Q&A: Coding based on clinical criteria
Q: I remember reading an article, or a Coding Clinic, that basically said it was inappropriate for facilities to set internal policies for coding conditions based on clinical criteria, but I can’t find the reference. Can point me in the right direction?
A: You are correct. The decision to code or not to code cannot be based on clinical indicators but must be based only on physician documentation. It is commonly referred to as “Guideline 19” from the Official Guidelines for Coding and Reporting, which can be found on p. 13 under the heading “Code Assignment and Clinical Criteria:”
The assignment of a diagnosis code is based on the provider’s diagnostic statement that the condition exists. The provider’s statement that the patient has a particular condition is sufficient. Code assignment is not based on clinical criteria used by the provider to establish the diagnosis.
Practically speaking, this means that a coding professional reviewing a record in which a physician uses Sepsis-3 criteria cannot omit the sepsis code based on the fact that they prefer Sepsis-1 criteria. Similarly, a coding professional reviewing a record with the diagnosis of acute kidney injury cannot omit the code if they think the patient recovered too quickly or the creatine was suspect.
What coding professionals can do is hold the documentation against the Uniform Hospital Discharge Data Set (UHDDS) guidelines that say reportable diagnoses must be:
- Evaluated
- Treated
- Diagnostically tested
- Use additional nursing services
- Extend the length of stay
Therefore, a coding professional can omit a code if, let’s say, a physician documented sepsis, but there’s no evidence of the diagnosis being evaluated, treated, or tested, and there’s no evidence of the diagnosis extending the patient’s length of stay or expending additional nursing services.
Guideline 19 does not supersede the UHDDS guidelines. A coding professional has to work to reconcile these two seemingly opposing edicts as best they can in their coding process.
There is one important additional piece from the American Hospital Association’s Coding Clinic, Fourth Quarter 2017, p. 110, that you may find helpful (emphasis added):
Facilities should also work with their medical staff to ensure conditions are appropriately diagnosed and documented. If after querying, the attending physician affirms that a patient has a particular condition in spite of certain clinical parameters not being met, the facility should request the physician document the clinical rationale and be prepared to defend the condition if challenged in an audit. The facility should assign the appropriate code(s) for the conditions documented.
This means that if a documented diagnosis does not appear to be clinically supported, the facility is justified in requesting the physician document their rationale via a clinical validation query. They cannot, however, censor the reported codes on the coding side. Usually this happens most often when the physician has a perfectly legitimate reason for making a diagnosis, but the patient has presented in an atypical way which may not be apparent to the non-physician reviewer on the initial review. Generally, all that is needed is further physician explanation of the atypical nature of the presentation and why it is a legitimate diagnosis even though it may appear to the untrained eye to be unsubstantiated.
Remember, you absolutely can have your physician advisor set clinical criteria for when a query should be placed. Query criteria are unrelated to the problem of code assignment based on clinical criteria. We certainly do want our physician advisors involved in the use of appropriate criteria for query purposes in order to avoid unnecessary, unreasonable, or leading query practices.
Nonetheless, there are always exceptions (referring to Guideline 19). For example, I do not advise one blindly follow the index on documentation such as “cardiac injury” to a traumatic injury code when that is clearly inappropriate. I also cannot advise one to blindly follow a 1992 Coding Clinic that suggests cardiac injury be reported as a myocardial infarction (MI) when the new fourth universal definition of MI clearly suggests that cardiac injury is for use in cases where infarction is not present.
As stated in Coding Clinic, Fourth Quarter 2018, p. 90, “ Coding Clinic itself doesn’t advise you follow the ICD-10 index entry for emaciation to the nutritional marasmus code but rather they recommend defaulting to the reporting of cachexia when in doubt. (See Coding Clinic, Third Quarter 2017.)
Editor’s note: Allen Frady, RN, BSN, CCDS, CCS, CRC, CDI education specialist for HCPro in Middleton, Massachusetts, answered this question. Contact him at AFrady@hcpro.com. For information regarding CDI Boot Camps, click here.