By Emilie J DiChristina, MBA for PracticeFirst
Over the last several weeks I have had the opportunity to work with medical records in the capacity of:
- Clinical coding auditing;
- Clinical Documentation Improvement; and
From the experiences above, you may be expecting a rant about “those darn providers”, or “cutting and pasting in EMRs” and so forth, and in truth, all of those comments may be valid for those involved in coding at any time. The issue that has struck me the most regarding coding over the last several weeks is the potential for variance among coders, among providers, among revenue cycle managers when it comes to the codes themselves.
A simple would be the coding for a patient getting a vaccine. Some coders include a “nurse visit or 99211”, some include a low level provider visit, while others code just the vaccine. In large part this is practice preference, however when coders go from one provider to another, these variances may be set-in.
Another example are those coders who are clinically trained as they really dig into the chart for specifics to diagnoses, or even worse – read into the record “what they think the provider is trying to say”. I’m as guilty of this as anyone.
It is safe to say that coding variances are more the rule than the exception.
In 2008 at the Perspectives in Health Information Management, Computer Assisted Coding Conference Proceedings; M. Elliott Familant, PhD, et al discussed the need for a Gold Standard in medical coding. Their specific comments on coding variability are as follows:
“Multiple studies have shown that medical coders will frequently differ in the number and type of codes they apply to medical charts. This is true for ICD-9-CM and ICD-10 coding and CPT (Current Procedural Terminology) codes, including E&M coding. Coder variability is a significant challenge in developing any standard of “correctness” for medical coding.”
Also, in the AAFP publication, Fam Pract Manag. 2005 Jun, Cindy Hughes CPC and Trevor Stone discussed similar concerns. In their article, they note:
“Coding is more an art of interpretation than an exact science. Nowhere is this truer than with E/M services. In 2000, the Journal of Family Practice published a study in which four faculty physicians, six resident physicians and six professional coders audited 1,069 charts.1 The results indicated that professional coders agreed with the codes assigned by the physicians far more often than did the physician or resident auditors. Further, when auditors disagreed with a code selection, the documentation justified a higher level of service four times more often than it suggested a lower level of service.
Another recent study involving a review of E/M codes assigned to 125 visits by pediatricians found that 44 percent of the visits were undercoded and 1 percent were overcoded.2 Before correction of the undercoding and overcoding, the code distribution showed a bell curve with the predominance of claims at the midlevel (99213). After correction based on the proportion of error found by the independent review, a more equal distribution of midlevel (99213) and upper level (99214) codes was demonstrated.
In a third study, 600 family physicians were sent surveys asking them to assign codes for six different clinical scenarios.3 The 205 physician responses agreed with expert CPT E/M code assignment in 52 percent of the scenarios for established patient visits. Undercoding was the most common error, identified in 33 percent of the cases. Again, the bell curve showed a more equal distribution of codes 99213 and 99214 after correct coding was applied.
Yet another study found that family physicians undercode their services at least as often as they overcode (22 percent undercoded, and 20 percent overcoded).
These studies indicate that coding of E/M services is not a clear-cut process. Physicians have difficulty choosing correct codes for these services based on the guidelines available, and even trained coders often disagree.”
So what does that mean for coders, auditors, revenue and compliance staff, and the providers overall? It means there needs to be clear cut practices, internal auditing, and a coding staff which knows your specialty, and gets routine feedback from a lead coder or coding manager.
It is impossible to eliminate every variance and there is an understanding that coding, particularly in the area of E&M may have a +/- swing of 1 code (which in some cases in considered unavoidable, but still undesirable), however the use of only highly trained, educable coders, well versed in your specific specialty is a goo