Updated on: March 21, 2016

Doc-to-Doc: Fear of the Unknown

By Robert S. Gold, MD
Original story posted on: February 9, 2015

 We’re getting closer to the probable implementation date for ICD-10, and we all know about the fear of the unknown that’s so prevalent in the healthcare industry.

 The major issues are the following:

1.  It’ll happen – but when?

2.  The considerations of a zillion new codes are fodder for consultants to sell multimillion-dollar programs or educational tools that speak coder language – yet doctors don’t speak coder language. And most of the training out there involves algorithms tailored to coders, not clinical thinking.

3.  Many of the new codes are intuitive – they’ve been requested by the physician specialties and are, to the greatest extent, representative of just how we physicians think of diseases and treatment of diseases. The information has to be presented that way.

4.  The idiocy of the environmental codes that have made headlines is staggering – SOMEONE wanted to know more about how something happened, such as catching fire while waterskiing. Doctors aren’t responsible for these things. Other people in the system will take care of these upon intake of the patient into the system.

What is important for physicians to know is that the better you describe what’s wrong with your patient, the better everyone else can treat your patient and take care of your patient’s needs. The better we can explain in the medical record that we know what’s wrong with the patient, what caused it, and what effect it is having on the patient’s other diseases or other organs, the better the patient will understand what’s happening, the better the family will understand what’s happening, and the better the insurance companies will understand what‘s happening.  I sort of equate it to how a physician normally tells a family member what’s wrong with mom – he or she usually does an admirable job of explaining all of that. Now, the physician is going to bill with the complexity of ICD-10 codes, and everyone wins.

The worst that can happen when the doctor takes a few seconds to think along these “complexity” lines is that we have a better concept of a more holistic way to treat the patient. And in the era of value-based purchasing and care integration and all the other buzzwords, this way, everyone is satisfied, from the consumer of healthcare to the payer for healthcare.

To say that you’re seeing a teenager for a follow-up of “DM” doesn’t tell anyone much. To say that the teenager has type 1 diabetes with progressive CKD, now at advanced Stage 3, and now needs attention to the anemia of the CKD (which explains the patient’s tiredness toward the end of the school day), and that we’re going to start erythropoietin, says a heck of a lot more. One simple ICD-10 code isn’t worth a heck of a lot. Several interrelated ones show far more complexity associated with the medical decision-making.

Learning the needs of the people we treat is a bonus, but if we know how to properly express in words what’s wrong with our patients, whether dealing with Medicare or Medicaid or Blue Cross or MetLife, an elderly or young patient, retired or actively working, everyone will do better and the coders will have an easier job interpreting what we intend to mean – and the data will show that we know what we’re doing and we deserve to get paid for it.

I recently visited a hospital system that was considering contracting with my company to provide training and updating of documentation needs for their medical staffs. During the visit, I was invited to see a demonstration of how some of the system’s IT-influenced physicians had been working with their electronic medical record (EMR) to ensure adherence to documentation and coding requirements for certain diagnoses. The goal was to make it easier for the physicians to select proper elements in their progress notes on the fly to ensure that everything needed for the proper coding of particular diseases was available for abstracting. They were interested in core measure practice guidelines that would validate their participation, ensure proper identification of diseases associated with extremes in severity of illness, and address risk of mortality and complications of care, all in order to demonstrate value basis of their care delivery. I was really impressed with the desires voiced by these members of the medical staff to do the right thing for the right reasons.

The first demonstration case was one of non-ST elevation myocardial infarction affecting a patient who had presented with chest pain and who had stage 4 chronic kidney disease. It was wonderful to see how the words that were needed appeared in the documentation, with minimal work needed by the doctor in the morass of EMR complexity, and how it all could be facilitated with fewer keystrokes. 

Then it hit me as I listened to the physicians before they went to the next demo case – the fifth digit of the code for non-ST elevation MI was wrong. The words were there – words that the physician would grab onto and click – yet it was that the words you could not see in the code description that were wrong. They had the system select 410.70, non-ST elevation myocardial infarction, unspecified episode of care.  

