Clinically Speaking: To code or not to code: That is the question
Decipher between impossible, improbable, and unlikely

By Robert S. Gold, MD

ICD-9-CM diagnostic and procedural code assignment usually follows certain patterns. This is because several elements are typically predictable: we know our docs and their quirks, we know their level of documentation, and we know the rules for assigning these codes. But at times, assigning diagnostic codes is muddy at best.
For example, a patient presents to the hospital with a set of signs or symptoms that may represent several different disease processes. The admitting physician usually takes one of the following three actions:

• He or she documents the abnormal lab findings, symptoms, or the physical findings
• He or she documents a list of differential diagnostic considerations
• He or she documents “rule out”

Then, either alone or with the help of additional tests and a group of consultants, the physician comes to one of the following four conclusions:

• The patient clearly has the condition that everyone has agreed upon
• The patient has a differential diagnosis, each element of which has been proven to be contributing to the presenting signs and symptoms
• The patient has a differential diagnosis, but the cause of the presenting signs and symptoms is unclear
• The patient’s condition is unclear (no diagnosis found) and the physician cannot come to a conclusion

We all know that the principal diagnosis is that which, after workup—led to the need for the inpatient hospital stay. Sometimes, secondary diagnoses may fall under the same decision tree. This is why it is important to make the determination only after workup.

MS-DRGs and documentation
Under MS-DRGs, there are a decreased number of diagnoses that CMS still considers to be complications and comorbidities (CC). A hospital’s financial dependence on the identification of a proper principal diagnosis, as well as a CC or MCC, means that coders may be under additional pressure to capture all conditions that will help yield a higher-paying MS-DRG. And sometimes one specific coding guideline gets in the way of truth and honesty.
Consider the following reference from Coding Clinic May/June 1984:

If the diagnosis at the time of discharge is stated as suspected, questionable, likely, or probable, code the condition as if it existed or was established. A diagnosis at the time of discharge prefaced by the term Rule out is to be translated as "suspected." A diagnosis stated as "Suspected __________________, but not confirmed" needs further consideration to determine whether it refers to a "suspected condition" at the time of discharge or a "suspected condition" at the time of admission but not established during the episode of hospitalization.

And we know the following from Coding Clinic Second Quarter 1988:

Coding guidelines for inconclusive diagnoses (probable, suspected, rule out, etc.) were developed for inpatient reporting and do not apply to outpatients.

We also know about the rules for comparative and contrasting conditions, which state the following:

Two or more comparative or contrasting conditions
In those rare instances when two or more contrasting or comparative diagnoses are documented as "either/or" (or similar terminology), they are coded as if the diagnoses were confirmed and the diagnoses are sequenced according to the circumstances of the admission. If no further determination can be made as to which diagnosis should be principal, either diagnosis may be sequenced first.

A symptom(s) followed by contrasting/comparative diagnoses
When a symptom(s) is followed by contrasting/comparative diagnoses, the symptom code is sequenced first. All the contrasting/comparative diagnoses should be coded as additional diagnoses.

Each of these references can spark discussion and potentially lead to controversy.

When a physician admits a patient and documents, “rule out MI [myocardial infarction],” the issue is that the patient has presented with signs and symptoms that might represent an MI, but they could also be due to a lot of other conditions. If the presentation is “typical,” it involves chest pain. If the presentation is “atypical,” it could manifest itself as shortness of breath or weakness.
It is up to the physician to determine whether the presentation—either typical or atypical—actually represents an acute MI. And the physician may perform EKG interpretation, cardiac enzyme determinations, and echoes and stress tests to see whether something acute is actually occurring.
What we need is for the physician—in face of normal enzymes, no change in the EKG from former tracings, and no signs of acute ischemic damage on the scans—to tell us whether the MI was ruled in or ruled out. But what do we do when we get nothing except, “Discharge—return to the office in two days?”
Do we code this as an MI? Not at all. We are not allowed to assume that just because there is no evidence whatsoever that the patient did not have any heart problems that the physician ruled out an MI. We have to go back to the physician and ask, “After all is said and done, what condition do you believe caused the patient’s presenting signs and symptoms?” If the physician is proactive, you’ll get the answer.
I received a question similar to the above just the other day. When a patient is admitted with “R/O ACS” (acute coronary syndrome) and there is no conclusion, the coders had been advised to assign code 411.1 because the physician used the tem “rule out.” That’s wrong!
There are professionals out there who will tell you to assign certain codes because the “coding guidelines say you can do it.” But if you read the citation a little more carefully, you’ll determine that this is not always the case. Consider the following excerpt from the ICD-9-CM Official Guidelines for Coding and Reporting:

Uncertain diagnosis
If the diagnosis documented at the time of discharge is qualified as "probable", "suspected", "likely", "questionable", "possible", or "still to be ruled out", or other similar terms indicating uncertainty, code the condition as if it existed or was established. The bases for these guidelines are the diagnostic workup, arrangements for further workup or observation, and initial therapeutic approach that correspond most closely with the established diagnosis.

This reference states that you should only code uncertain diagnoses as if they existed or as if the physician had established them when, AFTER WORK-UP, he or she still believes that the condition existed even though it could not be proven. If you are left with no conclusion, you have to go after one. You have to encourage the physician to validate that the condition was or was not acute coronary syndrome. If the doctor suspects that it was gastroesophageal reflux disease, you can code that when it is stated that this was probably due to GERD. If he or she thinks that it was probably the patient’s costochondritis, you can code that when it is stated that way. This is where the possible/probable/likely comes into play.  The doc believes, after work-up, that the symptoms were caused by this disease or condition, even though he didn’t prove it by tests.

Comparative/contrasting conditions
Let’s talk briefly about the comparative/contrasting issue. We spoke about a differential diagnosis when a patient presents to the hospital. It might be right upper quadrant abdominal pain with a list of possible gallbladder disease or ulcer disease or diverticulitis.
When workup shows that the patient has none of these, then the right upper quadrant abdominal pain should be the principal diagnosis. When the patient is found to have all of these conditions, but the physician isn’t sure which one caused the episode of pain, code all three of them and sequence them according to your policies and rules of ICD coding.
This is the rule’s intent. The intent is not to give a patient three diagnoses, none of which exist. However, if, at the time of discharge, the physician doesn’t know which condition the patient has, but still thinks that it could be any of them, code them as though they all exist. The main point to remember is that the physician’s uncertainty must occur at the time of discharge—and not before the workup is complete.