ICD-10: Increased Specificity Should Clarify, Not Confuse
By George Schwend and Brian Levy, MD
The sheer volume of diagnosis codes found in ICD-10-CM—roughly 68,000—often leads to the mistaken perception that ICD-10 will create more confusion in the healthcare industry than the current 13,000 ICD-9-CM codes. In reality, however, the specificity inherent in ICD-10 should clarify a great deal about care processes. The key is to ensure that provider documentation completely and accurately maps to codes with precise descriptors rather than those that are unspecified.
For most of the 30-plus years ICD-9 has been used in the United States, HIM professionals have tried to encourage providers to document in greater detail in the clinical record—with varying degrees of success. The problem, of course, is that providers—already overburdened with large caseloads—have little time to relearn what they need to write in the patient chart to ensure accurate coding.
Yet the only way to take full advantage of all the clinical and administrative promise of ICD-10 is to ensure the highest degree of specificity in both documentation and code selection. To benefit from ICD-10 specificity, providers must be offered tools and strategies that improve documentation with minimal effort and within current workflow. HIM professionals must then have a way to turn that documentation into the most precise code possible—whether in ICD-9, ICD-10, or any other future code set.
The Limitations and Lessons of ICD-9
Even with the comparatively small ICD-9 code set, too often coders today are left selecting either unspecified or minimally descriptive diagnoses. Take, for instance, diabetes mellitus code 250.0x, which describes patients with type 2 diabetes who do not suffer complications. A fifth digit of 0 indicates diabetes that is not stated as uncontrolled. A fifth digit of 2, on the other hand, describes an uncontrolled condition.
Providers often document blood glucose values in the patient record that clearly indicate—due to their clinical knowledge—that the condition is uncontrolled. Yet coders are not allowed to infer whether diabetes is controlled or uncontrolled based on clinical data such as lab values. A provider must write the word “uncontrolled” in the chart before a coder may assign the most appropriate 250.02 code. Failure to write uncontrolled requires coders to default to the less specific 250.00.
This is just one example of why improving documentation and code specificity is so important. With diabetes, controlled vs. uncontrolled status has major ramifications to treatment plans and potential clinical outcomes. The diagnosis code reflects this significant difference. As a result, coding 250.00 instead of 250.02 can adversely affect everything from reimbursement and care coordination to clinical research on this serious disease.
One lesson learned from ICD-9 is that the ability to use data to improve clinical care and lower costs requires accurate and complete information. It is difficult to track, coordinate, and manage care—whether for individual patients or large patient populations—without the ability to correctly report and share clinical details. ICD-10 captures even more of the granularity expected to help move clinical best practices to new heights.
Enhancements and Caveats of ICD-10
ICD-10 allows a more comprehensive view of patient care. A single leg fracture code in ICD-10, for instance, can disclose not only which bone is broken and whether the break is open or closed but can also reveal which leg is involved, whether the encounter is an initial or subsequent visit, and how the fracture is healing.
ICD-10 also brings diagnosis codes in line with the current practice of medicine. Much has changed about medicine since ICD-9 was introduced in the United States in 1979, of course, but unfortunately ICD-9 has been unable to adequately reflect many of those changes. Asthma is one example. Current ICD-9 codes still classify this chronic disease as extrinsic (493.0x) and intrinsic (493.1x), despite the fact that these factors are considered somewhat irrelevant in the new clinical definitions of asthma. Providers instead now focus on whether a patient’s asthma is intermittent or persistent. These two terms greatly impact the prescribed drug treatments, and ICD-10 asthma codes such as J45.20 (mild intermittent asthma, uncomplicated) and J45.30 (mild persistent asthma, uncomplicated) more accurately capture this terminology.
In fact, ICD-10 provides greater accuracy through both the terminology used and more specific individual code descriptors. There are simply more codes to describe current and relevant clinical information.
However, even with all the benefits of ICD-10, it is important not to overlook its shortcomings. ICD-10 remains a single, hierarchical classification system just like ICD-9. It carries many of the same rules and conventions as well. Ultimately, providers do not want to have to think about the coding rules of ICD-10 any more than they want to think about them in ICD-9. Consequently, many healthcare organizations have begun looking to HIT to ease the burden.
Keep in mind that in addition to clinical use, diagnosis codes also are at the heart of reimbursement. They supply the medical necessity for doing procedures and are the foundation of the diagnosis-related group prospective payment system for inpatient admissions. Therefore, the move from ICD-9 to ICD-10 could significantly impact future reimbursement.
This financial implication has caused a few healthcare organizations to begin coding in ICD-10 already, alongside ICD-9. With the help of HIT applications such as LEAP I-10, these early adopters are beginning to assess how the coding change will impact reimbursement. A few health systems already are working proactively with their payers to manage the transition.
While payers work to adjust their policies to reflect the increased specificity of ICD-10, providers can do the same. Providers also can use ICD-10 to finally receive accurate reimbursement for all the work they do. First, however, they must have a way to capture the increased level of specificity in their documentation.
Tools such as the Provider Friendly Terminology (PFT) content set, for example, can be used within an EHR at the point of care to prompt providers for the documentation necessary for accurate coding. A provider treating a fracture, for instance, may be asked to specify malunion or nonunion where applicable. In addition, providers and coders can search on the colloquial terms and phrases they are used to using, and PFT can help them code these terms to ICD-9, ICD-10, and even SNOMED CT.
This kind of utility not only helps providers improve the specificity of their documentation within workflow but also eases the mapping of that documentation to a variety of code sets. It can be used to generate reports that help pinpoint precise documentation training needs. Furthermore, it helps manage more than just the ICD-10 conversion—it helps manage change itself.
Managing Continual Change
The pace at which medicine is moving forward is nothing short of astounding. From EHRs to value-based purchasing and meaningful use, healthcare organizations increasingly will use health information to better inform financial, operational, and clinical decisions. In fact, there is a natural synergy between today’s efforts to achieve the meaningful use of HIT and tomorrow’s transitions to ICD-10, ICD-11, SNOMED CT, or other as-yet-unknown nomenclatures. The whole idea behind meaningful use is to encourage the adoption of HIT to ease continual improvements to patient care, no matter what form they may take.
When we inaccurately group lots of patients under the same ICD-9 code today, we lose important information and critical differentiators. ICD-10 will better describe our patients, allowing better clinical studies, better clinical decision support, and better patient education. With the proper HIT in place, healthcare organizations can avoid much of the upheaval now associated with the ICD-10 transition and instead begin embracing the promise that diagnosis specificity holds.
— George T. Schwend is cofounder, president, and CEO of Health Language Inc, a global provider of software for managing and updating standard and localized healthcare terminology.
— Brian Levy, MD, is senior vice president and chief medical officer of Health Language and a practicing board-certified internist who has spent years in the field of medical informatics. He has helped develop terminologies, clinical content, and the use of Web-based software by patients and physicians to improve care delivery.