Settling on a Secondary Diagnosis: Part II

EDITOR’S NOTE: This is the second in a two-part series on the subject of secondary diagnosis.

Related to last week’s discussion on the Uniform Hospital Discharge Data Set (UHDDS), the more pressing question is what to do with a relevant condition that is a preexisting comorbidity but is not necessarily evaluated, monitored, and treated or increases nursing care or length of stay (key criteria).

Well-meaning coders too closely follow those narrow key criteria and discard legitimate secondary diagnoses. Let us presume that you are not questioning the clinical validity, but whether the diagnosis can be considered a secondary diagnosis.

According to the ICD-10-CM Official Guidelines for Coding and Reporting FY 2017, Section III. A. Previous Conditions, “If the provider has included a diagnosis in the final diagnostic statement, such as the discharge summary or the face sheet, it should ordinarily be coded…” It goes on to recommend discounting “resolved conditions or diagnoses and status-post procedures from previous admission that have no bearing on the current stay.”

There are conditions that enter into the healthcare provider’s (HCP) thought process and affect how a patient is managed. Although these are certainly relevant, they may not meet those key criteria. This relates to the phrase, “…all conditions that coexist at the time of admission…”

In fact, many “personal history of” or “family history of” conditions fall into this category. The Official Guidelines anticipated this and inserted the following into the rules: “However, history codes (categories Z80–Z87) may be used as secondary codes if the historical condition or family history has an impact on current care or influences treatment.” The concept of impact on current care or influence of management is the crucial point.

Let me give you a few examples:

1. Early pregnancy in a patient who has been in a car accident
2. Paroxysmal atrial fibrillation (PAF) in a patient currently in sinus rhythm, complaining of light-headedness, diagnosed with dehydration and discharged from the emergency department (ED)
3. “Depression” in a patient for any other principal diagnosis who is on no maintenance medication
4. Neurofibromatosis in a patient with new-onset seizures

In the first example, Chapter 15 ICD-10 guidelines indicate that, unless a provider specifies that a pregnancy is incidental, which would permit the use of Z33.1 (pregnant state, incidental) it should be assumed that an O code is appropriate. As any clinician will tell you, being pregnant is almost invariably a factor in deciding on a care plan, regardless of gestation. Can you use that class of medication in pregnancy? Does the patient need to have an ultrasound to ensure that the pregnancy is intact? Even if the answer to those questions is no, the complexity of the patient is kicked up a notch just by the mere fact that she is pregnant, and her care is riskier. 

In the second case, if the patient and provider are convinced that the patient had been in, and converted out of, paroxysmal atrial fibrillation, and it is documented as a certain diagnosis, you code I48.0. PAF is, by definition, intermittent. If the patient had PAF in the past, and the HCP can’t assert that it had definitively happened during this incident, you would code Z86.79 (personal history of other diseases of the circulatory system) because you can only code to the highest degree of certainty.

Depression was one of those conditions that saw a shift between ICD-9 and ICD-10, and I think this condition may need further elucidation by query if it is not clear whether it is significant in this patient’s case. A patient may have had a single depressive episode many years previously (when does “in full remission” become “history of”?), and it really may not contribute to the calculation, or a patient may have recent significant depression and be undergoing ongoing counseling. Do you need to be sure that the treatment of principal diagnosis will not elicit another episode of, or worsen a current diagnosis of, depression? Choose a different medication? This latter situation would certainly be appropriate to capture the diagnosis. The former scenario is not as clear-cut.

The last example is the one which piqued my interest in this topic. I had reviewed a case like this and was astonished that both the coder and the coding supervisor did not feel that neurofibromatosis met secondary diagnosis criteria because they were using the more restrictive key criteria. There are conditions that may not have any effective therapy (like amyotrophic lateral sclerosis or Alzheimer’s disease) and may not increase nursing care early on in the course of the illness. These conditions still meet the “condition that coexists at the time of admission” criteria and are assuredly clinically relevant.

I will confess that I had trouble thinking of examples for you. Each time I thought I had one I would reject it because I found it really did meet those key criteria:

• Right-sided hemiplegia post-CVA (cerebrovascular accident)…increases nursing care
• Hyperlipidemia…treated with medication
• Long-term use of steroids in a patient, increasing likelihood of an opportunistic infection…treatment with those steroids and alteration of choice of antibiotic
• Diet-controlled diabetes…affects diet order
• Ankylosing spondylitis…depends on principal diagnosis. Could be relevant if it relates to a traumatic injury or if the patient takes pain medication.
• Computed tomography (CT) for pancreatitis with finding of diverticulosis…not relevant.

Eureka! The crux of the matter is relevancy. 

Relevant means that the consideration of the condition is pertinent to this encounter. Examples of conditions that may not be actively under treatment but are relevant: History of malignancy, long-term use of hormonal contraceptives (affects choice of antibiotic), GERD (gastroesophageal reflux disease) (on no medications, but may inform provider’s decision about other medications), pregnancy, alcohol dependence in remission, and psychiatric or neurologic conditions without effective medication or treatment.

Clinicians are mentally taking into consideration hundreds of details as they craft and adjust their treatment plans. The problem arises because they document precious few of the details. We are not going to get the HCP to lay out every idea he or she has; in fact, some of those thoughts are at a subconscious level, but are still critical to making good therapeutic decisions.

The presence of these conditions increases the complexity of the patient. If the clinical support isn’t obvious as to whether a condition is appropriate to capture as a secondary diagnosis, the coder may require more information, and the coder or clinical documentation improvement specialist (CDIS) may need to query.

It is hard to remember an acronym if it gets too long, but at the SAME TIME, it might be helpful to have a mnemonic:

S tudies

A ffect thought process/Altering or influencing treatment/Afflicted with at admission

M onitor

E valuate/Enlist consultant’s assistance

T reat/Therapy

I ncrease nursing care

M edicate

E xtend length of stay (LOS)

I don’t think you can memorize this easily, but you can print it out and post it by your computer. Try testing the patient’s condition against these criteria. If you can’t convince yourself that you have evidence that one of these has been satisfied, and the condition has relevance, ask the provider to substantiate it. He or she may not realize it, but the act of thinking about the condition in the context of the patient may actually improve the patient’s medical care.

To proactively help providers decide what needs to be in their progress note assessment and plan list or discharge summary, have them ask themselves three questions:

1. Is the condition relevant to this encounter?
2. Is the condition active (under active treatment or monitoring)?
3. Is the condition acute?

Providers need to be taught that conditions that are actively being treated but are not acute are likely considered chronic in coding parlance. Examples of this would be: Crohn’s disease, hypertension, and asthma not in exacerbation. HCPs need to be broken of the habit of referring to everything as “history of…,” which to them means, “This patient has a medical history that includes…”

I am visual, so I made the decision tree below for you. 

FIGURE 1

If you aren’t sure from the documentation, you may need to have a discussion or compose a query. Our collective goal is to capture only relevant diagnoses that contribute to the severity of illness and complexity of the patient and to ensure that our patients are receiving the highest quality of medical care. I hope this helps!