Proceed with caution: Why code crosswalks are built to fail

Code crosswalks are everywhere in healthcare, but there’s no safety in numbers. In fact, this practice is full of pitfalls that have far-reaching effects.
Crosswalk

Crosswalking is a common practice for those who work with healthcare data. It is meticulous, detailed work that translates or compares data between disparate systems such as standardized code sets and multiple versions of the same code set. In essence, code crosswalks help to identify equivalents, if and where they exist.

Why are crosswalks needed?

Healthcare data is diverse and complex, in large part because there are so many ways to document this information. For instance, lab data is captured with a different terminology and code set(s) than patient problems. And procedure data requires different codes than the ones used for genetics and genomics. In each case, important patient data is being captured, but each one uses a different “language” – often for a different purpose.

Every day, across the healthcare ecosystem, there are numerous scenarios in which crosswalks are employed, a few of which are described below.

  • Code sets are regularly updated – some multiple times a year. Crosswalks can be used to bridge smaller temporal gaps such as updating SNOMED CT® codes from January to July or larger jumps like ICD-9-CM to ICD-10-CM.
  • Crosswalks can be used to help with billing, translating from one code set to another to meet specific billing requirements.
  • Similarly, regulations may require reporting in one code system over another, necessitating a bridge between the two.
  • Data from disparate sources within a given system or received from an external source may require crosswalks if there is more than one industry standard code set per domain.
  • And for health IT organizations that aggregate and help share clinical data between sources, crosswalks are common, especially when bridging SNOMED CT to ICD-10-CM or CPT® to SNOMED CT.

The inevitable pitfalls of code crosswalks

Despite the clear need for code crosswalks, these bridges are temporary – and tenuous. Here are a few reasons why.

Crosswalks get outdated. Fast. Each code set has its own cadence for updates. For example, ICD-10-CM makes its changes on April 1 and October 1, while LOINC® updates in February and August. And each time these changes happen, crosswalks involving updated codes can break. Staff need to be on top of the update schedule, the actual update releases, and must make individual fixes to impacted codes…if they can find them. In addition, they need to know which version of a crosswalk they are using before making any modifications.

Crosswalk maintenance is manual and specialized. Organizations that don’t outsource this work need to build highly skilled teams of coders, terminologists, mappers, and clinical informaticists to get the job done. The process is time consuming, highly detailed, and expensive. And the cycle never ends.

There’s not always a right answer. Crosswalks aren’t always a one-to-one situation. They can be, but often they require approximations; or it’s a one code to many codes relationship; and in some cases, there’s simply no match. As for local codes, they can be crosswalked to an industry standard but that requires additional time, effort, and knowledge of what the local codes represent. And to keep things interesting, while you can crosswalk from a specific code to a more general one, it’s not necessarily the same result in reverse.

Loss of specificity. Since each code set is unique, crosswalks often result in a loss of specificity. The detail captured by one code set may not translate to another – a situation often found when bridging SNOMED CT to ICD-10-CM.

A crosswalk is a judgement call. Often the decision to create a specific crosswalk is made by a single person. It’s an opinion – hopefully an informed and measured one. The implications, however, reach far beyond the individual. Even if a single person makes the judgement call, a process needs to be put into place to track, share, and socialize these decisions. After all, the next round of updates is likely just a few months away and that crosswalk may not work once the codes are revised.

When good crosswalks go bad

Given all the ways that crosswalks can fail, and the likelihood that they will break unless carefully and continuously maintained, it’s important to understand the implications. Ultimately, broken crosswalks will result in inaccurate, incomplete data. But what does that really mean?

For patients, incorrect data can impact everything from the care they receive to specialist referrals to how they are billed. For hospitals and hospital systems, poor quality data will inevitably yield poor quality analytics. Those involved in cohort identification for initiatives like clinical trials will struggle to find the right patients if coded data is flawed. Health IT organizations that aggregate and share clinical data will have to manage inconsistencies and information gaps. And population health initiatives that rely upon data built on faulty crosswalks will likely generate faulty outcomes.

So, how do smart healthcare organizations solve for precarious crosswalks and constantly changing codes?

The short answer is, they don’t. They don’t fight this losing battle or deal with crosswalks at all. Instead, they look to a trusted vendor with a skilled and experienced team that not only manages a robust clinical terminology, but ensures each term or clinical concept is always mapped to the right codes and industry standard code sets. And with terminology at the heart of code mapping, the need for crosswalks is eliminated. It’s really that simple.

To learn how IMO’s clinical terminology makes crosswalks obsolete, click here

SNOMED and SNOMED CT® are registered trademarks of SNOMED International.
CPT is a registered trademark of the American Medical Association. All rights reserved.

Ideas are meant for sharing.

Sign up today and have Ideas delivered straight to your inbox.

Latest Ideas​

Hear how the Piedmont team approached a successful implementation of IMO Health’s surgical scheduling data solution and take away helpful best practices
In May, the AMA met to discuss adding more CPT codes to the RPM section – but updates and revisions have been
Learn how value sets impact data use and EHR workflows, plus how organizations can enhance their creation and maintenance with innovative tools.