Guest post: Three things to consider before you query

by Howard Rodenberg, MD, MPH, CCDS

I know this will come as a shock to my many (two) readers, but I’m not quite as perfect as I make myself out to be. I, too, have character flaws. One is that I ask questions. A lot of questions. I do so in order to find out information (“How was your day?”), to make sure someone is certain in their life choices (“Are you sure you don’t want Rice-a-Roni?”), and to hear myself talk. Truth be told, I really love me.

That’s why CDI and I are a good fit. We’re all about questions, although we say it’s a query to make it sound more important. But since the Dental Empress and our boys get exasperated with my questions (and they supposedly like me), it’s not hard to understand why physicians get query fatigue. It’s important for everyone to know that when we query a clinician, we’re doing so for the right reason. In an earlier piece for the ACDIS Blog, I proposed a three-part test for a query. Does the CDI query make a fiscal, regulatory, or clinical difference? If not, the query probably shouldn’t be asked, as it’s simply spinning wheels, driving metrics and demanding specificity simply for their own sake.

I’m very fortunate that we have an extremely high physician response rate to our CDI queries. Still, it only takes one or two inappropriate queries to a provider for them to spread the word that the entire CDI process is nothing but a pain in the tail. And as the CDI physician advisor, that’s when I hear about it. (Most physicians realize that the power dynamics of hospitals mean that they can’t really complain out loud about staff, as it can be seen as bullying and harassment. But complaining bitterly to another doctor? Just another day at the office.)

My solution, of course, is to ask more questions. And as a result, I’d like to propose three more questions to consider once you’ve decided a query’s in order.

Is this the correct timing?

The first question to ask has to do with timing. One of the more frequent comments I hear is that doctors are being asked to answer a query when they simply don’t have the information to do so. In CDI world, we want to get that query out as soon as possible. The sooner we get a diagnosis on the record, the more chances it will “stick” and be continually reinforced with additional documentation throughout the chart.

We often forget, however, that diagnoses are only established “after study,” and the amount of study required varies. But if clinicians are asked to answer a query when they truly can’t, nothing good happens. They either have to note something that may be inaccurate to make the query go away and remove it from their workflow; they can document that the diagnosis is “undetermined,” which doesn’t clarify anything and prevents us from asking a second query for the same problem even if more information is available; or they can simply not answer, which results in a flurry of follow-up and reminders which only serves to build resentment towards the CDI program.

One of the hallmarks of CDI, as we differentiate ourselves from coders, is that we apply clinical judgement to our query process. So, what I’m asking of my staff is to think as a clinician. If there’s not enough information in the record for you to be able to answer the query, why would there be enough for the physician to do so? Maybe best to wait a day or two for the case to flesh itself out before making the ask.

Issues of timing are also related to who receives the query. Recently I had an intensivist call me because three weeks into an admission, he had just come on service and was sent a query asking if a condition was present on admission. Wouldn’t that be better directed to the physician who admitted the patient? If there’s a question pertaining to a consult or a procedure, should not that query go to the appropriate specialist to clarify? “Why are you sending this to me?” is the most frequent whine I hear from my clinicians.

What’s the real question?

It’s easy for us to lose sight of the very questions we need to ask. The CDI specialist is asked to be part clinician, part coder, and part literary editor. As we stack these hats upon one another, trying to integrate often conflicting information from various sources, it’s almost inevitable that sometimes we miss the forest for the trees. A query that’s off-target may yield an inappropriate answer and deny you a second chance to ask the question you really needed answered.

One of the best examples I’ve seen of how the real question gets lost is the query for symptomatic or asymptomatic HIV disease. We’re all familiar with the fact that the presence of seropositive HIV infection without AIDS yields a specific code (Z21), while the presence of HIV disease/AIDS gives another (B20). The different codes have significant clinical and fiscal implications. If someone is known to be HIV positive and presents with an infection, the question one wants to know is if the infection is related to HIV or not. If so, you get code B20; if not, it’s Z21.

Easy, right? But Coding Clinic and other references confuse the picture by using terms such as symptomatic and asymptomatic HIV disease, which clinically mean nothing. If I had an HIV-associated infection last year, but now I’m doing fine, which of these do I have? The Centers for Disease Control and Prevention would still classify me as having AIDS (once you have an HIV-associated illness, the diagnosis stays in place) so I might be symptomatic. But I’m doing fine, so I’m asymptomatic, right? These terms simply get in the way of the real question, which is if the current infection is HIV-related. Always ensure you know what you really want to ask before you write the query.

It’s always helpful for me as I review queries to imagine I’m not talking to physicians, but what’s the bottom-line piece of information I want from the folks at the Quik-Trip. The way I would ask the question of the roller dog professional is probably how I should ask the doctor as well.

Is it integral?

We’re certainly familiar with the term “integral.” It’s Coding World’s excuse to not code certain diagnoses, and Payer World’s excuse not to fund their care.

I’m not certain we use the term properly. The book definition is that something is “integral” when it is “necessary, essential, or fundamental.” This is a key point that differentiates something that’s integral—you might think of it as required for the diagnosis—from something that is associated with it. For example, is metabolic acidosis integral in diabetic ketoacidosis (DKA)? I think we’d say yes. It is ketoacidosis, after all. But is metabolic encephalopathy integral to the diagnosis? It’s true to altered mental status may be associated with DKA, but it’s not a part of the of the case definition, so I would contend that it should be coded separately. To use a more subtle example, signs and symptoms of volume overload may be integral to congestive heart failure (CHF), but as volume overload can happen from extracardiac causes without deterioration in cardiac function, CHF may not be integral to volume overload.

Thinking of “integral to” as “associated with” leads down rabbit holes. A patient with vomiting presents with an elevated creatinine. Acute kidney injury is associated with dehydration. Dehydration is associated with vomiting. Suddenly vomiting becomes your principal diagnosis. As there is virtually nothing in medicine that can’t be associated with something else, using the term “integral to” in this fashion leads nowhere.

You can stop the spiral with clear and concise institutional definitions. These definitions spell out what is integral to a specific diagnosis and what is not. These definitions can vary between institutions (some hospitals may favor Sepsis-2; others like Sepsis-3). If your definitions differ from your neighbor down the road, that’s okay. Differences in institutional definitions don’t represent rights and wrongs, just variabilities in clinical practice. But they do help clarify the limits within which the term “integral to” should be applied and, more importantly, when it should not.

(A special thanks to my friend Lori Drodge, RHIT, CCS, DRG Coordinator at MaineHealth, who asked me to speak to her CDI Task Force this month. Preparing my presentation helped formalize the ideas I’ve shared. I also do weddings and Bar Mitzvahs. Thanks again!)

Editor’s note: Rodenberg is the adult physician advisor for CDI at Baptist Health in Jacksonville, Florida. Contact him at howard.rodenberg@bmcjax.com or follow his personal blog at writingwithscissors.blogspot.com. Opinions expressed are those of the author and do not necessarily represent those of ACDIS, HCPro, or any of its subsidiaries.