Ask ACDIS: Escalation policies and clinical validation queries

Association of Clinical Documentation Improvement Specialists, September 1, 2015

Q: There seems to be an abundance of articles on posing clinical validation queries and escalating such situations to the CDI physician advisor; however, there is not a wealth of information on what to do when the query is answered and the answer still does not support the diagnosis.

For example, if, after query response, the physician advisor doesn’t feel the documentation supports a diagnosis such as acute respiratory failure, could an internal policy allow us to not drop the bill with that code on the chart? Even after peer review, the physician may verbally agree but never amend the chart. What are other facilities doing to avoid denials on the back end and code the chart accurately the first time with a clinically supported diagnosis?

A: You bring up an excellent point. If the clinical indicators do not seem to support the diagnosis, it is the responsibility of the CDI/coding team to submit a query to the physician.

There are a few things you and your organization need to consider, however. First, the Official Guidelines for Coding and Reporting gives the responsibility for assigning diagnoses to the attending provider. As such, if the attending states there is acute respiratory failure and stands by the diagnosis even if the physician advisor disagrees, the Guidelines would support the coding of the diagnosis. Unfortunately, CDI does not have a process similar to the utilization review (UR) department. With the UR committee, members can “override” the admitting provider in regard to patient status. The Guidelines do not allow an override process when it comes to the attending provider and the patient diagnosis.

Best practice would be to make your queries a permanent part of the medical record to demonstrate CDI/ coding efforts in obtaining clarification regardless of the outcome. CDI/coding can’t be held accountable for the actions of the attending provider. I think many organizations fail to realize the importance of keeping the queries as part of the medical record. Even queries which aren’t answered by the provider or instances where there is disagreement with a query should be saved because they show transparency of the efforts to accurately assign codes to the record. Furthermore, maintaining a transparent process allows the CDI team and/or the denials management team to review the record and reinforce the importance of CDI efforts to the physician should that record’s claim later receive a denial.

I think a major issue is how to construct the clinical validation query. A multiple-choice query is often best in these situations. For example, I would write a query as follows:

Please clarify the status of the diagnosis “acute respiratory failure” as documented [where/when] in this patient who received a maximum of three liters of oxygen [whatever clinical indicators make you doubt the clinical validity of the diagnosis]. The acute respiratory failure was:


Ruled out

Without clinical significance

Unable to determine

Other: _____________

This approach allows providers an “out” if they were mistaken when they wrote the diagnosis, as they can always clarify the condition was without clinical significance and therefore shouldn’t be reported. However, if  an attending provider confirms the diagnosis even with what CDI/coding feels is a lack of clinical support, the diagnosis must be reported. This claim may receive a denial, but CDI/coding clarified the diagnosis with the provider, which is what they are required to do.

Unfortunately, many coders and CDI overstep their roles when they decide not to code a condition documented by the provider. There is no guideline or guidance that says the coder can choose not to report a diagnosis if it meets the definition of a reportable diagnosis according to UHDDS definitions. This issue was addressed in the 2008 AHIMA practice brief, “Managing an Effective Query Process,” with the following guidance:

Codes assigned to clinical data should be clearly and consistently supported by provider documentation. Providers often make clinical diagnoses that may not appear to be consistent with test results. For example, the provider may make a clinical determination that the patient has pneumonia when the results of the chest x-ray may be negative. Queries should not be used to question a provider’s clinical judgment, but rather to clarify documentation when it fails to meet any of the five criteria listed above—legibility, completeness, clarity, consistency, or precision … In situations where the provider’s documented diagnosis does not appear to be supported by clinical findings, a healthcare entity’s policies can provide guidance on a process for addressing the issue without querying the attending physician.

However, the 2013 ACDIS/AHIMA guidance titled “Guidelines for Achieving a Compliant Query Practice” states “generation of a query should be considered when health record documentation ... provides a diagnosis without underlying clinical validation.”In addition, the brief states:

The focus of external audits has expanded in recent years to include clinical validation review. … When a practitioner documents a diagnosis that does not appear to be supported by the clinical indicators in the health record, it is currently advised that a query be generated to address the conflict or that the conflict be addressed through the facility’s escalation policy.

Again, the brief doesn’t say the code should be omitted, only that it should be addressed. I would suggest that best practice is to have an organizational definition (including coders, providers, CDI specialists, compliance, and quality) for high-risk diagnoses like acute respiratory failure, severe malnutrition, etc., that include treatment parameters because the treatment is often the most powerful clinical indicator differentiating a minor condition from a more severe one.

Once definitions are established, I would query when a diagnosis is documented but doesn’t meet the established criteria. If the provider confirms the diagnosis after a clarifying query, I would report the diagnosis, keeping the query as part of the health record to demonstrate clarification was sought by the provider. If the provider habitually fails (i.e., determine how many is too many times—three, five, seven?) to use organizational definitions for diagnoses, that should trigger an escalation process (physician advisor, peer review, etc.)

Physician education may be a useful tactic as you may want to educate providers to document why the patient has acute respiratory failure when the “standard” clinical indicators aren’t met. Maybe the patient has underlying lung disease, etc., so the provider might document “unable to administer more than 40% oxygen as treatment of acute respiratory failure due to underlying COPD” or whatever the concern may be. Medicine doesn’t follow a recipe book, so the provider’s best course of action is to always document why he or she has deviated from the “expected” protocol/standard. If this documentation exists and it is reasonable, it will often support an appeal as CMS recognizes the provider’s clinical judgment.

Editor’s note: Cheryl Ericson, MS, RN, CCDS, CDI-P, CDI education director at ezDI, answered this question. Contact her at cherylericson@comcast.net.