All diagnoses should be clinically valid.
Denial of reimbursement for medical care occurs for a multitude of reasons, including incomplete or inaccurate information, lack of demonstration of medical necessity, and coding-related issues. But documentation and coding denials comprise a small percentage of denials (fewer than 5 percent).
Historically, coding denials were based on Diagnosis-Related Group (DRG) validation. The auditor considered whether the correct codes and sequencing had been rendered, according to the documentation. The question asked and answered was, “was the encounter coded correctly?”
These types of denials still occur, but there has been a fundamental shift to clinical validation. Clinical validation (CV) is a determination of whether conditions documented in the medical record were really afflicting the patient. It is an assessment as to whether the diagnoses are supported by the clinical evidence.
There are several ways CV concerns enter into medical practice. The first is providers not making correct, accurate diagnoses. This can be a function of having incomplete data on which to base a diagnosis. For example, a provider may diagnose “possible Gram-negative pneumonia” in a patient from a nursing home. If the sputum and blood cultures were to return as pneumococcus, up crops a discrepancy. If the documentation doesn’t evolve to “pneumococcal pneumonia,” a CV issue has been created. Payors care, because this inaccuracy changes the DRG and inflates the reimbursement.
CV concerns can also arise from faulty understanding of clinical criteria, absence of definitive criteria, or honest mistakes. Providers may not be current with updated medical literature. Labeling every patient with hypoxia as suffering from “acute hypoxic respiratory failure” is an example.
Clinicians sometimes use terminology to justify their professional fee, unaware of the larger DRG implications. An example of this is use of the phrase “acute postprocedural respiratory failure” to justify critical care billing in the postoperative period, when the patient does not meet criteria for respiratory failure.
A more insidious or nefarious way that CV questions can be applied to the record is due to administrative overzealousness or provider exuberance. Administrators are aware that certain DRGs or capture of tier-shifting comorbid conditions or complications (CC/MCCs) result in higher reimbursement, and they sometimes encourage providers to document those conditions even if they are not clinically supported. Consider the documentation of acute systolic heart failure for every transcatheter aortic valve replacement (TAVR) patient, or “severe” protein-calorie malnutrition when “moderate” is more clinically appropriate.
But I would venture to say that far and away the most common cause of CV issues is substandard documentation. The adage “if you didn’t document it, you didn’t do it” is patently false. It can be labor-intensive and time-consuming to “think in ink.” However, without an explanation as to why a provider believes a patient may have a condition, despite clinical indicators to the contrary, the only conclusion an auditor may draw is that the condition didn’t truly exist.
Overreliance on templates can result in internal inconsistencies. Documentation of a “well-developed, well-nourished” patient with a templated normal musculoskeletal examination does not support a diagnosis of “severe protein-calorie malnutrition.” “In no acute distress” calls into question a diagnosis of “sepsis.” The question becomes, is the “well-developed, well-nourished” part correct, and there is no malnutrition, or is there malnutrition, and the patient is really not in such good shape?
Probably the biggest culprit is copying and pasting. Empiric diagnoses made early, before data has been accrued, get indiscriminately propagated without revision and perpetuate ruled-out diagnoses. Laundry lists of assessment and plan (A&P) repeat conditions ad infinitum.
Discharge summaries also may backpedal on diagnoses that seemed to be definitive, suggesting questionable clinical validity. Clinical documentation integrity (CDI) queries may cause providers to choose diagnoses that are not sufficiently supported, thus creating a CV concern. I always recommend that query responses be made in the medical record proper, rather than just appearing on a clarification request document.
The basis for the significance of CV is the False Claims Act of 1863, which makes it a crime for any person or organization to knowingly make a false record or file a false claim regarding any federal healthcare program. If a patient doesn’t have a diagnosis being proffered, this can be considered a false claim.
In September 2020, the U.S. Department of Health and Human Services (HHS) Office of Inspector General (OIG) released a report noting that millions of dollars had been paid to Medicare Advantage (MA) organizations for patients recorded as having acute strokes, when what they really had was a history of stroke. In my opinion, this likely stems from the coding-clinical disconnect of the qualifier “history of,” but the point is that CV is not only for inpatients. It relates to the HCC methodology as well. Clinical validity is important wherever and in whatever status patients are being treated.
What’s an organization to do?
