Published in the June 2012 issue of Today’s Hospitalist
MEDICINE FEATURES MANY FAMOUS TRIADS: Cushing’s, Hutchinson’s and Virchow’s, just to name a few. Most tend to be obscure, board-exam issues that you seldom encounter in daily practice. When’s the last time you saw keratitis, notched teeth and deafness on the same hospital ward, let alone in the same patient?
In contrast, we have an epidemic of Coder’s Triad, which is so common that every patient seems to suffer from some form of it these days. Like many conditions, it appears to be a consequence of modern living. In fact, medical historians are fairly certain that the first reported case occurred two decades ago.
Coder’s Triad consists of history, physical examination and medical decision-making. Although there are theoretically 64 possible “presentations” for any given patient, everything eventually collapses down into three levels of service for the common series of codes that hospitalists use (99221’99223 and 99231’99233).
History and physical exam don’t tend to cause coding and compliance problems. (See our May 2010 “Translating a diagnosis into RVUs” for a quick refresher.) Most hospitalists know how to count, and the organ systems and/or body areas reviewed are either documented (key word) or not. Medical decision-making, on the other hand, has added layers of complexity and may even admit some shades of grey. Dollars to donuts, this is where most coding problems occur, even if your documentation is otherwise OK.
Medicine is a thinking profession. Patients’ welfare “and, by extension, our salaries and the hospital’s bottom line “depends on the choices we make. Although there is clearly a lot of texture to most medical decisions, the Centers for Medicare and Medicaid Services (CMS) parses our work in that category into three broad components:
- the clinical scenario (the number of diagnoses or treatment options);
- the data we play with (the amount and complexity of information reviewed); and
- the risk of bad things happening (complications and/or morbidity or mortality).
Overall, you determine decision-making by mixing and matching among these categories, and the final decision-making level you choose must meet or exceed certain levels in two out of those three. This two-of-three rubric can help or hurt you, depending on the clinical scenario.
Take, for example, a somatically preoccupied patient with some acne, a hangnail, a few paper cuts, and unwanted hair in her nose and ears. While there are multiple diagnoses, you won’t score any points in the data and risk departments. Even a compulsive internist will have a hard time getting anything other than straightforward medical decision-making out of the encounter.
Conversely, one single bad diagnosis “MI, stroke or PE, for example “could potentially land a patient in the moderate-to-high complexity range because of all of the data to be considered and the significant chance of complications and risks.
You say low, I say high
On its face, the framework for determining the level of decision-making seems a bit squishy. Chicken Little, who sees impending catastrophe in benign circumstances, assesses risk differently than the rest of his partners. And the average hospitalist and, say, cardiac electrophysiologist probably have a slightly different take on a run of V-tach. To some degree, risk is in the eye of the beholder.
The CMS, not surprisingly, doesn’t really go for arbitrary and vague. Although it acknowledges that the “determination of risk is complex and not readily quantifiable,” its matryoshka-like system provides additional, nested guidance. (See “How do you determine risk?“). This still permits some wiggle room, but it reduces practice variation considerably.
The guidance makes it clear, for instance, that “the highest level of risk in any one category … determines the overall risk,” and that’s a fairly reasonable standard. Moreover, commonplace interventions like prescription drug management and IV fluids containing potassium are in the moderate-risk range. Parenteral controlled substances, about which we’re pretty cavalier, are at the highest end of the scale. For hospitalized patients, it’s hard to imagine many low-risk situations.
A few words of caution: Never collect data without medical necessity. One of the most common errors coders see is documentation that “old records were reviewed,” without any indication of why you needed to review old records or any summary of what information you were able to glean.
And selecting an E/M service level must be based on both clinical relevance and medical necessity, not on the volume of documentation you can produce. That said, hospitalists probably leave a lot of money on the table. Take the patient receiving IV fluids containing potassium. This intervention alone could (and probably should) prompt you to tailor your history and physical exam accordingly.
You could reasonably evaluate the patient for dehydration, fluid overload, and complications such as local infiltration, phlebitis and infection. Depending on your approach, your history and physical exam could meet the requirements for billing a higher level of service in terms of subsequent visits (99231-99233).
That’s not piling on to take advantage of the system. Instead, that’s working backwards from decision-making to history and exam “and working smart, putting how we think and make decisions in charge.
David Frenz, MD, is a hospitalist for HealthEast Care System in St. Paul, Minn., and is board certified in both family medicine and addiction medicine. You can learn more about him and his work at www.davidfrenz.com. Sue A. Lewis, RN, CPC, PCS, is a compliance consultant with HealthEast Care System.