Updated on: November 28, 2016

More Doesn’t Always Mean Better

By
Original story posted on: March 25, 2016

The nation’s healthcare industry is expected to experience an influx of new codes when regular updates to ICD-10 resume as of the start of October 2016.

But will more codes translate to a better way of doing things?

That’s the focus of “Getting Specific: New ICD Codes – Will They Make a Difference?” – a newly published white paper by Joseph C. Nichols, principal of Health Data Consulting.

“A robust set of codes that define the key parameters of the patient health state provides the opportunity for valuable information to improve the health of the population and improve outcomes of care. This opportunity will not be realized, however, if these codes are not used,” Nichols wrote. “The critical path to data specificity does not lie in the number of codes, but rather in the consistent and appropriate use of those codes by providers as part of the delivery and documentation of the care they provide.”

It was a topic broached, Nichols noted, during a recent special edition of ICD-10monitor’s Talk-Ten-Tuesdays – specifically, a broadcast that covered the latest National Committee on Vital and Health Statistics (NCVHS) ICD-10 steering committee meeting  about proposed additions and changes to the current ICD-10-CM code set.

The broadcast featured discussion about the ability of these new codes to provide much-needed detail about certain conditions, plus the opportunity to clarify the nature of each condition and to identify potential differences in risk, severity, and complexity for different patients with similar types of conditions. 

A few of the areas addressed included the following, Nichols noted:

  • Standard categorization for reporting the level of blindness and low vision;
  • Detailed anatomical localization of breast lumps;
  • Classification of myocardial infarctions by type;
  • Additional clarification of the severity of non-pressure ulcers; and
  • Addition of a specific code for the Zika virus.

“These and many other ICD-10 codes provide the opportunity for a much richer data environment to understand the key parameters of the patient condition that make dramatic differences in assessment of risk, severity, cost, injury causes and patterns, patient safety, disease trends, quality and outcomes, policy impacts, and any other use of population data,” Nichols wrote in his white paper. “Opportunity does not, however, imply implementation.” 

“The advantage of this increased level of detail is not realized if we don’t see this detail reflected in clinical documentation and the codes that capture these important clinical concepts. Unfortunately, our track record for capturing and coding key clinical concepts is not good,” Nichols continued. “Unlike many other industries that put a premium on the quality, completeness, and accuracy of transactional data, healthcare has often considered this data merely an administrative burden required to get paid.”


Joseph C. Nichols, MD

Nichols also referenced a recent ICD10monitor article covering the issue of leveraging ICD-10 rather than simply mirroring ICD-9; the article focused on the suggestion that unless there is a significant change from old data capture habits, the additional opportunity for better data would never be realized.

Health Data Consulting recently performed an analysis of three years’ worth of payer data across all lines of business, the article noted, and based on this data, 5 percent of all ICD-9 codes were used for nearly 75 percent of all claim charges.

“Codes that would be considered ‘unspecified,’ ‘other,’ or ‘symptom and findings’ codes were used as the primary code for 54 percent of all outpatient (professional) claims. While there has been significant improvement in the ability of ICD-10 codes to capture more specific information about breast cancer, for example, historically we have coded breast neoplasms as ‘unspecified type’ or ‘unspecified site’ over 80 percent of the time,” Nichols wrote. “More detailed codes have always been available, but they are rarely used. There is little reason to believe that without some change in documentation and coding habits, that data will be any better in ICD-10.”

There are a number of factors that play into this less-than-optimal use of diagnostic codes, Nichols noted. Specifically:

  • Clinicians believe that these codes are simply for payment purposes and have no other value.
  • Most clinicians don’t see any incentive to code more specifically. In some instances, there is a disincentive, because vague codes may actually be more likely to pass edits than detailed codes.
  • Transactional data goes into a black box from the provider perspective, since very few providers have a robust ability to analyze their own claim data and compare their coding patterns for the same conditions to their colleagues.
  • There is often a lack of clinical consistency in the application of key parameters across similar clinical domains within the ICD-10 code set.
  • Some code search tools in EHRs (electronic health records) and billing systems change descriptions, so they don’t represent what the code really means to the clinician entering the code.
  • The use of code description terminology may not be consistent with terminology familiar to the clinician.
  • Combination codes are used with widely varying levels of granularity, so the clinician may be uncertain about the number of codes to use to represent various aspects of the patient condition.
  • Coding tools frequently search for terms or words rather than medical concepts.
  • When searching for key terms, the clinician may run into a wall and settle for a non-specific code.

“The changes occurring in a data driven, value-based purchasing environment are creating new incentives for providers to take the accuracy of data more seriously. If we assume, however, that there are strong incentives in place for more detailed data, there is still the challenge of finding the right code in the nearly 70,000 codes currently in play,” Nichols wrote.

The solutions, he added, are many, luckily enough.

Since the structure and terminology of codes varies greatly across different clinical domains, for example, a specialty-focused approach is critical. Nichols suggested that clinicians should be educated in the variation in use of codes that combine a number of parameters into a single code (combination codes); where terminology may be inconsistent with common use, or even inconsistent within the ICD-10 language; how to drill down on code sets without hitting a wall; what additional parameters are available and need to be documented to further define a condition; and what’s missing in certain clinical areas from the current code set.

Furthermore, Nichols said:

  • Concept-based tools to assist the clinician in finding the right codes based on an ontology of clinical concepts are evolving. These tools build clinical knowledge into the code mapping so they are intelligent enough to include codes such as “Colle’s fracture” and “Smith’s fracture” into the results of a code search that’s looking for codes related to “extra-articular” “fractures” of the “distal” “radius.”
  • Clinicians should have access to meaningful analysis of their claim coding patterns and how their patterns compare to some expected benchmarks.
  • A thorough analysis of the workflow of clinical data capture and system input should be done to identify process improvement opportunities.
  • Clinician incentives should be created to reward accurate, complete, and specific definitions of the patient health conditions they are managing.

“A vision for understanding current incentives and challenges for provider documentation and coding must be a primary focus. Based on this understanding, there must be a commitment to actively and openly identify opportunities for improvement that do not increase the administrative burden on the clinician,” Nichols concluded. “There must also be a commitment to act on those opportunities. As I have gone around the country, I repeatedly hear great ideas for improving both care delivery and data capture processes from many sources within the same organization. I also repeatedly hear from both the clinical and administrative side the frustration that ‘no one is listening.’”

Disclaimer: Every reasonable effort was made to ensure the accuracy of this information at the time it was published. However, due to the nature of industry changes over time we cannot guarantee its validity after the year it was published.
Mark Spivey

Mark Spivey is a national correspondent for ICDmonitor.com who has been writing on numerous topics facing the nation’s healthcare system (and federal oversight of it) for five years. 

Related Stories

  • Things Your Mother Never Told You About HCC: Version 23
    The 2019 CMS risk adjustment model is version 23. The Centers for Medicare & Medicaid Services (CMS) released, in April, the latest update to the CMS-hierarchical condition category (HCC) Risk Adjustment Model (V23).  It applies to payment year 2019.  As…
  • Random Thoughts about ICD-11
    New classification system noted for granularity. Several of my colleagues recently attended an ICD-11 presentation by Kathy Giannangelo[i] at the American Health Information Management Association (AHIMA) Convention & Exhibit. Kathy has been in the trenches with ICD-11’s development for some…
  • Understanding Presumptive Linkage for Code Titles “With” or “In”
    Sharing insights on assumptive coding  When I was a physician advisor, I used to offer a diabetic Charcot joint as an example of why we must be explicit with linkage. Years ago, if a provider listed diabetes mellitus and a…