Inpatient study: ICD-10-CM coding impacts MS-DRG assignment and LOS for malnourished patients

An inpatient study recently published in the Journal of the Academy of Nutrition and Dietetics demonstrates the importance of accurate ICD-10-CM reporting for malnutrition to ensure accurate Medicare severity diagnosis-related group (MS-DRG) assignment and the establishment of appropriate comparison benchmarks such as expected geometric mean length of stay (GMLOS).

Under the MS-DRG system, CMS classifies illnesses by severity level (i.e., a complication or comorbidity (CC), a major complication or comorbidity (MCC), or non-CC) based on secondary diagnoses and procedures as documented using ICD-10-CM codes. A relative weight (RW) is assigned to each MS-DRG that reflects the average cost to provide care for inpatients with a specific diagnosis relative to the average cost to provide care for all Medicare patients.

The GMLOS score, which is largely influenced by assigned diagnosis codes, is one of the components that Medicare considers when determining RWs reimbursement for MS-DRGs. Because hospitals do not typically receive additional reimbursement for hospital stays that exceed the expected length of stay, the goal is to discharge patients before they exceed this calculated average.

To examine the effect of malnutrition coding on expected GMLOS, researchers collected clinical data from a community hospital where patients are screened for nutritional risk within 24 hours of admission. Reports on malnourished patients are generated monthly by the hospital’s financial analysts and include admission and discharge dates, assigned MS-DRGs, CCs or MCCs, and their associated ICD-10-CM codes.  

Researchers analyzed the hospital’s monthly reports for malnourished patients admitted between March 2015 and June 2017. During this time, registered dietician nutritionists identified 1,817 records for malnourished adult patients. Of these patients, 1,171 (64.4%) of them were not coded for malnutrition.

The Medicare expected GMLOS based on assigned MS-DRGs was 3.5 days. However, the average length of stay for this group was 5.3 days. Researchers estimate that if malnutrition had been coded appropriately, the expected GMLOS would have been 5.2 days, which is a much closer to the actual average length of stay for this subset of patients.

According to the researchers’ findings, of the malnourished patients who were not coded for malnutrition, 475 or 40.6% of them would have benefited from proper coding to change the MS-DRG, RW, and expected GMLOS.