Q&A: Escalating the query process to include clinical accuracy

A: While we do not expect nor ask coders to vet every diagnosis written in the chart, we do expect them to send obvious concerns our way. Here is some background information representing my view as to how the problem has evolved and how we handle it.

 

First, all auditors (Recovery Auditors, or RACs), Medicare Administrative Contractors (MACs), and private payers, are allowed to question a physician’s clinical judgment while coders are not. Coders are only supposed to code what is written in the chart by the physician regardless of whether or not the physician is correct in his or her diagnosis. I find this to be extremely unfair and ridiculously favorable to the auditors.

 

Second, while stopping waste, fraud, and abuse are noble goals, the entities CMS hired are private companies whose primary concerns remain with the success of their own business and their own bottom line. 

 

Third, the last set of eyes to see a chart before it is submitted for reimbursement are the coder’s.  Therefore, the coder represents the last line of defense in the complex process of denial prevention.

With this in mind, the process at our facility evolved to the following:

 

1)      If the patient is still in the hospital and one of our CDI specialists sees a documented diagnosis that is not clinically present or does not meet an accepted clinical definition, they will either:

A)  Query the physician for clarification

B)  Send me (as the physician advisor) an email to let me know that a physician needs some gentle reminding about our “house-wide” accepted definitions

C)  Send a note to the coder not to code that particular diagnosis because it is not actually present

 

As to which of the above option or options are chosen depends on the circumstances in which the diagnosis is written. If the diagnosis in question is the only CC or MCC on the chart, the CDI specialist might perform all three options. If there are several other CCs or MCCs present, the CDI specialist might only choose Option C. If the physician routinely diagnoses something incorrectly, then I (again in the role of the physician advisor) will conduct an educational intervention.

 

2)    If the patient has already been discharged and the coder runs across a diagnosis that suddenly appeared in the chart and doesn’t seem to fit with that hospitalization, the coder will ask me to review the chart. If I make the determination that the documented diagnosis was not present, then we tell the coder not to code it. While we have been teaching our coders the clinical definition of various disease processes, they do not make the final decision as to whether or not a particular documented diagnosis existed.

 

Again, while we do not expect coders to be responsible for vetting every diagnosis in the record, we do expect them to send obvious concerns our way.

 

For example, if a patient has been in the hospital 10 days and all of a sudden “acute renal failure” shows up out of the blue, but the creatinine never increased, that should be a red flag to the coder that I need to review that chart.

 

As another example, we just had a case where a resident started documenting “sepsis present on admission” on the seventh day of a patient’s hospital stay, the patient went home on the ninth day. The resident documented this diagnosis on hospital day eight and nine and in the discharge summary as well. However, the patient was not admitted for an infectious issue and the coder accurately identified that the patient’s vital signs never met the definition of sepsis; the white blood cell count was essentially normal, and no antibiotics were ordered during the hospitalization.

 

The coder sent me an e-mail saying that the sepsis diagnosis would be the only MCC on the chart, that she did not believe it was clinically present, and that she did not want to code it. I reviewed it, the coder was correct, and we did not code “sepsis” on that case. This means that an unnecessary potential denial may have been averted.

 

In the interest of full disclosure, these scenarios tend to only happen with residents and generalists (hospitalists, traditional internal medicine physicians, and family practice physicians). Rarely does this happen with a specialist like a nephrologist because they were all part of the decision as to what parameters our facility uses to define acute renal failure.

 

Unfortunately, I believe the days of the coder blindly coding what the doctor wrote are over.  Coders are going to have to become more clinically orientated, partner more cohesively with the CDI team, and ask more questions of physicians. While this may stand in stark contrast to common practice, we cannot afford to have the facility and physician payments denied nor our hospital dragged through costly audits.

 

Editor’s Note: This Q&A was provided by Trey La Charité, MD, a hospitalist at the University of Tennessee Medical Center at Knoxville, and an ACDIS Advisory Board member. His comments and opinions do not necessarily reflect those of UTMCK or ACDIS. Contact him at clachari@utmck.edu.