The pervasive culture must be that the purpose of documentation is to tell the story, and tell the truth. All providers must strive for accuracy and completeness in the medical record. Policies must foster this, and electronic health records should enable it.
CDI professionals must be empowered to perform CV queries. As is pointed out in the American Health Information Management Association (AHIMA) Clinical Validation Practice Brief, updated January 2019, the CDI professional is not “performing clinical validation. Rather, they are highlighting a potential gap between a documented diagnosis and clinical evidence in the health record.” The actual validation is done by the healthcare professional answering the query. Providers must understand that CDI personnel are not prejudging a diagnosis as invalid. CV queries pose the question, “is this diagnosis clinically valid?” Providers must not view CV queries as an affront to their clinical judgment.
The coder must consider whether a condition is reportable. For an inpatient, was it present and relevant at the time of admission, or was there clinical evaluation, therapeutic treatment, diagnostic procedures, extension of length of stay, or increased need for nursing care or monitoring? For an outpatient, did it MEAT criteria (Monitor-Evaluation-Assess-Treat)?
The RAC (Recovery Audit Contractor) Statement of Work contends that “clinical validation is beyond the scope of DRG validation, and the skills of a certified coder. This type of review can only be performed by a clinician,” but that establishes the necessity of a clinician for the RACs, not for the healthcare organization. If a coder has enough knowledge and experience, he or she is perfectly competent to recognize a CV issue. In fact, I would go so far as to say that if any coder has a question as to whether a diagnosis is clinically valid, that record should definitely be reviewed. If a lay coder wonders, so will an auditor.
There are only two acceptable answers to a CV query: “yes, the diagnosis is clinically valid, and here is my support for the diagnosis” or “no, the condition was ruled out, and I am now removing it from the record.” But is there a way to be proactive, and avert a CV crisis?
The practice of “evolve, resolve, remove, recap,” converts an uncertain diagnosis to definitiveness. Once it has resolved, that should be evident in the documentation. A patient does not still have “acute hypoxic respiratory failure” on the day of discharge, or they wouldn’t be suitable for discharge. The diagnosis should then be removed from the A&P, but reappear in the discharge summary. If the uncertain diagnosis is ruled out, it should be removed from the daily list (or, at very least, be marked as “resolved”). There is no need to “rule out pneumonia” forever if no infiltrate develops and the empiric antibiotics are discontinued.
Copying and pasting should be reframed as copying and editing. Every act of copying forward should have a review and edit to ensure that it is accurate for today’s entry. My colleague Dr. Ronald Hirsch likes to say, “if you didn’t ask it, review it, examine it, or think it, it should not be in that day’s note.” Propagation of inaccurate information should be strongly discouraged. The problem list and A&P should be curated and updated. This is not just a CV situation; it is a patient safety and optimization of clinical care issue.
We might even try to get providers to do a documentation timeout. Before they document on the patient today, before they copy and paste everything from yesterday or use a template that might need extensive editing, they should stop for a moment and put “MENTATION” into their documentation. Are the diagnoses they are documenting today accurate and correct? Guide them to perform their own clinical validation in real time while having the watchful CDI eye perform concurrent CV review in the course of their daily activities.
But if something still slips through the cracks, there are predictable high-risk MS-DRGs, diagnoses, and services that might warrant second-level CDI review for retrospective CV queries. These represent the conditions and codes most vulnerable to denials. Each facility should determine which of these scenarios might be of highest yield for their institution:
- RAC/MAC/OIG targets;
- Cases with a single CC or MCC;
- Short-stay diagnoses with high severity of illness for principal or secondary diagnoses; or
- Services that have resulted in CV denials recently (e.g., sepsis, malnutrition, respiratory failure, AKI).
Creation of institution-specific clinical guidelines can be useful, but they shouldn’t be immutable. They should be regularly reviewed and updated with the most current guidelines and criteria. There must always be room for clinical judgment, with the caveat that the thought process must be explicitly laid out in the documentation. The medical staff must be aware of these guidelines and able to readily access them.
“A good practice is for the person performing clinical validation to ask themselves whether other providers would come to the same conclusion based on the same information. Is the diagnosis a reasonable conclusion based on the totality of the health record?” a 2016 AHIMA publication read.
If the answer is no, then either the diagnosis should be removed, or the documentation should be beefed up. In the interests of best clinical practices, all diagnoses should be clinically valid.