ICD-10: Design for change

Sept. 21, 2015 / By Rhonda Butler

In Neil Stephenson’s latest novel Seveneves, one of the main characters says, “politics” is a word nerds use when they feel impatient about the human realities of an organization. When I read that sentence, I got a shock of recognition—yes! That is me, exactly! (I didn’t even mind the nerd part, because some nerds are pretty cool. Steve Jobs. Ruth Bader Ginsburg. Dr. Who).

Over the past dozen years, I could have developed some patience with the human realities of getting organizations to prepare for ICD-10. But no, the closer the implementation date gets, the more impatient I become.

I think I know why that is—I see organizations treating ICD-10 implementation on October 1 as if it were a test they can cram for. And the crazy part is, I am one of the ones trying to help them cram (My recently reprised DIY ICD-10 cheat sheet blog is a prime example). While one half of me is helping people cram, the other half wants to wave my arms about and spout classic junior high school teacher clichés like, This is the 21st century, people! Get with the program! It makes me feel like I have multiple personality disorder.

Organizations trying to simply get through the transition are missing out. The disruption of business as usual is actually a great time to take a look around, not just at what you do but at how you do it, and figure out how it can be done better, faster, cheaper. To flourish in this world, your organization needs to design for change, so that it is no big deal, so that frequent change can be accommodated with minimal disruption.

How do you know if you crammed for ICD-10 instead of designing for change? If you are a large-ish organization, and in the process of implementing you uncovered this egregious thing: programs for handling data or policies are written in a computer language that no one who works in your organization can read anymore. Did you pay a retired programmer whatever he asks to come back and tweak the programs? That is classic cramming. In your defense, designing for change would likely be expensive in time/money—using your in-house team to rewrite the programs from first principles in a current, maintainable, standard language.

A less egregious but more common bad thing that organizations of all sizes can run into is discovering reports, policies or documents that contain “code ranges” in ICD-9 (for example “diabetes is 249xx-250xx”). Cramming is—translating your ICD-9 code ranges to ICD-10 code ranges (e.g. “diabetes is E08-E13”). This is risky, because the classification has changed dramatically in some places, and you risk getting diagnosis codes (and therefore types of patients) included in the ICD-10 version of your policy that were not included in the ICD-9 version.

Designing for change in this case is highly recommended. From now on, store your code-based content as complete, exhaustive lists, spreadsheets or databased tables. In order to come up with a correct translation of all of the ICD-9 codes in a range, you need to list them out as individual codes and translate to ICD-10 anyway, using the GEMs or a GEMs-based tool. After that, you have the complete, exhaustive list of ICD-10 codes, so formatting them as ranges is an extra step that can introduce error into the process.

If your report, policy or document refers to ICD-10 procedure codes, you can safely and efficiently represent them as “hyperslices.” (Hyperslices refer to a PCS code as it could be expressed in a database query. It has additional meaning for software engineers who use high-powered math in their work, for instance with auto-coding—plus it sounds cool).

Here is an example of a hyperslice: If you are noting all of the PCS codes for closed biopsies of the lung, and you don’t want to list out all of the codes separately, you can capture the same information efficiently and accurately in a single statement, where each axis of classification in the seven-character code is spelled out: 0BB[CDFGHJKLM][38]ZX. The letters and numbers in brackets contain all of the valid “values” for the fourth and fifth axis of classification in your code. The fourth axis specifies the body part, and in this case contains all of the separate “lung” body part values. For example value C specifies the right upper lobe of the lung. The fifth axis specifies the approach, and the numbers 3 and 8 specify the percutaneous and endoscopic approaches respectively.

We live in a world in which cheap data storage has transformed so many aspects of our lives, and where lots of information and tools are freely shared. Things that used to be difficult when information was expensive and hard to come by are not that difficult anymore. Do not make the mistake of thinking something is hard because historically it was hard. That was history.

Change is the new normal. If the ICD-10 experience has anything to teach us, it should be to design for change.

Rhonda Butler is a senior clinical research analyst with 3M Health Information Systems.

The ICD-10 finish line is closer than ever.  How did your organization prepare for the change?  Share your journey and access valuable ICD-10 resources here.