular diagnoses, or durable medical equipment (DME) services. Nonetheless, certain MA plans and vendors include diagnosis codes in their risk adjustment submissions even though they appear only on those invalid sources of documentation.

Chronic Conditions: While it is true that some conditions (such as Parkinson’s) never go away, this does not mean that the diagnoses can be submitted to CMS every year. Risk adjustment rules explain that a condition may only be submitted for reimbursement if it is actively treated (or affects other treatment) in the year in question. It is not enough that the patient was diagnosed or treated for the condition at some point in the past.

Targeted Coding: Some organizations pressure coders to focus on identifying high-value diagnoses, rather than coding just what is in the medical record. Some common high-value targets include:

• Cachexia/Malnutrition (HCC 21) – value of $7,800 per year
• Old myocardial infarction (MI) (HCC 83) – $2,200 per year
• Diabetes with complications (HCC 15) – $4,600 per year
• Major depression (HCC 55) – $3,200 per year

Know the Red Flags

If a coder involved in chart reviews or an audit related to risk adjustment sees any of these activities, there is a strong likelihood the coder is dealing with fraud. If someone tells a coder to use a diagnosis code that doesn’t meet ICD-9-CM standards and says it is OK because “risk adjustment coding is different than regular coding,” that is a major red flag indicating the health plan or vendor is engaged in fraud.

At its core, risk adjustment coding is “regular coding,” but stricter. Even where a diagnosis meets traditional ICD-9-CM standards, it may not be submitted for risk adjustment purposes unless the diagnosis is: (1) documented by the provider in the medical record as having been treated or as affecting the patient’s treatment; (2) made during a face-to-face encounter; (3) submitted to the MA plan from a qualified provider type; and (4) made during the specified calendar year.

Risk Adjustment Fraud and the False Claims Act

At a May 31, 2012 MA compliance conference, federal prosecutor Robert Trusiak noted that MA fraud—in particular risk adjustment fraud—is a “hot button issue” for the Department of Justice (DOJ). Trusiak further noted that MA plans face potential liability under the federal False Claims Act (FCA) for false risk adjustment claims, even when the upcoding or other fraud was perpetrated by a vendor on the plan’s behalf.

The FCA says any person who submits a false or fraudulent claim to the United States or causes someone else to submit a false or fraudulent claim may be liable for three times the amount of the false claim, plus an additional penalty of up to $11,000 for each false claim. To encourage whistleblowers to report fraud, the FCA contains a qui tam provision awarding whistleblowers 15-30 percent of what the government recovers as a result of whistleblower lawsuits they file against individuals and entities committing fraud.

The government has already begun enforcement against unscrupulous MA plans attempting to game the risk adjustment system. In United States v. Janke, the government sued an MA plan under the FCA for submitting upcoded (or non-existent) diagnosis codes for risk adjustment payments. The DOJ settled with the MA plan and its owners for $22.6 million in November 2010.

Be Cautious and Speak Up

As Trusiak cautions, the FCA targets not only the person or organization submitting a false claim, but also anyone who “causes the submission” of a false claim. This means that MA plans are not the only ones who face potential liability under the FCA for false or fraudulent risk adjustment claims. Hospitals or physician groups could be liable, as well, if they submit false information about their MA patients’ diagnoses to MA plans and that false information is used to submit a false risk-adjustment claim to CMS.

To avoid this risk, you should ask to review rules used by vendors when those vendors are identifying “new” diagnoses. Don’t hesitate to speak up if the standards being used by an outside reviewer don’t line up with the established CMS coding rules your organizations are using. Providers should insist on reviewing any code submissions made for their patients—especially when an MA plan or vendor has reviewed the providers’ medical record and identified new diagnoses—to ensure the patient actually had that particular diagnosis and was treated for it during the visit. Coders, administrators, and providers can all take steps to prevent or stop risk-adjustment fraud.

Mary A. Inman, JD, and Timothy P. McCormack, JD, are partners at Phillips & Cohen LLP, a law firm representing whistleblowers (www.phillipsandcohen.com). Whistleblower cases brought by the firm involving Medicare and Medicaid fraud, and other types of fraud against the government, have returned more than $8.5 billion in civil settlements and related criminal fines to federal, state, and local governments.

Proper hierarchal condition category (HCC) classification depends on a plan’s ability to obtain accurate diagnostic HCC information and report that information accurately to the Centers for Medicare & Medicaid Services (CMS).

