By Joe Bormel, M.D., MPH, for Healthcare Informatics
The 83rd annual convention of AHIMA is getting underway as I write this, and with it HIM is at the center of the biggest regulatory reforms in most of our careers, ARRA/MU and ICD-10. At this event, new and evolving approaches to coder’s workflow in the MU era, provenance standards, documentation methodologies incorporating natural language processing (NLP), Computer-Assisted-Coding (CAC), Quality Reporting, and related services like Clinical Documentation Improvement (CDI) will be discussed. In many ways, it all gets down to billing and clinical codes. So I thought this would be a good time to take a look at the challenges and issues.
Quality healthcare through quality information?
Physician, policy advocate and certified coder, Dr. Mike Stearns was recently describing the distinction between billing codes and clinical codes. Billing codes are intended to get you into the neighborhood of the specific condition, explains Stearns. The clinical codes are needed for care, quality reporting, identifying appropriate order sets, and creating adequate clarity for medication management and reconciliation.
It will come as no surprise to readers of Healthcare Informatics or this blog that billing codes and clinical codes are fundamentally different. But the temptation to use them interchangeably can be overwhelming. When they don’t overlap, which is often, problems arise. No group of people understands this challenge better than the Health Information Management professionals who have been managing our medical records.
Billing and Clinical Coding
Let’s take a look at diagnosis billing codes, most commonly synonymous with ICD-9 and soon, ICD-10. The other billing codes are the procedure codes; there are several families of these, CPT-4 being the most recognizable set in the US. The simple, bottom line is that those codes are intended to describe the “neighborhood” of a clinical problem, not a specific, clinically accurate diagnosis. Your could think of a clinically accurate diagnosis as a “driveway within the general neighborhood of the diagnosis billing code.”
ICD-10 helps build on ICD-9 by creating more and smaller neighborhood definitions, and are often more anatomically and physiologically specific. That said, even in ICD-10, the notion of accuracy is still a bit schizophrenic in the lay, two-headed sense of the word. Accurate for billing is not the same as accurate for clinical care. When there are two or more possible billing codes, the one that a payer will reimburse will be used, even when another billing code might be a closer depiction of the clinically accurate situation.
One recent example of this comes from a real ENT physician who diagnosed jaw pain coming from the TMJ (the joint between the jaw bone and the skull bone, just beneath and in front of the ear). In order to get the clinical encounter “covered” or reimbursed, he had the give a billing diagnosis of pain coming from the ear. His clinical diagnosis was clearly something very different, pain coming from a boney joint. The ear pain was the neighborhood in this analogy; the TMJ syndrome pain was the specific clinical “driveway within that neighborhood.” But as you can see, that ear pain “driveway” was in a different neighborhood on a nearby street.
Those kinds of inaccuracies are a deliberate consequence of how we pay for and account for care, versus how we classify the clinical condition for diagnostic and therapeutic purposes. This “accuracy for different purposes” dichotomy is well understood and accepted as common by both HIM professionals and clinicians.
When clinicians select codes
When a patient presents to a clinician in a care delivery organization, there ends up being a clinician who declares the clinical diagnostic, as well as billing codes to describe that encounter. That declaration process in the age of EHRs and Meaningful Use is not simple. In addition to impacting reimbursement from payers as it always has, even in the paper world, the declaration has a broader impact. Specifically, quality measures and workflow are directly impacted; the former is required for MU attestation, the latter for clinical adoption.
The relationship of accuracy on Meaningful Use and Adoption
From an accuracy and simply a data perspective, the clinical codes now need to populate an Up-to-Date Problem List (required by MU in Stage One). From a broader MU perspective, these codes also take part in MU quality reporting. Hopefully, it’s obvious that clinical accuracy is far more important than billing accuracy for patient safety. That said, it can be expedient, albeit risky, to use billing codes, even when they are clinically not accurate. And making it easy, fast, and straight-forward for clinicians to even enter or pick an accurate clinical code is often hard or impossible.
Here’s where HIM innovations to be discussed at AHIMA get interesting. There has been increasing interest in using the clinician’s free-text narrative as the primary source of capturing the accurate clinical code, as well as the defensibly accurate and neighborhood-to-driveway accurate billing code. This often involves a combination of voice-recognition, NLP, CAC, and related services CDI, support and real-time compliance adjudication.
With any innovation, there are ultimately three sequential questions that need to be addressed. First, does the innovation actually work? For coding, the measure is coding accuracy. Second, is the innovation fast enough? In healthcare terms, what are the effects on the productivity of coders, abstracters, doctors, nurses, and health data specialists serving internal and external reporting needs? Third, is it polished and pretty enough? Obviously, if the first two questions aren’t properly addressed, polish is irrelevant. However, when we roll out technologies that work and are fast enough, but aren’t sufficiently polished, we often aren’t afforded the opportunity to go back and polish. Think about the consequences and please share your thoughts with me.