From 3M Health Information Systems
E/M coding: Can we agree to disagree?
In the past, I’ve written several blogs about the complexity of E/M coding and how to find your way through. The first step is to establish the place of service, then the visit type and finally the level of care. These three categories are not always easily determined and level of care is the most complex of all, including three sections of guidelines and multiple subsections to calculate that level based on the rules associated with the visit type.
Everyone agrees that E/M coding is the most complex coding category in CPT, that E/M coding makes up a large percentage of medical services, and that these services are usually coded by providers, not coders. Additionally, no one seems alarmed that after years of scrutiny–auditing, education, EMR-suggested codes that are often ignored in favor of the provider-selected codes—that expert E/M coders still disagree with each other around 30-40 percent of the time. Those of us that know E/M coding inside and out understand why that is, (contradictory rules, vague/multiple guideline, different payer interpretations), yet we are tasked with telling providers that they have selected the “wrong” code. How credible are we when we can’t even agree with each other about E/M coding?
In a recent article in Healthcare Business Monthly published by AAPC (February 2017), the author Thomas Field, CPC, CEMC, wonders about the benefit of E/M coding. He makes a well-supported argument that the administrative burden of E/M coding is too high when compared to the eventual reimbursement and I have to agree with him. When I think about how much effort goes into getting the “right” code selected for any given E/M services with an average of $95 reimbursed per service, I am stunned. As a coder, my opinion has always been that physicians should not spend their time coding. I know many, maybe even most, providers would agree with me. I hold this opinion, not because providers are bad at coding their E/M services, but rather because auditing and education doesn’t work for most providers.
Back in the day, when I was doing compliance work, I knew who would “fail” their audit. I knew because that provider had “failed” in the past. When I talk with practice administrators, HIM directors, compliance officers at any, literally any clinic, they also know who their “risky” providers are in terms of coding. This tells me that my experience is a common one. Audit, educate, feedback, repeat: This will only work for a small group of providers. The remainder are going to go about their business, providing the best medical care they can and pay little attention to the coding of their services.
Yet there are changes in reimbursement for professional services on the horizon. Some of these changes are in relation to HCCs, which are based on ICD codes rather than CPT. And we have a shiny new set–well new to us–of ICD-10 codes to work with. We have over 71,000 ICD-10 CM codes, compared to five, or sometimes only three levels of care. Our difficulty with the E/M codes is that there are so few of them with which to code the entire human condition’s worth of services. But this isn’t a call-out for more E/M codes, or a revisit of the directions on how to level an E/M service. There has been far too much of that and in my mind, it was wasted time. No, we need to agree to disagree on E/M coding, not be so punitive when a provider gets the code “wrong” and concentrate on getting the most specific ICD-10 code possible assigned to the visit. Some specialists will have a limited number of ICD-10 codes to work with, but others will not. Will they take the time to find the most specific ICD-10 code possible, or will they select a catch-all unspecified code, which might not point to an HCC?
Only you know how well your providers are selecting their ICD-10 codes. And now it matters, because reimbursement will be effected by the specificity of the codes submitted on many claims.
Rebecca Caux-Harry, CPC, is the CodeRyte product specialist for cardiology with 3M Health Information Systems.