Goodbye “Outpatient CDI,” Hello “HCC CDI”

EDITOR’S NOTE: The following is the third and final installment in a three-part series by Dr. Remer on outpatient clinical documentation integrity.

In the first two parts of this series, we talked about risk adjustment in general, the shift to population health management, and how quality metrics and reimbursement are linked to risk adjustment models. In this installment, we will explore the hierarchical condition categories (HCCs) model, and I will explain why I don’t like the term “outpatient CDI (clinical documentation improvement).”

The HCC model organizes diseases and conditions into groupings that have similar characteristics and longitudinal resource utilization. The “hierarchy” comes into play when there are related conditions; only the most severe manifestation enters into the calculation, so lesser or duplicative conditions are nullified.

There are multiple HCC models. The Centers for Medicare & Medicaid Services (CMS) uses the CMS-HCC model, whereas commercial insurers may utilize the HHS-HCC model (U.S. Department of Health and Human Services). This is analogous to the MS-DRG in relation to the APR-DRG; the model needs to be tweaked to account for a younger, healthier population as compared to the Medicare and disabled population.

The CMS-HCC is a prospective model. This means that diagnoses ascribed to a patient this year predict next year’s consumption of resources. No conditions carry over; after Jan. 1, all diagnoses need to be reattributed to a patient for them to enter into the calculation. Alternatively, the HHS-HCC model is concurrent, because younger folks are more likely to have episodic healthcare, through which resources are utilized and then the condition resolves, without the prediction of future resource consumption. The perfect example of such a scenario is pregnancy.

There are a few other differences between the two models, but let’s focus on CMS-HCC, because it was the premier methodology, used in Medicare Advantage plans, and it is illustrative. I am going to simplify the system a bit because the exact scores are not crucial here; the concepts are key (that sounds better than “the actual calculation is too complex for me to be able to wrap my mind around.”)

The patient starts with a baseline score given according to demographics, presence of disability, and eligibility for Medicaid. The risk adjustment may differ if the patient is from the community or is institutionalized. The CMS-HCC model used today was created in 2014, so don’t go looking for an updated one from 2017.

Diagnoses deemed HCC-worthy are high-cost conditions, common conditions with high frequency of care, and complex, long-term disease processes. Sixty-one percent of seniors have at least one HCC, as compared to only 23 percent of patients younger than 65 years old. Diagnoses are organized into disease groupings, similar to major diagnostic categories (MDCs) in the DRG system. These are then broadly combined with clinically related disease processes with similar costs/resource consumption into condition categories (CCs) – couldn’t they come up with a different acronym? The hierarchies are then imposed.

Remember that the information in this article is current as of 2017. HCC RAFs adjust periodically, so take the numbers in this article with a grain of salt. You will need to check the CMS website for the actual current numbers: https://www.cms.gov/medicare/health-plans/medicareadvtgspecratestats/risk-adjustors.html.

Each HCC has a risk adjustment factor (RAF) coefficient that is a numerical representation of the weighting of the expected consumption of resources. For instance, HCC 23 is titled “Other Significant Endocrine and Metabolic Disorders;” the community RAF is 0.228 and the institutional RAF is 0.337. Encompassed in this HCC are conditions such as idiopathic hypoparathyroidism, primary hyperparathyroidism, diabetes insipidus, and glycogen storage diseases. If you were unfortunate enough to have two of these endocrinopathies, you only get the RAF once – no double dipping!

Let me explain hierarchies with an example. There are five HCCs in the cancer hierarchy grouping (RAFs are the community coefficients):

If a patient currently has breast cancer that has metastasized, she theoretically could have both HCC 12 and HCC 8, but HCC 12 would be nullified by the higher hierarchical condition, and you would only calculate the 2.625 figure into the risk adjustment score (RAS).

Over the course of the year, all the conditions a patient has are documented and captured. It is advantageous to be specific, as this may land a diagnosis into the appropriate higher HCC within a hierarchy. For example, gram-negative pneumonia is in HCC 114 [RAF 0.599], pneumococcal pneumonia is found in HCC 115 [RAF 0.221], and J18.9, pneumonia, unspecified is not found in a HCC at all.

The final factor is the disease interaction adjustment. When certain diagnoses are present in concert in a patient, there is an additional contribution. For instance, if a patient has both congestive heart failure (CHF) and chronic obstructive pulmonary disease (COPD), there is an additional 0.190 accrued, whereas having cancer and a documented diagnosis falling under “immune disorders” yields a RAF of 0.893.

The RAFs get added up to produce the total RAS. This is adjusted by some yearly determined normalization factor intended to establish the average RAS as 1.000. The patient RAS is analogous to the relative weight of a DRG. The population or institutional average RAS is analogous to the case mix index (CMI). The average RAS can be monitored and trended, and may indicate the diagnostic precision of a given provider or group.

Forty-two percent of HCCs are also comorbid conditions or complications; 16 percent of HCCs are MCCs, and 42 percent of HCCs are neither CCs nor MCCs. Inpatient CDI has been predominantly focused on ensuring capture of CCs and MCCs, with a secondary goal of attempting to optimize severity of illness (SOI) and risk of mortality (ROM). Forty-two percent of diagnoses that pertain to the HCC risk adjustment are not on the inpatient CDI specialist’s radar, though. Although the HCC model has been utilized in some quality metrics all along, now providers are going to be attuned to the fact that their finances are directly dependent on documentation and the codes derived.

HCC diagnoses can be gleaned from inpatient principal and secondary diagnoses, hospital outpatient clinics, physician offices, other clinically trained professionals (like podiatrists and psychologists), and both technical and professional claims. It is optimal to document conditions multiple times, because if a claim is denied for some reason, a diagnosis only found on that rejected claim may be eliminated from the patient’s set.

Surprisingly, payors are invested in getting the risk adjustment correct as well. They may be getting premiums and subsidies that are dependent on accurate RAS, and they want to anticipate expected expenditures correctly. The alignment of the payor and the provider is a welcome change.

And so, finally, we get to the crux of the matter. Why do I not fancy the term “outpatient CDI?” Because diagnoses are not exclusively found in the outpatient realm, and inpatient CDI specialists can help with the objective of capturing acute and chronic disease processes. It may very well be strategic for an Accountable Care Organization (ACO) to hire outpatient CDI specialists to staff their office settings as well, ensuring that all significant diagnoses are captured, especially on yearly preventative encounters. However, educating experienced inpatient CDI specialists to consider CCs, MCCs, and HCCs may benefit everyone involved.

Perhaps the term “HCC CDI” would be more precise?