Risk adjustment is a newer payment methodology that is designed to allow for a proactive approach to managing healthcare. It does not seek to audit against current claims or encounters. It is a yearly collection of diagnosis codes to accurately build a total RAF (Risk Adjustment Factor) score so that we may become more precise in healthcare dollar spending while also becoming more proactive in disease management and patient care. While it collects diagnosis codes, it is completely separate from other medical billing and claims.

The most used and well-known payment method is Fee-For Service (FFS) which assigns a dollar value for certain procedure codes and services being performed. Under FFS, the only use of a diagnosis code is to show the medical necessity of the procedure or service that was billed.  Diagnosis codes can be used for other things too, for example identifying patient groups such as cancer survivors, disease management program opportunities, identifying commonly found diagnoses based on geographical locations, socioeconomic groups, ethnic groups, and more.

Some of those diagnoses are flagged for better management opportunities, but many of the diagnoses meant to be included in risk modeling may not always meet the criteria to appear on a claim to support what was billed for that claim but are appropriate supplemental diagnoses from which to build a clinical profile. This profile is what supplies funding to help pay for those FFS claims and are meant to prepare for the patients’ needs in the future. 

Risk adjustment models are really a modified capitation approach. Frankly they are brilliant if we use them properly. No longer chasing after bills for past treatments, we have an opportunity to create an individualized approach to medicine. Risk adjustment models do not care about any of the procedure codes or services provided. Their focus is on the accurate collection of all current diagnoses for each patient for each year as specifically and completely as possible.

Note that each diagnosis is counted only once for the whole year, and there are no points for assigning the same diagnosis or another one in the same diagnostic family just because they appeared more often on medical claims. Hierarchical Condition Categories (HCC’s) are only applied to certain diagnoses that are specifically chosen chronic conditions which are costly.

Each HCC carries a risk adjustment factor (RAF). RAFs are fractional numbers, much like those of an RVU (relative value unit) for procedure codes. They are added together for a total RAF for each patient. This RAF will then be applied against the patient’s PMPM rate for next upcoming year in prospective models like CMS Medicare Advantage. Instead of paying a flat rate of say 800.00 PMPM, we adjust that PMPM rate according to how sick each patient truly is, as documented by the diagnoses they carry each year in anticipation of total estimated future health costs and needs.

To establish an expected standardized average cost of care, many brilliant people came up with algorithms that are applied to the model. The data is calibrated to hold the average in line and in comparison, with the Traditional Medicare FFS model. This is one of the yearly calibrations that take place. The average RAF is set to 1.0. This means that anyone with total diagnoses and conditions, once all HCCs are accounted for, who have fewer medical conditions, the RAF will be lower than 1.0, thereby reducing the PMPM rate. Alternatively, for patients with higher (greater than 1.0) RAF scores, then the PMPM will be increased to account for the management and treatment of those many conditions. 

Risk adjustment is identifying all known diagnoses (not just the ones we see every 3-6 months) and calculating a unique RAF score customized to each person. There are many conditions for which a limited number of primary care providers have time to see in an already overfilled appointment book. There are many ongoing active diagnoses that will only be found in a list of some sort within the record.

We manage frequent visits for common chronic conditions like diabetes, COPD, CHF, CKD, hypertension, hypercholesterolemia, and heck oftentimes several of these diagnoses in multiple patients. There are many other common chronic conditions which are actively managed and usually followed every 3-6 months depending on individual patient circumstances, but there are also many chronic conditions found in risk adjustment models that do not require a yearly visit to manage or address anything at all. They should be reported so that we are ready for their manifestations and expected effects on each person’s health. 

Risk adjustment models currently require us to re-document all diagnoses each year.  Even those permanent lifelong conditions for which there is no treatment. Even genetic disorders for which there is no cure. I have always thought of this requirement as a built-in quality check to make sure that the RAF profile is accurate each year, by checking those diagnoses previously noted, and identifying any newly assigned diagnoses, or new complications or manifestations by newly reported ICD code specificity. If we neglect to report them and only report them when we monitor, assess, evaluate, or treat them, then patients themselves suffer unnecessarily.

There are more than 68 diagnosis codes that are commonly only documented in an ongoing problem list or PMH list of some sort. When they are not collected for risk adjustment purposes a chain reaction occurs. First, there will then be no financial forecasting done for the uncollected chronic condition, and now we will have unplanned expenses. This will come as a surprise when the patient comes in next year or maybe 3 years from now with a manifestation or complication of some uncollected chronic condition for which we were not prepared financially nor did we expend any disease management effort for the invisible diagnosis, but here we are with expense of a manifestation or complication, and the patient has deteriorated due to our negligence.  We could have minimized or even possibly prevented the manifestation or complication that would have been both a better health outcome as well as saved healthcare dollars had we just known all of them.

Identifying all current comorbidities is the first step necessary for risk adjustment to succeed. The second step is what we do with this data. How we develop tracking programs and disease management protocols that are designed to keep complications and manifestations at bay or possibly even circumventing them altogether as we allow for a proactive view of healthcare delivery using risk adjustment modeling where there are fewer surprises. We should also be thinking ahead and outside of the box for how we can better manage healthcare delivery. Disease management programs can begin to pivot to more proactive measures to mitigate manifestations and complications and as we progress in technology, I am sure we will develop new opportunities to improve, revise or even reverse conditions. 

“When a patient has a cardiac event, the quickness of an AED (automatic external defibrillator) can make a huge difference in the patient’s outcome. When a shock is needed, the sooner it can be delivered, the better. This is why we now see AED’s in so many public places and AED training is now a part of CPR. We learned to be proactive. We can do lots more of this in the interest of healthcare delivery through risk adjustment modeling.” 

Brian Boyce