Pulse dose steroid experience among hospitalized patients with systemic lupus erythematosus: a single-center feasibility study

Abstract

Introduction/objectives: Pulse intravenous (IV) methylprednisolone (MEP) is often used for severe SLE manifestations requiring hospitalization. However, the accuracy of pulse dose documentation extracted from the electronic health record (EHR) is unknown. We assessed the feasibility to study pulse steroid dosing among hospitalized patients with SLE at our institution. Method: Using the Stanford Medicine Research Data Repository (STARR) extracted from the EHR, we identified patients with ≥ 1 SLE ICD code before/during hospitalization receiving steroids (1/2008-12/2017). SLE diagnosis required rheumatologist confirmation. For our feasibility study, we randomly sampled 40/747 patients meeting search criteria. Pulse IV MEP was defined as ≥ 200 mg. Pharmacy dispensation data required EHR confirmation. Results: Forty adult and pediatric subjects were identified, passing initial criteria screen; 6 pediatric patients were excluded as EHR pharmacy confirmation was unavailable. Of the 34 adults, 14 had SLE confirmed. Among 5 adult SLE patients with pulse documentation, 3 occurred while hospitalized, for the following indications: acute renal transplant rejection (2 patients, 2 hospitalizations) and lupus flare (1 patient, 2 hospitalizations). No discrepancies were observed in pharmacy dispensation documentation of pulse dosing between EHR and STARR for all 4 hospitalizations. Conclusions: Assessment of pulse steroid dose dispensation among hospitalized patients with SLE can be reliably ascertained from the extracted portion of the EHR designed for research. Reliance on a single ICD code for SLE in the EHR may lead to high rate of false-positive diagnoses of SLE among hospitalized patients. We document the importance of supplementing one ICD code with additional clinical information when confirming SLE diagnosis.

Key Points • Assessment of pulse steroid dosing dispensation among hospitalized patients with SLE can be reliably determined from the extracted portion of the EHR designed for research purposes. • Reliance on a single ICD code contributes to a high rate of false positive diagnoses of SLE among hospitalized patients. • Supplementing ICD coding with additional clinical information is vital when confirming SLE diagnosis.