CDI Queries Work Best if the Recipient is Kept in Mind

I had an epiphany the other day while discussing compliant query composition with a very knowledgeable clinical documentation integrity specialist (CDIS). It will be easiest to explain if I share the original query first:

The following clinical indicators were noted in this patient’s medical record:

A 70-year-old female was admitted with sepsis, pyelonephritis, urinary tract infection, and documented “worsening altered mental status.” There was an infectious disease consult. The urine culture grew Klebsiella. The patient was treated with IV antibiotics. 

Please clarify the patient’s altered mental status.

Based on these clinical indicators and your professional judgment, please document in the medical record whether you believe any of the following conditions are present:

• Acute metabolic encephalopathy
• Septic encephalopathy
• Confusion only
• AMS with no further specificity
• Delirium
• Other (specify)
• Unable to determine

When a CDIS composes a query, they should be providing the provider with the clinical indicators they need to make a thoughtful, informed decision. The CDIS can pick and choose which clinical indicators to offer, but they should give both clinical indicators that support the condition they might be hoping to get in response AND clinical indicators that might not be consistent. The intent is to get the right answer, meaning the condition that is clinically valid and significant. The fact that a blood culture grew out streptococcus might be very pertinent in a clinical validation query regarding “probable gram-negative pneumonia.”

This was a made-up scenario, but other clinical indicators that might have been relevant could have been the results of blood cultures, information from a neurology consult, and whether the final mental status returned to baseline. And what did the discharge summary say?

Next, ensure that the question being asked is the question you want answered. In this case, the CDIS wants to know if the “altered mental status” could be categorized as some comorbid condition (not used in this context here as a CC or MCC), as opposed to a sign/symptom.

(As an aside, a symptom is a manifestation of a condition subjectively reported by the patient, whereas a sign is a manifestation which the provider objectively perceives, e.g., “felt feverish” versus T 39° Celsius)

The questions asked were: “Please clarify patient’s altered mental status” and do you “believe any of the following conditions were present?” The provider may think to himself/herself: saying “altered mental status” is pretty clear. The reader may disagree. Altered mental status could mean lots of things, including lethargy, confusion, or difficulty understanding or expressing oneself.

An alternate way to pose the query could have been, “Based on your clinical judgment, is there a more specific diagnosis that clarifies the patient’s altered mental status?”

We then honed in on the offered choices. My colleague felt we could eliminate “septic encephalopathy,” since it gets coded as metabolic encephalopathy anyway. This was emblematic of one of the key points of this article. Doctors don’t really do their documentation for coding. They do it for clinical communication. In fact, they probably don’t even know (or would particularly care) that “septic encephalopathy” is compliantly coded as “metabolic encephalopathy.”

But I wouldn’t remove that choice, because there may be providers who do use that terminology, and would feel it clarified the altered mental status. It also might serve as support for acute sepsis-related organ dysfunction (establishing sepsis). So, I would leave two choices that get coded the same way. I want the verbiage to feel authentic, in their voice.

I would also remove the “acute” from “acute metabolic encephalopathy.” I don’t want to leave words in choices that might make a clinician hesitate or scratch their head. What if they felt it had developed over two or three days and they really thought it was “subacute.” Would offering a choice with “acute” in it stymie them?

If the provider had described the altered mental status as “confusion” somewhere, then “confusion only” would be acceptable (even if it is undesirable!). If they had not, I would not potentially put those words in the provider’s mouth.

I also wouldn’t use “AMS” in a choice because I can’t compliantly index that to R41.82, Altered mental status, unspecified. I wouldn’t use an initialism here; I would type out “altered mental status.”

Another aside (from the CDC):

• Abbreviation: truncated word; e.g., “min” for minutes
• Acronym: made up of parts of phrases it stands for and pronounced as a word; e.g., SIRS for Systemic Inflammatory Response Syndrome
• Initialism: Similar to acronym, but pronounced by enunciating each letter; e.g., SOB for shortness of breath

What about delirium? Should we introduce a new condition that wasn’t mentioned in the record? It depends. Is it consistent with the clinical indicators? If the nurses or different providers mentioned waxing and waning attention or a fluctuating course, I would present that in my clinical indicators and then offer that a selection of “delirium” would not be inappropriate.

Lastly, I HATE “unable to determine” as a choice in multiple-choice queries. If you give an “other” or free-text option, you don’t need to use “unable to determine.” It is appropriate and “required” in POA and yes/no queries, per the Compliant Query Practice Brief. I don’t like setting myself up for the provider choosing an option that is uncodable, sets up more questions, or is not clarifying.

My advice is to make sure that every query is for a purpose (to clarify the record and make it as accurate and specific as possible) and ensure that it is understandable by the clinician. It doesn’t help the CDIS’s metrics and productivity to generate a query if it just confounds the provider and doesn’t result in a useful response.