(099) Distinguishing Hospital Encounters as "with" vs "for" COVID-19 Using A Large US Administrative Claims Database

Background: Real-world studies defining COVID-19 hospitalization as any hospitalization with a COVID-19 diagnosis code or a positive SARS-CoV-2 test may misclassify some people with incidental findings of COVID-19 (“with COVID”) as being hospitalized “for COVID.”

Objectives: To determine the proportion of claims-based COVID-19 diagnoses in the hospital setting that are likely “with COVID” rather than “for COVID.”

Methods: People with an inpatient ICD-10-CM U07.1 COVID-19 code in California or Louisiana were identified from April 2020–October 2023 using HealthVerity administrative claims data. Hospitalizations “with COVID-19” were defined based on published literature and expert medical opinion as having both i) a U07.1 code and ii) at least one additional code in the same inpatient encounter for a condition or procedure that was likely to cause hospitalization and be unrelated to COVID-19, including unintentional injury, physical trauma, non-cardiopulmonary surgery, childbirth with < 2 days length of stay [LOS], psychiatric admission, alcohol & substance abuse, or poisoning & toxic exposure. Hospitalizations “for COVID-19” were defined as those with a U07.1 diagnosis that did not have an additional incidental code, or a U07.1 hospitalization where treatment used solely for COVID-19 (e.g., remdesivir) was identified regardless of accompanying diagnoses.

Results: We identified 357,643 COVID-19 hospitalizations and classified 108,363 (30% of cases) as hospitalization “with COVID-19.” Of which, psychiatric (49%) and physical trauma (30%) diagnoses were the most common, with unintentional injury (8%) and childbirth codes (1%) less commonly co-occurring among those identified as being hospitalized “with COVID.” The proportion of COVID-19 diagnoses that were incidental varied over time, ranging 28-33% from April 2020 through December 2021, 36-42% from January 2022 through May 2023, and then falling steadily to 22% by October 2023.

Conclusions: In this study, 70% of COVID hospitalizations were likely for COVID-19. Of the remaining 30%, half (15% overall) were likely incidental and most often associated with hospitalization for severe physical trauma or unintentional injury. Psychiatric admissions with a COVID-19 diagnosis accounted for the last 15% of COVID-19 hospital admissions. These admissions merit further investigation to distinguish between clinical conditions that could be “with COVID-19” (e.g., psychiatric admissions) or “for COVID-19” (e.g., altered mental state or confusion). Limiting the definition of COVID-19 hospitalization to those without claims for unintentional injury, physical trauma, poisoning, or childbirth may aid future real-world studies, although the proportion of incidental COVID-19 diagnoses may recently be declining.