Home  |   Subscribe  |   Resources  |   Reprints  |   Writers' Guidelines

September 2013

Risk Adjustment Creates Challenges
By Brian Boyce
For The Record
Vol. 25 No. 12 P. 6

In a world where providers consistently have attempted to meet criteria for levels of care for evaluation and management codes, now there is a new documentation need that further illustrates the complexity of care providers deliver to each patient. Risk adjustment has become known in health care circles largely because of its effect on diagnosis coding for Medicare Advantage plans. Still, there are many health care professionals who aren’t sure exactly what it is or how it works.

Risk adjustment models were in use long before Medicare began looking at implementing them for its Advantage plans; one of the largest groups using a risk adjustment model is Medicaid plans. Risk adjustment is a process of collecting all diagnosis codes from patient charts and using these illnesses (along with their comorbidities and complications) to determine the ICD codes, which drive risk. Diagnosis codes are a good starting point for analyzing known conditions and speculating on potential risks.

Medicare Advantage plans focus on these diagnoses to illustrate the need for higher reimbursement rates for patients who have more serious conditions or problems to manage. Utilizing the risk adjustment model likely is one of the most groundbreaking ideas health care has seen in years. Instead of focusing on levels of care, which can vary from visit to visit, health plans can concentrate on what actually is wrong (by ICD code) with each patient. While it may be difficult to predict with much accuracy the next stroke or heart attack victim, each current diagnosis can be successfully identified and enable more preparation from a care management and financial forecasting perspective for patients in general. 

While Medicare Advantage plans are collecting all known diagnoses to ensure accurate funding for their patient panel care needs, Medicaid and even commercial plans now are interested in risk adjustment models and how they can assist in predicting future patient needs and plan for potential complications. This level of precise diagnosis information is paramount to understanding each patient.

The Medicare-adopted Hierarchical Condition Categories (HCC) model assigns a numeric or HCC value to the diagnosis codes, each of which carries a risk adjustment factor similar to the relative value unit found in CPT coding.  In the Medicaid chronic illness and disability payment system (some commercial plans have adopted the Medicaid model as is or with some adjustments), there are far more diagnosis codes identified. While the codes carry a numeric value, they also are rated as being high, medium, or low risk.

All in the Family
Both forms of risk adjustment use a family and hierarchy for diagnoses. For example, in the diabetes family, diabetes with a serious complication such as kidney disease is considered to be a high-risk diagnosis as opposed to a lower-risk diagnosis of diabetes with neuropathy. Taking it a step further, diabetes with no known complication is low risk.

As such, documentation becomes a major influence on risk adjustment. For years, diagnosis codes have been misused across many levels of care. In fact, one of the largest drivers of diagnosis code selection continues to be which problem matches the CPT codes being billed to demonstrate medical necessity. Coding guidelines state that providers should document all diagnoses that are a part of the medical decision-making process for each visit. For example, if a patient presents with possible strep pharyngitis and also has diabetic neuropathy, it would be appropriate to choose the former as the primary diagnosis code. However, diabetic neuropathy also should be included because it surely was taken into consideration when treating the first condition. 

There has been a general decline in diagnosis coding, which sometimes leads to selecting the first code found, even when it may not be the best choice.

Relationship Problems
The proper documentation of complications and comorbidities also plays an important role in risk adjustment. Many providers still do not realize that coding guidelines largely prohibit coders from assuming cause-and-effect relationships. If these are not clearly documented in the medical record, they are lost in the translation.

Diabetes is a particular challenge.  Many providers have memorized the basic code for diabetes, yet that primary unspecified code should be used only for diabetics with no complications. Typically, there are complications (which are weighted higher in risk adjustment models), but they frequently go undocumented. This not only skews the patient’s medical history but also hampers data collection on a larger scale. Many providers do not enter all of the current diagnoses on encounter forms or superbills, which means they subsequently don’t appear in electronic systems where they can be tracked and credited appropriately.

Another example of problematic diagnosis codes is hypertension which, like diabetes, has potential cause-and-effect conditions that may lead to a different code. Many providers also do not understand the special rules for coding stroke.  Once a patient has been discharged from inpatient treatment, it is no longer appropriate to use the stroke code. Instead, a history of stroke code should be used.

The Offshore Option
The combination of a heavy coding need and a perceived coder shortage has led some practices to opt for offshore coding services.

The economy in India is such that physicians and nurses can become certified coders and increase their salary by coding records for US clients.  The average chart review fee for record review in the United States can range between $15 and $40, depending on the type of review or audit. When completed overseas, the cost per chart can drop to between $5 and $10. Some companies use the disparity to their advantage. For example, they will have the initial work done overseas and quality review performed by US coders.

There are some concerns, however, because many offshore coding companies lack the strong training necessary to accurately code what is documented in the record. Also, while English often is a strong second language in these countries, there still is a great deal of miscommunication when it comes to interpreting medical record documentation, especially when American medical abbreviations and medical slang terms are used.

On the security front, business associate agreements may be in place, but privacy concerns persist. Because HIPAA itself can’t be enforced overseas, the actual business contract and its language become crucial to anyone using offshore solutions. There may be future guidance from the US government on how such companies may or may not be used.

Beyond the overseas coding challenges, there also are onshore challenges to consider. The medical coding profession is rapidly growing, and because of increasing interest in the profession, there are many community and career colleges offering coding courses and certificates. Unfortunately, many of these programs take advantage of unsuspecting new students. Many require prerequisite courses such as medical terminology or anatomy and physiology before students are allowed to take a coding course. While these prerequisites may seem like a good idea, often they are only there to collect more federal education assistance dollars.

Students should be cautious of any education program not affiliated with AAPC or AHIMA, organizations that often can do a better job of preparing students for national exams. It’s also probably a good idea for vendors to favor coders who have been certified by either AAPC or AHIMA.

Life in Risk Adjustment
Certified coders are not necessarily knowledgeable on every topic. For example, risk adjustment is relatively new, so there’s a good chance training will be necessary. While many companies are hesitant to hire CPC-A (certified professional coder—apprentice) candidates, it’s important to note that these newcomers often have yet to develop bad habits and are eager to learn. With proper training, new coders with less experience actually can be an asset.

It is not unusual for two certified coders to disagree on code selections regardless of their areas of expertise. The majority of disputes typically are related to the manner in which different providers specifically document in the medical record. The way in which a diagnosis is written or described ultimately can change code selection, such as use of the term “chronic” in diagnosis coding.

Coders are accustomed to submitting diagnosis codes on claims for the purposes of reimbursement validation for services rendered. However, that is not the case in risk adjustment models. Instead, codes are submitted to properly calculate each patient’s risk score, which affects the financial reserves that will go toward future treatments related to these existing diagnoses.

These diagnoses are collected not for the purpose of submitting a claim but rather to include them in a supplemental file that will chronicle the patient’s condition throughout the year. The diagnosis codes are converted to an HCC or Chronic Illness Disability Payment System value (risk adjustment factor), resulting in an adjusted risk score composed of diagnosis values that are used to budget for annual patient care.

Some have argued that risk adjustment is inflating previously unreported diagnoses. However, the concept is requiring providers finally to pay close attention to how diagnoses are documented. The subsequent surge in more specific diagnosis codes will serve to better illustrate the nation’s overall health.

— Brian Boyce is an AAPC-approved professional medical coding curriculum instructor, an AHIMA-approved ICD-10-CM trainer, and head of the consulting firm ionHealthcare.