If a plan focuses solely on disease management to decrease costs (neglecting to develop an effective HCC strategy), it runs the risk of losing money due to under-reported HCC codes. Although the plan may still save $150-$250 per member, it will be deficient if it does not factor HCC coding into its business model and work aggressively on a two-pronged approach that incorporates both prospective and retrospective HCC capture.

Prong No. 1: Retrospective

A plan generally relies on algorithms (risk adjustment software) to search for unreported diagnosis codes via chart reviews. A plan’s coding staff, or a third-party vendor contracted by the plan, extracts large numbers of charts from network physician offices to capture chronic disease processes. After review, any previously unreported codes are submitted to CMS.

Prong No. 2: Prospective

Plans provide ongoing education to assist physicians in the process of developing a complete and accurate member profile that resonates with all current ICD-9-CM codes identified at each encounter. Taking a prospective approach increases a plan’s ability to capture more accurate data.

Providers must report all diagnoses that affect the patient’s evaluation, care, and treatment, including:

• Nature of the presenting problem
• All chronic conditions (such as atrial fibrillation, congestive heart failure (CHF), chronic kidney disease (CKD), rheumatoid arthritis, diabetes with manifestations, chronic obstruction pulmonary disease (COPD), all active cancers)
• History on any relevant past conditions
• V codes (factors that influence health/status codes)
• E codes (external causes of injury and poisoning)

HCC scores on individual members determine CMS reimbursement to the plan. Diagnosis and demographic information should be captured at each face-to-face encounter to obtain a health-based measure of that member’s future medical needs.

Knowledge = Recovered Reimbursement

Consider the following:

• More than 50 percent of a plan’s revenue comes from captured HCC codes.
• More than 30 percent of HCC codes do not pass the CMS validation process, due to lack of supporting documentation in the medical record.
• Providers do not report greater than 40 percent of active chronic conditions.

With those disturbing statistics, it is imperative that a plan employs certified coders who have a thorough understanding of CMS’ HCC methodology and HCC coding process to ensure capture of all documented chronic conditions that risk adjust to HCCs. Coders must also be able to identify documentation deficiencies and review with network providers for improvement.

HCC coding processes include:

1. Assessments, plans, all active chronic conditions, and diagnosis codes documented in charts annually.
2. Coding precision and specificity: Coders have the ability to conduct prospective chart reviews to capture missed chronic conditions that have been documented, but not submitted, by the provider or group.
3. The provider’s ability to submit at least eight diagnosis codes to maximize HCC reporting to plans (CMS has accepted eight diagnosis codes since 2007). You may claim 99080 Special reports such as insurance forms, more than the information conveyed in the usual medical communications or standard reporting form for providers who submit to Medicare Advantage plans to report additional diagnosis codes for chronic conditions. Some providers who use an electronic health record (EHR) may not have the ability within the EHR to submit more than four diagnosis codes to a plan. Code 99080 has no relative value units (RVUs), and may be used as an adjunct to the evaluation and management (E/M) code to capture additional diagnosis codes without skewing the provider’s accounts receivable (A/R) report.
4. The plan sends to risk adjustment processing system (RAPS) diagnosis codes that are converted to HCC codes.
5. CMS factors the plan’s risk adjustment.

This process allows plans and providers to deliver better benefits and care. For example, at the plan level:

• CMS reimburses health plans on a risk-adjusted basis.
• The sicker a member is expected to be, the more CMS pays a plan.
• Diagnoses reported in one year affect payments for the next year.
• Increased reimbursement from CMS (due to better and accurate reporting from providers) allows the plan to provide richer benefits to members for the following year, and allows for bonuses and better reimbursement to providers for fee-for-service (FFS)/capitation models.

Providers also are better able to:

• Completely and accurately assess member’s health status.
• Monitor and document all active diagnoses, past illnesses, and status conditions.
• Monitor readmissions to hospitals.
• Review medication.
• Identify potential new problems early.
• Reinforce self-care and prevention strategies.

Plans that implement a two-pronged approach (prospective AND retrospective) to capture HCC codes will see increased revenue and cost containment through better disease management by including a defined HCC coding initiative. A plan that combines both approaches can potentially increase revenue anywhere from $1,500-$2,500 per member.

