This is the second installment in my series of articles addressing clinical validation denials. Clinical validation denials (CVDs) are the derivatives of diagnosis-related group (DRG) coding validation denials. The latter is a rejection of a claim on the basis of reviewing physician documentation and determining that the correct codes or sequencing were not present. The former occurs when the rejection stems from someone’s subjective opinion that the documented medical conditions were not present. 

Previously, we addressed encounters that really did have unsubstantiated diagnoses, for which the submitted DRG truly wasn’t accurate. The approach to those CVDs should be to try to avert them in the first place with concurrent clinical documentation review and pre-bill auditing. However, if a prudent practitioner in the same geographic location would not draw the same conclusion and make that diagnosis, you should graciously accept the CVD and return the overpayment.

Today, let us address situations in which the condition is probably present, but the evidence and/or documentation is suboptimal. If the preponderance of evidence makes you believe that it is more likely than not that the condition legitimately existed, you should at least try one appeal.

Realistically, these denials are becoming so commonplace that you may not have the resources to craft multiple denial appeals for cases with weak documentation. My personal experience is that composing a unique rebuttal takes between one and two hours. Your institution or system will need to decide how much time, energy, and money to expend, and this will determine how far you will take this type of CVD into the appeals process.

Keep in the back of your mind that the original reason for denials was to combat fraud, abuse, and waste. The theory is that if we take care of a patient with a constellation of conditions and diagnoses, we are entitled to be accurately and adequately compensated. That is the concept behind the DRG classification system. Payors were not supposed to withhold reimbursement because someone forgot to cross a “t” or dot an “i.”

I feel obliged to remind you all that I am a physician at my core, and I do not always agree with coding rules and guidelines, although I do abide by them. I do not appreciate the distinction between “uncontrolled” and “poorly controlled,” and I think it is whimsy that one indexes to “with hyperglycemia” and not the other. I don’t get why “leaking abdominal aortic aneurysm” can’t be translated to “with rupture,” because what else would it be?

I point this out because you may find that your providers align philosophically with me. You need them to understand that, just like they have to be compliant or they can get into trouble with the regulators, so must the coders. They understand that in order to take the best medical care of patients, everyone should be on the same page. Coders are not clinicians, nor are they mind-readers. And, as a last resort, if you need to appeal to your provider’s baser instincts, inform them that very soon their compensation will be inextricably bound with the hospital’s (if it isn’t already), and that it is in everyone’s best interest to avoid denials.

Why do I seem to be digressing? Because the examples I will give you as “suboptimal” documentation often seem completely adequate in the healthcare provider’s mind. It may take a bit of explaining to them why what they typed isn’t conveying what they think it is. And I think the approach to this intermediate type of CVD is three-pronged. First, you should try to avert it with concurrent clinical documentation improvement (CDI); next, you attempt the appeal/s; and finally, give the providers feedback to try to prevent a similar incident in the future.

Consider the following example:

An elderly gentleman with frequent urinary tract infections from pelvic mass with suprapubic catheter presents to the ED for altered mental status with “some element of cognitive dysfunction,” having reportedly “acted strangely.” A CT of the head was performed. After initiation of treatment, he became alert, answered questions appropriately, and “did not recall” his abnormal mentation from the previous day.

He was ultimately diagnosed with a pseudomonas urinary tract infection. A query was generated to see if there was clinical significance to his altered mental status with a response concluding in “probable metabolic encephalopathy.”

Now consider another example:

A demented elderly patient is admitted for first time in our system with creatinine of 2.13 (i.e., no old comparison level) and dehydration. The best creatinine level recorded over course of admission was 1.33, after receiving hydration, with serial electrolytes and renal function to monitor. Diagnosis of acute kidney injury denied.

An observation about CVDs, which have suboptimal support of the diagnosis, is that they often result from a condition which was skated around or implied, but the definitive codable diagnosis was only elicited by a CDI query or only noted in the discharge summary. This may mean that there is a single mention of the diagnosis in the encounter. It also heralds an opportunity for education regarding that condition and/or verbiage, because there were multiple missed opportunities to capture the diagnosis along the way.

Additionally, these CVDs often result from providers who are practicing the art of medicine and may be treating a patient who did not read the textbook. If you make an astute diagnosis, but an auditor with less clinical knowledge or experience than you can’t follow your reasoning, he or she may be inclined to conclude that the diagnosis was not valid.

In my final article in this series, I am going to describe how I approach crafting a denial appeal. I would like to address how to give providers formative feedback in this piece. This is not a query, so you can compliantly include quality or reimbursement impact if it serves your purposes.

Here are the things I keep in mind when composing my email:

  • There is evidence that learners only take away three to seven points from any given educational interaction. I try to limit my feedback to three points, and I present it in bullet points, as short and sweet as possible.
  • I attach the denial and my response. This means that, if they read my response, they may be getting some of the feedback more than once, in different phrasing. Repetition is good.
  • I try to provide the clinical evidence as opposed to being very coding-centric or theoretical.
  • If I have an appropriate CDI tip, I attach it to give the clinicians information so they can make educated decisions or diagnoses in the future (see an example at “Tip of The Week, ).
  • If I am sending feedback to the provider/s who actually took care of the patient, I leave the identifiers so they can review the actual encounter. If I am trying to disseminate feedback to an entire group or service line, I redact the protected health information (PHI). Even though this may be exempt from HIPAA because you are using it for teaching and quality purposes, why take a chance?
  • I also take the opportunity to let the clinician know that we, the administration, are acting on their behalf to protect them, and that no action is necessary on their part regarding this particular denial.

Let me give you an example:

Dr. Frankenstein:

Many clinicians are unaware that auditors and third-party payers may attempt to remove diagnoses from a patient months or years after the encounter has been concluded and billed.

We received the attached denial on _______________ (insert date here) regarding ________________ (insert patient name and medical record number here), and I wanted to share it with you. We have created a rebuttal on your behalf. Please read this email and the attached information, but you do not need to take any further action.

Takeaway points are:

  • Although we clinicians recognize that altered mental status is clinically important, the risk-adjusting term is “encephalopathy.” Please see attached CDI tip.
  • A best practice is to document a condition in codable format multiple times in the record. It is optimal to do so when it is first diagnosed, as you treat or resolve it, and then tie it up in a bow in the discharge summary.
  • If you receive a query and you agree, a best practice is to not only give the desired verbiage, but to bolster it by supporting it with your clinical indicators. For instance, had I answered this query, I might have documented something like:

“The initial cognitive dysfunction and strange behavior was so clinically significant that we obtained a CT scan of the brain. It ruled out any intracranial pathology, but we concluded the altered mental status was due to metabolic encephalopathy. We appropriately treated the dehydration and infection, and the encephalopathy resolved.”

If you have any questions, please feel free to contact me. We are working hard to make sure that you can continue to demonstrate the quality of excellent medical care you provide to our patients. Thank you for your hard work!


Me, physician advisor to my facility




CDI Tip on Altered Mental Status

Obviously, I adapt the points to address the deficiencies in that particular denial. You can feel free to template some of the text, but I always try to make something personal, if possible. I don’t want the provider to think that the entire email is cookie-cutter, or it might be dismissed out of hand. If I were disseminating this feedback, I would adapt the email accordingly.

If I find that a provider or service line needs the same type of feedback repeatedly, I approach the administration with an offer to educate. Some education is best accomplished face to face.

Let me conclude with the admonition that giving good feedback is time-consuming. I just know it is valuable and worth the effort. To paraphrase George Santayana, those who don’t learn from their mistakes are doomed to repeat them.

Share This Article