The confusion about when to code severe malnutrition is likely to be back in the spotlight as malnutrition screening becomes a mandatory electronic quality measure creating a contradictory message from The Centers for Medicare & Medicaid Services (CMS) as hospitals continue to have this diagnosis removed from their claims. Becker’s Hospital Review reported that CMS will require hospitals to report the Malnutrition Care Score beginning in payment year 2028, finalized in the FY2027 Inpatient Prospective Payment System rule.
This is the first nutrition-focused electronic clinical quality measure CMS has mandated in the Hospital Inpatient Quality Reporting (IQR) Program.
The Malnutrition Care Score began life in 2013 as the Global Malnutrition Composite Score, developed by the Academy of Nutrition and Dietetics with Avalere Health from four component measures: risk screening, a nutrition assessment for at-risk patients, a documented diagnosis when malnutrition is present, and a care plan to treat it. CMS adopted it as a voluntary IQR measure for the CY2024 reporting period. It is now becoming a mandatory measure, layered onto a complex, aging population where, by the CMS estimate, malnutrition is diagnosed in fewer than one in ten patients who are likely to have it. The Academy of Nutrition and Dietetics called mandatory reporting a milestone, quoting Academy President Tamara Randall’s view that the measure “will help make identifying and addressing malnutrition a consistent part of high-quality hospital care.”
Why is CMS implementing a measure that will likely increase the reporting of malnutrition when its own auditors have spent more than a decade claiming severe malnutrition is overreported within claims data? To be clear, those are two different assertions, not a contradiction. Malnutrition overall appears to be underdiagnosed relative to its true prevalence, while the specific severe malnutrition codes reviewed by the U.S. Department of Health and Human Services (HHS) Office of Inspector General (OIG) reviewed were often assigned more aggressively than the documentation supported.
The OIG began by examining 2,145 inpatient claims across 25 hospitals where Kwashiorkor was reported. They found it was incorrectly reported on all but one claim. The result was an error rate above 99 percent, costing Medicare roughly $102 million between 2006 and 2015. Their perspective was that Kwashiorkor is exceptionally rare in Medicare patients and, therefore, was not substantiated on these claims. Although that assessment was likely accurate, much of the reporting of Kwashiorkor was related to how it was indexed within coding guidelines, not an attempt to defraud the federal government.
Fast forward to 2020 when the OIG reviewed severe malnutrition claims, which did not have the same indexing issues as Kwashiorkor. They found errors in 164 of 200 sampled claims. These findings were extrapolated to more than $1 billion in overpayments for fiscal years 2016 and 2017. Hospitals pushed back arguing that the OIG reviewers had not applied “nationally accepted evidence-based standards” for diagnosing malnutrition, but OIG stood firm. Severe malnutrition has been a PEPPER outlier target ever since.
CMS is now requiring every hospital to screen, assess, diagnose, and treat malnutrition as a mandatory quality measure. Fulfilling these requirements is likely to result in malnutrition diagnoses being included in more claims because the condition will meet the Uniform Hospital Discharge Data Set (UHDDS) reporting guidelines due to being evaluated and, perhaps, treated. But the same issues identified in the last OIG review likely persist.
One of those issues is that code assignment should not be driven by whether a patient meets the ASPEN criteria, the newer GLIM consensus criteria, or any numeric threshold. Code assignment is driven by the provider’s documentation of malnutrition when it meets reporting requirements regardless of the criteria used to arrive at the diagnosis.
Or, more precisely, the criteria the registered dietitian used to arrive at the diagnosis. Hospitals who rely upon ASPEN or GLIM will need to reconcile these criteria with the screening tool used to meet this requirement. CMS did not mandate a specific screening tool instead recommend options such as MST, MUST, MNA-SF, or NRS-2002.
The diagnosis of malnutrition cannot be coded from the dietician’s documentation, but can it be used as supporting evidence within a query if a process is not already in place for the provider to validate the dietician’s findings. Section I.B.14 of the ICD-10-CM Official Guidelines, Documentation by Clinicians Other than the Patient’s Provider, lets a coder pull data like BMI from a dietitian’s note, but Coding Clinic clarified as far back as Fourth Quarter 2008 that a BMI value alone cannot be coded unless the provider separately documents a nutritionally related diagnosis, such as malnutrition, underweight, or obesity. A dietitian can screen, assess, and apply GLIM or ASPEN criteria all day but it does not become a reportable code until a provider puts the diagnosis in the chart.
What the OIG has consistently wanted to see, and what a defensible severe malnutrition diagnosis still requires, is a treatment plan that matches the reported severity. Although treatment is not required to report malnutrition, severe malnutrition is something that a provider cannot ignore because it requires treatment.
The problem is there is little agreement as to what is reasonable treatment. Some argue that it must be more than supplement drinks. Reasonable treatment for moderate malnutrition typically means a documented nutrition care plan, oral nutrition supplements, and a calorie and protein-modified diet, monitored by a registered dietitian throughout the stay.
Severe malnutrition may include enteral tube feeding or parenteral nutrition when oral intake cannot meet estimated needs, and, critically, monitoring for refeeding syndrome consistent with the ASPEN 2020 consensus recommendations whenever nutrition is reintroduced to a significantly depleted patient. However, the documentation must clearly support the reason for a feeding tube or other strategies is to treat malnutrition, not to prevent malnutrition in a patient who is experiencing dysphagia or has a risk of aspiration.
Just like intubation is not always used to treat acute respiratory failure, the placement of a feeding tube is not always to treat severe malnutrition. A chart documenting severe malnutrition with no or minimal dietitian involvement, no escalation of nutritional support, and no electrolyte monitoring is exactly the pattern that the OIG has flagged for fifteen years, and payers have learned it too.
So will the Malnutrition Care Score spike Clinical documentation integrity (CDI) and coding query volume? I would be surprised if it did not. Mandatory risk screening on every eligible admission means far more patients will get a dietitian assessment than currently happens, and every assessment confirming malnutrition without a matching diagnosis already in the provider’s own documentation is a query waiting to be written. Malnutrition will likely become the highest-volume clinical validation query categories and an even bigger target for auditors.
CDI leaders should not wait for the FY2028 payment determination to find out how this collides. Standardize now on one clinical criterion set, ASPEN or GLIM, rather than leaving providers, dietitians, and coders to reconcile three definitions mid-query.
Train providers now on the specific, quantifiable distinction between moderate and severe, before an auditor teaches that lesson retroactively with a demand letter. And treat the nutrition care plan as part of the diagnosis, not paperwork that follows it, because a mismatch between claimed severity and actual treatment is the fact pattern that has decided nearly every malnutrition audit CMS has ever won.
CMS has spent more than a decade telling hospitals it doesn’t trust how they code malnutrition, and it’s about to spend the next several years requiring every one of them to look harder for it anyway.
I don’t know which message CMS actually believes. I’m not sure CDI can afford to wait until FY2028 to find out.