Asking them to go back to the list of codes, it was easy to see why this happened. With the coding logic they used, the physician had the opportunity to select non-ST elevation MI, non Q wave MI, and subendocardial MI as the top three choices, all with the fifth digit of “0” – unspecified episode of care. Following that were the three adjectives, with the fifth digit of 1 indicating acuity and then the same three adjectives appearing along with the digit of 2 for subsequent episode. In other words, the doctors saw the same disease nine times in a row, which created little inclination to point out the importance of the fifth digits, because there were just too many words on each line. Naturally, the doctor in the field as well as the doctor performing the demo selected the wrong code.

We heard in the most recent edition of Talk-Ten-Tuesdays that demonstrations and ICD-10 coding tests have resulted in a certain number of cases being denied payment or having reduced payment because of lack of specificity. I think here we have a pretty good reason why.

The algorithms selected by the electronic health record vendors present diagnostic elements in numerical order by ICD code, whether ICD-9 or ICD-10. And in most cases in which there are variations of a disease, the first code almost always has a terminal digit of “0.” And this is provided in both inpatient electronic records and outpatient electronic records, both for hospital and office coding. But wait! The companies all say that you can tailor your “pick list” to your specialty and to the environment of your practice. So physicians take a list of the diseases they deal with most often and give the list to a coder or to an IT person – and they get to the more pertinent diagnoses for the doctors to consider. But these people don’t take a clinical perspective on how they will sequence the codes – they do it numerically, because both groups of people (coders and IT people) think numerically. Doctors don’t. They are setting themselves up for denials and failure. And nobody will know why.

Is it ever appropriate to select a terminal digit of “0?” Certainly! In some diseases listed, the terminal digit of “0” has specificity attached to it just as much as the 1, 2, and so on. But with other conditions, early in a patient workup, you probably don’t know the specificity yet. It takes time and work to identify the cause of a cardiomyopathy (“0” indicates that it’s dilated, but you have no idea about the disease that caused the dilation), the cause of hypertension (default code is I10, which presumes that it is essential, but it may not be after workup), or the severity of the thyroiditis (“0” indicates that it’s acute, but you have no idea what caused it).

I think it would be helpful if doctors had a chance to select those acute conditions with specific causes and “other specifieds” before the chronic conditions with specific causes and other specifieds – and before any of the “I don’t know” or “unspecified codes.” In the office environment, I’d sequence the chronic cases with specified causes and the chronic cases with other specified cases before the acute cases, and all of them before the “I don’t know” codes. The practice environment and the likely found clinical scenarios should take precedence in pick lists, leaving the “I don’t know” codes to the very end. I mean, that’s how doctors work. They may not know yet, and when they do, they are eager to let the world know what they’ve found.

Take the opportunity to start trying these things out at your facilities. The automated pick lists are not physician-friendly. The pick lists tailored for you by IT staff and coders are not physician-friendly.

Unless the doctors who perform the revision of the pick lists are told about this particular issue, they likely won’t think of it themselves, because it’s usually a coder or an IT person leading the initiative. And they don’t always know what we do. Even going back into the history of code selection will likely not work, as the benchmark will show unspecified codes leading the pack.

About the Author

Robert S. Gold, MD, is a nationally known physician, responsible for having championed clinical documentation with a peer-to-peer educational approach in hospital organizations. Dr. Gold is a cofounder and the CEO for DCBA, Inc., a consulting firm that concentrates on development of Clinical Documentation Improvement (CDI) programs that aid in proper data streams, proper communication within the medical records and proper reimbursement.

Contact the Author

Comment on this Article

 

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.

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…
  • Tracing the Evolution of Clinical Documentation Integrity
    CDI is the critical link to quality outcomes.   Any healthcare system’s overall performance relies heavily on a strong clinical documentation integrity (CDI) program, and therefore on the professionals at the heart of this discipline. From patient outcomes and mortality…
  • 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…