Risk adjustment (RA) is the most groundbreaking initiative involving reimbursement models since their inception and will likely become the standard in the future across all payers; therefore I applaud you for seeking out knowledge on risk adjustment.
Risk adjustment has been around for years. You are likely aware of being personally risk adjusted already – insurers have been using this methodology to determine our life insurance and car insurance for decades. But what is risk adjustment in healthcare? Risk adjustment is a modified version of the traditional capitation system which is fairly limited in its ability to differentiate payment for the healthy person who rarely sees a physician, and the person with multiple co-morbid conditions who requires complex management.
Plans are serviced through health insurers, which are paid by CMS based on a risk adjustment score instead of a DRG. The health insurer plan then pays the provider based on the RA score.
Using an actuarial tool to plug in a person’s current conditions and utilizing data collected over decades about those conditions, a forecasting perspective can be derived not only for future financial implications, but more importantly, to predict future patient care management needs and plan for potential complications. It essentially levels the playing field and allows each individual’s health to be reflected as it truly is.
Every one of us carries a level of risk. The healthiest of us – those without any chronic health conditions, are at the lower end of the risk scale. With the addition of certain health conditions, the risk increases. Other factors that impinge on risk are how impactful the condition is on the member’s health status, and the compounded nature of particular conditions. For example, a diabetic is at a certain level of risk, but a diabetic with nephropathy is at an increased risk.
The CMS-HCC Classification System, used for Medicare Advantage beneficiaries, begins with the 70,000+ ICD-10-CM codes and funnels into 805 diagnostic groups, which are further specified to 189 Condition Categories. It is at this point that hierarchies are imposed among related conditions, hence hierarchical condition categories. For payment year 2017, current version 22, of those 189 HCCs, there are 79 active and designated HCCs.
Each HCC has an associated value called the relative factor, similar to the relative weights that are used in DRG classification systems. Using the individual’s health conditions, a risk adjustment score is determined using a combination of demographic information (age and gender) along with disease information. The score is highest for the patients who are either sicker, or whose conditions utilize the most resources, as determined by the combination of several factors. There are also additional factors such as disabled status and Medicaid eligibility that can be included in the calculation. The accumulation or sum of the demographic and disease factors to predict costs for each individual is called the risk score.
While there isn’t necessarily a brief reference list, there is a crosswalk that maps every ICD-10-CM code, though some find it easier to ask 2 questions that assist in identifying potential risk adjustors:
- Which conditions are currently affecting the patient’s health status?
- Which conditions have the potential to affect the patient’s health status in the future?
These diagnosis codes for the patient are collected by the Medicare Advantage plan from hospital inpatient, hospital outpatient and physician office claims over a calendar year. The supported diagnoses are considered differently according to how they relate to each other.
HCCs reflect hierarchies among related disease categories. For example, if a person has diabetes with a chronic complication and develops an acute complication, there will not be risk scores for both HCC 18 (Diabetes with chronic complications) and HCC 17 (Diabetes with acute complications). The hierarchy of HCC 17 will be the only one of the two applied.
For unrelated diseases however, HCCs accumulate; therefore, in this way, the model is also additive. For example, a male with heart disease, stroke, and cancer has at least three separate HCCs coded, and his predicted cost will reflect payment increments for each of the three problems.
In addition to the additive terms in the model, the CMS-HCC model also incorporates some interaction terms for conditions where the costs are more than additive.
For example, it has been determined that the presence of both diabetes and CHF leads to higher expected costs than for DM and CHF alone. Therefore, the model includes a set of two-way interactions between certain pairs of disease groups, those which together have clinical validity and most strongly predict higher additional costs.
These disease interactions allow for higher risk scores when these particular disease pairs are present. There are interactions for disease & disease and interactions for disability & disease.
Because a single individual may be coded for none, one, or more than one HCC, the CMS-HCC model can individually price tens of thousands of distinct clinical profiles. The model’s structure thus provides, and predicts from, a detailed comprehensive clinical profile for each individual.
