Next presentation will be by Nikhil Sharma, Surgical CPT Coding Discrepancies and Analysis of Surgeons and Employed Coders, an argument I think we frequently have. Good afternoon. My name is Nikhil Sharma. I'm from the Hospital University of Pennsylvania, and I will be presenting on Surgical CPT Coding Discrepancies. I just want to take a minute to thank the AANS for giving us the opportunity to present our work. We have no financial disclosures. All right. Providers and billing technicians use current procedural terminology codes, CPT codes, to specify what treatment a patient receives for a given admission, as well as the cost associated for that treatment. At the Hospital of the University of Pennsylvania, the attending surgeon will first enter his or her codes subsequent to surgical intervention. Next, the employed coder who is blind to the surgeon-derived codes will review the surgeon's dictated op notes and then ascribe the codes that they see fit. If there are no disparities between the two lists, the codes are submitted to third-party payers. However, if discrepancies do arise, an email discussion ensues prior to submission, which is not delayed. At our institution, individual surgeons have recently reported an increase in disparity rate with employed coders. In theory, these two lists of CPT codes should match for an individual patient per admission, with generally less than 5% code change rate. So we hypothesize that the change in CPT coding is actually occurring at a higher rate. And this is important because if there are discrepancies between the two lists, they could have monetary consequences as differences would change the calculated work RVU, as well as the total RVU for each patient. They can also attribute to incorrect hours of work performed by the surgeon. All right, so to explore any differences, we retrospectively analyzed a heterogeneous cohort of 300, of 500 patients over three months. And inclusion criteria was all electively scheduled cases with preoperative HMP and precoding. We excluded any trauma cases, any patients under 18 years of age, and any emergency surgeries. So the two lists of CPT codes were categorized into four categories based on the changes that the coder made. They were addition of a distinct code, addition of a repeat code, which would be like adding a spinal level. Deletion of a distinct code, and a deletion of a repeat code. Any CPT codes that matched on both lists were excluded as we were primarily concerned with discrepancies. And to determine and quantify the impact of change or discrepancy between surgeons and billing technicians, the RVU differences were analyzed. We also surveyed coders and surgeons on cases where disparities occurred to assess the level of agreement on submitted codes. All right, so of our 500 patients, 150 of them had identical CPT code lists. And so these were excluded since we were interested in the disparities. However, 350 cases had some changes of some sort. When we looked at the individual categories of changes, we saw that the addition of a distinct code occurred in about 85% of cases with an average of 2.8 code additions per case. Repeat additions occurred in about 4% of cases with an average of 4.6 additions per case. Distinct deletions occurred in about 33% of cases with 1.9 deletions per case. And repeat deletions occurred in about 15% of cases with 2.1 deletions per case. When we looked at the overall 350 patients, we saw that 22% of surgeon-derived codes were deleted and 11% of repeat codes were deleted. Further, we saw that coders added a net of 600 CPT codes. This net addition of 600 CPT was comprised of 2.1 distinct code additions and .37 repeat code additions. When calculating the RVU implications, we saw a range between negative 11,000 RVUs and positive 20,000 RVUs over the course of three months. Now, this could mean that some months had a net negative shift while others had a net positive shift. But what's important is that there is a drastic and significant amount of disparity between the coders and the surgeons. One more interesting thing that was found was that when analyzing the individual cases between the surgeons and coders, the surveys that we administered, there seems to be multiple instances of continued disagreement about how a case should be coded. And this primarily happened with more complex cases. So the graph on the left is just meant to show the net difference that the coders made, while the graph on the right is the raw number of changes per category that we looked at. All right, so our hypothesis was correct in that there is a high disparity rate between surgeons and coders. The disparity analysis demonstrated that there is continued disagreement and that this is something that needs to be addressed. So the steps that we're taking at our institution to mitigate any disparities include enhanced coder-surgeon communications to enhance consensus. We require attendance of coders and surgeons at AANS, CNS coding education courses. We want to develop a pedagogically sound paradigm for coders and surgeons coding education. And for more complex cases with continued disagreement with high disparity rates, we should seek outside educational consulting agreements to assess these complex cases and develop a consensus for future coding. So I want to thank the AANS again and thank you very much for your time. Thank you.

Surgical CPT Coding Discrepancies

CPT codes

treatments and costs for patients

attending surgeon

employed coder