Breaking News
Home / Medicine & the Law / The ‘median coding’ phenomenon

The ‘median coding’ phenomenon

By Michael G. Calahan, PA, MBA.

Scenario A – Undercoding: Dr. Smith evaluates and treats his longstanding patient, Mrs. X, for diabetes mellitus and hypertension. Knowing her case well but evaluating some emerging new problems with her current medication regimen, he spends more time than usual with her reviewing and discussing salient points of her history, extending her review of systems, and modifying her plan of treatment with several new medications. He initially encircles 99214 on the superbill; Dr. Smith then changes his mind fearing 99214 “looks too high” for a well established patient, scratches it out and encircles 99213 instead.

Scenario B – Overcoding: Dr. Jones briefly sees and evaluates a patient during a second office visit. The patient reports total abatement of his flu-like symptoms and feels completely well. Dr. Jones quickly examines the patient in light of this news, performing a cursory auscultation check over a few key areas of his lungs, and adds this history and examination data to the record; she does not re-review the remainder of the history components (e.g., review of systems, etc.) and simply tells the patient to finish up his originally prescribed medications and return as needed if his symptoms resume. Dr. Jones checks off 99213 as a “typical follow up office visit” and does not consider the required components of this level of service.

It comes as no surprise to those of us who have been auditing evaluation and management (E/M) services by reviewing clinical records that some physicians have fallen into habitual coding patterns without regard to the actual required components of the E/M levels. In consulting we term this audit process as “leveling,” meaning assessing the three required components of each E/M service to ascertain at which level or code the particular service should be reported. The three required components of most E/M services are: history, physical examination, and medical decision making.

Each of those components have subsets of requirements to be met depending on the type of E/M service (e.g., office visit, inpatient hospital visit, consultation, skilled nursing facility visit, etc.), as well as the level within that service range (e.g., 99211-99215). With the focus on E/M services by the Office of the Inspector General (OIG) beginning in the early to mid 1990’s, followed by the confusing issuance and application of the 1997 E/M Guidelines and the subsequent choice of utilizing either the 1995 or 1997 guideline sets, many physicians simply opted to brush the confusion aside and play it safe by coding their E/M services in a “middle-of-the-road” or median manner. The rationale, though not always applicable, is that many services will be undercoded (“I cannot be fined for that!”) and that most services billed would not be overcoded. Further, even if an E/M service or two is overcoded, then “I’ve undercoded so many services at this point Medicare owes me money.” It is a miscalculation.

Unfortunately, each and every E/M service is adjudged individually, so the practice of habitual median coding resulting in many undercoded E/M services does not provide any type of leverage when having to answer for the other scenarios that might have been overcoded. The Centers for Medicare and Medicaid Services (CMS) via its Medicare carriers does request individual health records for the purpose of leveling or scoring those services for coding/billing accuracy, and does track physicians on an individual basis by the claims each submits for payment. The claim is the mechanism for tracking, and the clinical records are the documentation required to substantiate each E/M service. By now, this is all quite well known. What many physicians are not aware of is that CMS publishes it tracked information in a several formats (available on-line at Again no surprises: CPT code 99213, Office or Other Outpatient Visit requiring expanded problem focused history and physical exam and a medical decision making of low complexity, used so often as the median code for established patient office visits is reported far more than any other E/M code.

According to CMS data reported by Medicare participating physicians in 2005 (i.e., MedPar Data), code 99213 far exceeded other codes and represented nearly $5.7 million in allowed charges. Was each and every of the 110-plus million incidences of 99213 reported in 2005 appropriate? Individual cases would have to be audited to accurately answer that question. But many more cases might be audited in the future due to another coding phenomenon still emerging as the familiarity with modifier -25 becomes more firmly established.

Modifier -25, which represents “Significant, Separately Identifiable E/M Service by the Same Physician on the Same Day of the Procedure or Other Services,” allows an E/M service to be paid (instead of being denied) in cases when another service is performed during the same patient encounter, such as removing a skin lesion or providing a therapeutic injection. The over- and mis-reporting of modifier -25 was illustrated by the November 2005 OIG report “Use of Modifier 25.” Based on a review of only 450 claims from year 2002, the report suggests the misuse of modifier -25 in 35 percent of the cases audited.

Using CMS’ numbers for 2002, out of 29 million claims costing the Medicare program approximately $1.96 billion, this misuse accounts for about $538 million in claims reported to Medicare and subsequently paid incorrectly. Though not stated as such, by sheer volume one must suppose with the median coding phenomenon involving CPT code 99213 that many of the inappropriate -25 modifier cases also involved CPT code 99213. The OIG has charged CMS, via its Medicare carriers, to be more vigilant in their review of claims with modifier -25, and naturally this will involve more auditing and review of cases reported with CPT code 99213.

Ironically, consultants do find that many cases have been undercoded when reviewing an E/M case reported as 99213, therefore lost revenue is a very real issue. Education provided on-the-spot to providers is an essential function of the consultant, and during this educational interaction consultants find an almost palpable reluctance on the part of providers, physicians in particular as opposed to physician assistants or nurse practitioners, to code their E/M services at higher levels. There is a concern that coding at higher levels (e.g., 99205, 99215), even when appropriate, will cause the CMS or OIG spotlight to fall on that physician. It is a false sense of security however, and staying under the radar by coding at middle-of-the-road levels will not keep a physician safe from CMS scrutiny.

While it is true that CMS tracks provider patterns and pulls from those data aberrant coding practices, subsequently auditing those providers who fall outside the normal bell curve, it is also true that median coding can emerge as aberrant coding. For example, a recent audit of a specialty practice’s E/M services revealed all new patients coded as CPT code 99203 and all established visits coded as 99213. There was no deviation from this practice over the timeframe audited, which encompassed about six months’ worth of services. No aberrant bell curve is needed in this case; this coding pattern graphs as a straight line. Instead of staying under the radar this practice cast its coding patterns in neon.

This was not an isolated case. Consultants see this pattern, more or less, all of the time. The solution ultimately comes back to provider education and the implementation of “best practices.” Best practices for proper E/M coding is carried out on an individual basis for each and every patient encounter by thoroughly considering all required components for each E/M service, and is not done out of habit or out of a premeditated coding pattern to avoid federal scrutiny.

In fact, it is a myth that certain providers can avoid CMS medical record reviews for E/M services or the use of modifier -25 by median coding. Medicare carrier claims review initiatives are random in many cases, so even providers who fall squarely inside of the bell curve norm have the chance of being audited. It therefore makes no sense to continually perform median coding when appropriate coding, once learned and applied, is far less worrisome and generally more profitable.

Michael G. Calahan, PA, MBA, is Director, Healthcare Operations & Compliance Services, Parente Randolph, LLC.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.