Assessing the Validity of the Centers for Medicare & Medicaid Services Measure in Identifying Potentially Preventable Emergency Department Visits by Patients With Cancer.

ABSTRACT

PURPOSE: The Centers for Medicare & Medicaid Services (CMS) implemented chemotherapy measures (OP-35) to reduce potentially preventable emergency department visits (PPEDVs) and hospitalizations. This study evaluated the validity of the OP-35 measure in identifying PPEDVs among patients with cancer. METHODS: This is a cross-sectional study, which used data from the 2012-2022 National Hospital Ambulatory Medical Care Survey. ED visits are assessed and compared on the basis of three measures: immediacy using Emergency Severity Index (ESI), disposition (discharge v hospitalization), and OP-35 criteria. RESULTS: Between 2012 and 2022, a weighted sample of 46,723,524 ED visits were made by patients with cancer. Among reported ESI cases, 25.2% (8,346,443) was high urgency. In addition, 30.3% (14,135,496) of ED visits among patients with cancer led to hospitalizations. Using the OP-35 measure, it was found that 20.85% (9,743,977) was PPEDVs. A 21.9% (10,232,102) discrepancy between discharge diagnosis (CMS billing codes) and chief complaints was identified. Further analysis showed that 19.2% (1,872,556) of potentially preventable ED visits (CMS OP-35) were high urgency and 32.6% (3,181,280) resulted in hospitalization. CONCLUSION: The CMS approach to identifying PPEDVs has limitations. First, it may overcount preventable visits by including high-urgency or hospitalization-requiring cases. Second, relying on final diagnoses for retrospective preventability judgment can be misleading as they may not reflect the initial reason for the visit. In addition, differentiating causes for ED visits in patients with cancer undergoing various treatments is challenging as the approach does not distinguish between chemotherapy-related complications and others. Identification inconsistencies arise because of varying coding practices and chosen preventable conditions, lacking consensus and alignment with specific hospital or patient needs. Finally, the model fails to consider crucial nonclinical factors like social support, economic barriers, and alternative care access, potentially unfairly penalizing hospitals serving underserved populations.