Guest Post: Query questions: When should you not query?

by Robert S. Gold, MD

I recently attended the annual meeting of the ACDIS in Orlando where I heard a variety of perspectives during many informative presentations.

A common thread in many discussions was how to query physicians using information in the medical record and when it’s appropriate to do so. Coders and CDI specialists must remember that a valid diagnosis must be supported by evidence in the body of the medical record. What does this mean?

Let’s consider the often difficult decisions associated with three common diagnoses: Anemia, acute renal failure, and sepsis.

Anemia

Too many coders see the word “anemia” in a chart and automatically assume they can ask whether it’s acute blood loss anemia. It’s an automatic response that’s born out of incorrect and misinformed teaching.

First-before initiating a query-coders must determine whether the patient is anemic. Is the patient anemic if the hemoglobin is 12.5? The answer could be no, but it could also be yes.

A patient with sickle-cell anemia has anemia regardless of the level of hemoglobin. If a physician doesn’t explicitly document sickle-cell anemia, coders may be confused. A patient with end-stage renal disease (ESRD) who has anemia of ESRD-and is currently receiving erythropoietin-has anemia of ESRD regardless of the hemoglobin level.

Next, coders must review the myriad of diseases that can cause anemia. ICD-9-CM codes exist for most of these anemias due to other diseases.

Don’t assume a diagnosis of anemia based solely on hemoglobin level. If a resident records hemoglobin as 8.5, what do many coders do? They prematurely ask for the diagnosis of anemia or even acute blood loss anemia.

Coders should do their homework beforehand. They shouldn’t ask a question unless they already know the probable answer. Consider these questions about anemia before requesting clarification:

• Does the hemoglobin level belong to this particular patient, or did the resident document someone else’s hemoglobin level in the progress note?
• Are laboratory results correct? Does the laboratory slip include the correct patient name? Check the level. Is the hemoglobin truly 8.5 and not 18.5?
• What are the clinical circumstances? If there is evidence of an acute or chronic bleed, ask for clarification. If there is no evidence, don’t ask.
• Is there a trend in the hemoglobin level? If the last few hemoglobin levels in the patient’s record are 8.5, the anemia has likely been there for a while, and a physician should document its cause. If the cause is not evident elsewhere in the record, request clarification.

Consider this example: A patient whose past hemoglobin levels are normally higher is hypotensive and receives two or three liters of IV fluids over a few hours to raise the blood pressure. Ask what led to the drop in hemoglobin. The answer may be that it’s ­dilutional. In this case, a physician may ­order a few more complete blood count tests ­during the workup to address the hypotension, especially when it involves an infection. Then, the physician will wait ­until the cause of the patient’s hypotension ­resolves and the hemoglobin level returns to the baseline. Was the patient anemic? No. There was simply a change in a laboratory value. Should a ­coder report anemia based on one laboratory value? No.

Dilution or hemoconcentration are two ­circumstances in which levels of the bloodstream’s components can change dramatically. Physicians can’t do much about these changes; however, they can lower the serum level of sodium by administering large volumes of D5W or half normal saline. They can also raise creatinines in cases of moderate dehydration with IV fluids alone.

Renal failure

If a patient is admitted with an elevated creatinine ­level or develops one during the hospital stay, some ­coders ­instantly ask the physician to document acute ­renal failure or acute kidney injury (AKI). Is this a correct practice based on a solitary laboratory value? Absolutely not!

Coders first must identify the patient’s baseline ­creatinine. If it’s usually elevated at this level, the ­present elevated creatinine level likely represents chronic kidney disease, and coders may ask physicians to clarify its stage and cause.

If the patient’s baseline creatinine level is usually lower-although not necessarily normal-and is now significantly raised, determine what’s wrong with the ­patient. If the patient has sepsis or is in cardiogenic shock, chances are the elevation represents decreased perfusion of the kidneys. In this case, requesting clarification of a pathologic reason for the elevation or spike in the creatinine level may be valid.

If dehydration is the only evident reason for the creatinine elevation, coders must consider the treatment and the patient’s response to it. If the patient receives IV fluids, and the creatinine returns to baseline within 24 hours, there is no AKI, according to the Acute Kidney Injury Network definitions.

However, if the creatinine does not return to ­normal-or almost normal-within 24 hours with IV fluid resuscitation, the patient likely has AKI. In this case, asking for clarification about whether the ­patient’s clinical circumstances meet the physician’s definition of AKI due to dehydration is wise.

Sepsis and SIRS

Similarly, coders shouldn’t encourage physicians to document sepsis or SIRS due to a urinary tract infection (UTI) without doing their homework first.

For example, a patient is admitted to the hospital with syncope, and a workup in the ED demonstrates tachycardia and leukocytosis. A urinalysis demonstrates a UTI. I recently encountered this situation. A resident assigned to the case documented that the patient had two of four SIRS criteria and therefore had sepsis. Every other note for the next three days included the same language copied and pasted into it.

However, upon review of the chart, it became apparent that the cardiologist evaluated and treated the ­patient’s tachycardia atrial fibrillation with rapid ventricular response, and it resolved with Cardizem®. It ­became even more apparent that the patient’s leukocytosis was linked to his recent steroid injection to treat ­arthritis. It was extremely apparent in the ED note that the UTI was found incidentally on routine admission laboratory reports. The patient had been totally asymptomatic.

Did the patient have sepsis or SIRS? No. Was the resident’s initial assessment wrong? No. Was everyone wrong in this continued use of the terminology upon clarification of the reasons for the abnormalities? Absolutely.

Coders should always ­review the medical record to ensure that it includes ­clinical evidence to support the diagnosis for which they seek clarification. Do this before you query.

Editor's note: This article was first published in the journal Briefings on Coding Compliance Strategies’ June edition. Gold founded DCBA, Inc., in Atlanta, a consulting firm that provides physician-to-physician programs in CDI. The goals are data accuracy, profile management, and compliance in the inpatient and outpatient arenas.