Acuity and Specificity
Because ICD-10-CM codes are used in risk adjustment, the documentation of acuity and specificity can be significant. These are some examples of the increased specificity needs that are important to include in the documentation for risk adjustment:
- Hepatitis, acute hepatitis, unspecified viral hepatitis, alcoholic hepatitis > no HCC
- Acute hepatitis with hepatic failure > HCC 27
- Alcoholic cirrhosis > HCC 28
- Alcoholic hepatic failure without coma > HCC 28
- Alcoholic hepatic failure with coma > HCC 27
- Bronchitis not specified as acute or chronic > no HCC
- Chronic bronchitis > HCC 111
- Renal failure:
- Renal failure > no HCC
- Acute renal failure > HCC 135
- Obesity > no HCC
- Morbid obesity > HCC 22
- Arrhythmia > no HCC
- Specified – most > HCC 96 (at fib, at flutter, VT, PAT, SSS)
- Vfib, Vflutter > HCC 84
- Unspecified, Stage 1, 2, 3 > no HCC
- Stage 4 > HCC 137
- Stage 5, ESRD > HCC 136
Commonly Missed Conditions
- Diabetes and manifestations
- Secondary cancers
- Drug/alcohol dependence
- Amputation status
- Ostomy status
- Asymptomatic HIV infection status
- Renal dialysis status
- Ventilator dependence
CMS HCC Key Points
- The CMS-HCC model is a prospective model: diagnoses in one year are used to predict costs in the following year.
- CMS-HCC follows a calendar year.
- The slate is wiped clean every January 1st; therefore, all ongoing conditions must be addressed and documented again each calendar year.
- The encounter must be face-to-face.
- The CMS-HCC risk adjustment model is based on ICD-10-CM codes only, not PCS, CPT or HCPCS codes.
- Code assignment must be in accordance with the International Classification of Diseases (ICD), Clinical Modification Guidelines for Coding and Reporting.
- There is no sequencing in CMS-HCC.
- A record can have more than one HCC.
For those with a strong coding and clinical knowledge base, as well as well versed in clinical indicators, risk adjustment is not far from the responsibilities of the front-line CDI professional working in the inpatient environment. Strengthening the outpatient coding guidelines will increase your success in risk adjustment.
My hope is this article begins your self-learning journey toward success in risk adjustment.
Karen Newhouser, RN, BSN, CCM, CCDS, CCS, CDIP
Director of Education
CMS. April 4, 2016. Announcement of calendar year (CY) 2017 Medicare Advantage capitation rates and Medicare Advantage and Part D payment policies and final call letter. Note: Table begins on page 78 of this document. Retrieved from https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Downloads/Announcement2017.pdf
CMS-HCC 2017 Model Relative Factors. Retrieved from https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Downloads/Announcement2017.pdf
CMS. July 23, 2013. Risk adjustment101 participant guide. 2013 National Technical Assistance. Retrieved from www.csscoperations.com
CMS Medicare Advantage 2017 Midyear Final ICD-10 Mappings. Retrieved from https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors-Items/Risk2017.html?DLPage=1&DLEntries=10&DLSort=0&DLSortDir=descending
CMS. Risk adjustment. Chapter 7, Medicare Managed Care Manual. Retrieved from https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/Downloads/mc86c07.pdf
Smith, D.M. December 2015. Risk Adjustment: Leveraging HCCs. SC ACDIS PPT. 3M. Retrieved from http://blogs.hcpro.com/acdis/wp-content/uploads/2015/12/12.4.15-HCCs.pdf
Smith, D. & Gordon-Moore, L., 2016. 3M Best practice recommendations for ensuring complete HCC capture. 3M. Webinar retrieved from http://players.brightcove.net/2635130879001/Byb30H47x_default/index.html?videoId=5254178172001