National Countdown to ICD-10: A Solution

DISCLOSURE: Reminder that I am a practicing physician, hospital/health system administrator, and also the Chief Medical Officer of a CDI/ICD-10 vendor called ICDLogic.

Want to thank all of you for waiting so patiently for this follow up post to National Countdown to ICD-10: A Status Report.

Co-author: Brian Hannah, MD, MS, CPHIMS, CMIO, Mercy Health System, Conshocken, PA

SIGN THE PETITION

 

Introduction

As previously stated, despite heavy preparation and recent developments/concessions from regulators, I continue to have significant concerns about the upcoming transition to ICD-10.  My concern is primarily for the independent practitioner (of which I am one) and the already decreasing number of small private practices in the US- approximately 60% of our nations doctors.  I'm concerned about the impact of I10 on the routine workflow of doctors and is well-documented in my previous post, this post will focus on a potential solution.

This solution is not my idea, is not new, and is definitely not easy, but it is something that I believe is absolutely necessary for our great Nation to have the highest level of success with the I10 transition.  In brief, it definitively shifts the language of clinical coding from one standard to another.  Namely, ICD-10 to SNOMED-CT.

Overview

When Mr. Smith gets admitted to the hospital with a severe case of asthma, his doctors and nurses create documentation of his visit.  This documentation is written in clinical language that each provider has learned through their training.  Though there are small variations in the style of documentation, the quality of documentation is most often dependent on the provider's level of experience.  It generally takes 3-5 years for most providers to learn their communities conventions but most would not argue that there is only small variation between communities of providers.  This means that the description of asthma in Boston is pretty much the same in Florida or New Mexico.

Mr. Smith's asthma is documented in clinical language- nurses, doctors, payers, and many other officials use this documentation throughout his visit and afterwards.  Upon discharge, his chart is whisked away to the Hospital Information Management Department (aka Medical Records) where an entire world of other people begin reviewing the case and initiate/continue the process of medical coding.  Just like we are communicating in English right now, this medical coding is currently done in the ICD-9 code set.  The coders read the chart and take literal documentation and convert it into codes, ie. the words describing his asthmatic attack (status asthmaticus) would be converted into 493.21.  In ICD-10, depending on a bunch of other things (previously outlined), it might code to J44.0.  Whether you use 493.21 or J44.0, most physicians (including me) will say, "What the heck are you talking about!?!??"  You see, only payers and regulators "care" about this... well, I guess all the reform has also now engaged the patient... rats, I hate when Obamacare comes out on top!!

This process is fine and good in the hospital because we generally have the resources to provision an adequate number of qualified medical coders to abstract this data, but it is quite different in the private setting.  It is my opinion that our payers have very noble intention's to shift the responsibility of coding to physicians so that various issues (including errors of omission) in the abstraction process are minimized/eliminated- ICD-10 can do a fairly good job of that, but at significant expense to the provider (as previously documented).  I'm sure it has nothing to do with the fact that coders are "middle-men" in this process and would like to be paid for their service; after all, we all know how generous our Payers are...

The Issue

There are approximately 18,000 ICD-9 codes- this may sound like a lot, but it is actually not that many if you consider there are dozens of medical specialties that carve out some subset of this number.  So even if there were only 40 subspecialties, we're looking at approximately 500 codes per specialty- give or take...  Most docs have memorized at least 500 or so obscure facts in their first semester of medical school.  Even still, computer-based systems were probably not as good as a good pocket card or "Superbill" (which is just a big pocket card) that listed the most common diagnoses.  The problem is that we're now going from 15 to 50 to hundreds of codes per diagnosis.  To be more precise, from 13,000 ICD-9 CM codes to 68,000 ICD-10 CM codes and 4,000 ICD-9 PCS codes to 83,000 ICD-10 PCS codes- that's from 17,000 to 151,000 codes!!!  Using our previous calculations, even our brainiest docs will take quite a bit of time memorizing 4000 new codes... but the real question is, do we really want them too??

The Solution

So what if they didn't memorize codes??  What if they documented in SNOMED CT?  Like Spanish, SNOMED CT is just another language (for all you technogeek Informaticists, it's a Terminology).  The difference though is that doctors already KNOW SNOMED CT!!  It was developed from the very same textbooks that we all studied throughout our training.  It is the language that we are already speaking, reading, and writing in... it doesn't need to be abstracted by coders, it IS the code in and of itself.  It is the preferred way to INPUT data, it is not so good at REPORTING data, that is ICD-10's strength (hey technogeeks, it's a Classification).  In other words, SNOMED CT is great for inputting data where ICD-10 is great for reporting outcomes.

SNOMED CT has over 300,000 terms that can be combined in nearly unlimited permutations.  Though this may sound daunting to us humans, computers eat that kind of thing up!!  However, computer systems with this capability are non-existent and though I'm sure there are people working on this problem, some of our brightest programmers are working on the next knock off of Angry Birds (personally I'm a Plants vs Zombies guy myself...)

The good news is that Meaningful Use criteria are essentially requiring the use of SNOMED CT for certain documentation.  The bad news is that there is no "lobby" or special group out there that is powerful enough to convince Payers and regulators that they should invest in helping EHRs and other vendors to produce systems that can convert SNOMED CT into ICD-xx, LOINC, and other code sets of tomorrow.  EHRs are particularly woeful in giving clinicians tools to document in ICD-10 in the first place, so can essentially "bail-out" a lot of EHRs from having to invest in this activity.

The Proposal

So now that you know all of the above, let's do this:

1. Make CMS adopt SNOMED CT as the de facto standard for clinical documentation (it kind of is anyway)
2. Identify funding for EHRs and vendors to produce SNOMED CT to ICD-10 conversion systems
3. Continue to push Meaningful Use incentives
4. Live happily ever after...

If you think this is a reasonable proposal, please sign our petition with your support (no purchase required!!):

SIGN THE PETITION

Discussion

Clinicians and payers can both be satisfied by this proposal, but how often is there REALLY a WIN-WIN in healthcare administration?  Experienced professionals might say rarely and I would have to agree...

The real truth is that there is a third party in this story that is often under-recognized.  That is our friendly-neighborhood HIM professionals.  Unfortunately, just like the rest of us, this party has a silo of their own and might be quite content with things as they are- you see, coders generally work in the HIM department, and they are just as averse to change as the next guy.  Any changes to this portion of the revenue cycle will mean retraining for all of these colleagues...

So at the end of the day, if this proposal is successful, payers will be none-the-wiser (they get ICD-9/10), incentivize clinicians to document in SNOMED CT (which they are doing anyway), create jobs (and profits) for EHRs and vendors (after all, we don't want them to starve), and put a bunch of coders out of business (they are already used to starving).  

The hope is that patients will somehow benefit from all of this...  Maybe I'm a little biased, but as a patient, I really want my doctor to at least FEEL supported!  If nothing else, this proposal can at least do that.  All in a days work when it comes to clinical transformation.

SIGN THE PETITION

Glossary

CMS- Center of Medicare (and Medicaid) Services
EHR- Electronic Health Record

HIM- Health Information Management 
ICD-9- International Classification of Disease version 9
ICD-10- International Classification of Disease version 10
LOINC- Logical Observation Identifiers Names and Codes

SNOMED CT- Systematized Nomenclature of Medicine Clinical Terms