(089) Clinical characteristics and healthcare resource utilization among patients identified with post-acute sequelae of COVID-19 (long-COVID) in US administrative claims data
Background: Methods to identify patients with lasting symptoms after an acute COVID-19 infection or long-COVID in claims data are emerging. The constellation of known long-COVID symptoms and the evolution of medical coding practices add to the challenges of studying long-COVID in real-world data. Efforts to understand the characteristics of patients diagnosed with long-COVID in RWD and how those characteristics differ by index diagnosis are necessary as more people experience the lasting effects of COVID-19.
Objectives: To describe demographics and clinical characteristics of individuals identified with long-COVID in a US administrative claims database.
Methods: A cohort of adult patients with long-COVID was identified in a US claims database (HealthVerity) from 01JAN2021 through 07JUL2022. Patients were eligible if they had evidence of long-COVID via ICD-10 diagnosis code of U09.9, Post COVID-19 condition, unspecified, or (introduced Oct 2021) 2) B94.8, Sequelae of other specified infectious and parasitic diseases (code utilized prior to release of U09.9) and a diagnosis code for COVID-19 (U07.1) in the 365 days prior (required continuous enrollment period). Baseline clinical characteristics and calendar time of diagnosis were reported. Incidence rates of long-COVID symptoms and clinical characteristics were described through the follow-up period. Analyses were stratified by index diagnosis code and age.
Results: Among 188,232 adult patients who index on long-COVID, the majority (82%) indexed on the U09.9 code. Patients who indexed on B94.8 were more frequently hospitalized on the date of long-COVID diagnosis than patients who index on U09.9 (18% and 9% respectively) and had a higher frequency of cardiovascular-related baseline comorbidities, including hypertension (53% and 46% respectively) and diabetes (30% and 21% respectively). The most frequent incident conditions in follow-up were dyspnea, fatigue/malaise, myalgia, sleep conditions, and chronic pulmonary disease with incidence rates ranging from 14,000-49,200 per 100,000 person-years. Incidence rates of clinical conditions did not differ meaningfully by index code.
Conclusions: This study identified coded symptoms of long-COVID consistent with literature, and did not find evidence to support restriction of analyses to a time frame after U09.9 became available. Follow-up incidence rates of known clinical sequelae of long-COVID aligned with those reported in literature and did not differ by index code, indicating the differences in frequency of baseline characteristics may be due to variation in healthcare resource utilization over calendar time of index rather than clinical differences.