The 411 on Third-party Vendors

If a plan chooses to work with a third-party vendor to aid in the retrospective aspect of HCC capture, it should have a checklist clearly defining the plan’s expectations. A vendor’s ability to successfully conduct a majority of the retrospective coding initiatives (the first prong) is imperative, as it allows the plan to focus on prospective coding initiatives (the second prong).

When a plan has a targeted approach to HCC capture, it can better identify high-risk members and channel them into an appropriate disease management program. At the end of the day, when a plan is successful at HCC capture, it creates a win-win outcome for the plan, the providers, and ultimately the members who are served.

Here is a checklist of what a plan should look for in a third-party vendor to assist in retrospective reviews:

• Vendor has established relationships in physician network
• Current number of clients: Can the vendor handle your plan’s volume on time to scan appointments and minimize rescheduling?
• Ability to generate pursuits and set scan appointments
• Ability to identify what the extractions should or should not include (health care effectiveness data and information set (HEDIS) measures, special needs plan (SNP) forms, progress notes)
• Flexible chart retrieval services based on the specific needs of the plan
• Number of scan techs on staff: Does the geographic range and staff support the provider network area?
• Security and Health Insurance Portability and Accountability Act (HIPAA) compliance—equipment types (for example, flash drives, portable scanners, etc.): Do they bring paper if records have to be printed, so as not to use the provider’s resources?
• Diverse staff to meet different market needs: Excellent provider and plan relationship skills
• Ability to view the electronic images of all medical records
• Number of certified coders on staff (in-house and remote)
• Ability to generate accurate coding reports based on scans to minimize duplications and errors
• Ability to code each record using online magnetic resonance angiography (MRA) reporting, capture for diagnosis, or HCC codes
• Ability to accurately identify areas in the record that support HCC findings and risk adjustment data validation (RADV)
• Ability to identify provider deficiencies in documentation and coding, and report to the plan on results
• Annotate the electronically coded record with notes and report generation to assist the plan in targeting deficient providers
• Year-to-date, month-to-date, and real-time (within the past 30 days) report generation to identify low RAF score providers and providers whose HCC reporting is low in comparison to panel size
• Ability to identify members who have not had any HCC codes reported from a provider panel
• HEDIS reporting abilities to assist providers and plan to obtain four- and five-star ratings
• Pharmacy utilization and facility tracking
• Ability to identify members who have not been seen and are new to the provider panel within the past six months
• Turn-around time (TAT) from time of scan to coding, with report generation to the plan
• A vendor also should provide the plan with a monthly accounting that identifies errors and generates corrective actions from all pursuits. The report should contain at the very least the following:
• A list of members charts scanned from provider or group
• A list of charts that were coded from provider or group
• A list of charts that weren’t coded from provider or group with logic to pursue with provider/group
• By member, a list of captured HCC or prescription drug hierarchical condition category (RxHCC) codes that can be submitted
• By member, a list of dropped HCC or RxHCC codes that need to be addressed with the provider or group
• By member, a list of reduced HCC or RxHCC codes that need to be addressed with the provider or group
• By member, a list of new HCC or RxHCC codes

Understand HCC Methodology

HCC payment rationale was developed to mirror the individual health risk profile (HRP) of Medicare Advantage members, and uses ICD-9-CM information as the primary indicator to determine a member’s health status. Thousands of ICD-9-CM codes map to less than 100 HCCs, which are what ultimately drive risk adjustment factor (RAF) scores and per member per month (PMPM) premiums paid to a Medicare Advantage plan.

Holly J. Cassano has worked in practice management, coding, auditing, teaching, and consulting for multiple specialties for the past 16 years. She served two terms as an AAPC local chapter officer, maintains an online column for Advance for Health Information Professionals, writes for Justcoding.com, and is the host blogger for: Coding Notes for Consumer Media Network (CMN) www.medicalbillingandcoding.org/blog/welcome-to-my-new-blog/. This past April, she presented at the Third Annual HCC Best Practices for Proactive Medical Management from Generalities to Interventions to Outcomes for Physician Groups and Health Plans, in Jacksonville, via Opal Events. She works for Preferred Care Partners as a CDI specialist, based out of The Villages, Fla and is the founder of ACCUCODE Consulting, LLC (hjcpmg@yahoo.com). You can reach her at  holly@medicalbillingandcoding.org or follow her on Twitter @HollyCassano.