Background/Purpose: Systemic Lupus Erythematosus (SLE) is an autoimmune disease characterized by a heterogenous clinical phenotype that may present differently over time and between patients, which can create challenges for identification and treatment of individuals being seen at multiple care centers. Clinical data research networks (CDRN) aim to pool electronic health records (EHR) to provide more complete clinical information for patients shared across care centers. We sought to determine whether algorithms to identify Systemic Lupus International Coordinating Clinics (SLICC) classification criteria attributes, using structured EHR data, could be applied to a CDRN to describe people with SLE.

Methods: The attribute identification frequency of SLICC criteria are represented in Table 1. We identified 6,488 persons ≥ 3 SLE diagnoses and 1,201,999 persons with no diagnosis codes for SLE. The following SLICC attributes were identified from each group and identified at the following rates (No diagnosis vs. ≥3 diagnoses): Oral Ulcers: 2.1% vs. 3.5%, Alopecia: 2.5% vs. 5%, Neurological: 18% vs. 23%, Arthritis: 2.5% vs. 6.8%, Serositis: 7.1% vs. 17%, Acute Cutaneous: 0.17% vs. 19%, Chronic Cutaneous: 0.1% vs. 18%, Renal: 10% vs. 29%, Thrombocytopenia: 6.3% vs. 14%, Leukopenia: 67% vs. 77%, Hemolytic Anemia: 1.4% vs. 4.8, Antinuclear Antibodies: 2.2% vs. 26%, Anti-dsDNA Antibodies: 1.6% vs. 42%, Anti-Sm Antibodies: 0.078% vs. 6%, Antiphospholipid Antibodies: 0.26 % vs. 7.1%, Low complement: 0.56% vs. 37%, Direct Coombs Test: 0.057% vs. 1.4%. The percent of patients satisfying the SLICC definition of SLE 0.81% among those without a SLE diagnosis and 43% among those with ≥3 SLE diagnoses.

Results: The attribute identification frequency of SLICC criteria are represented in Table 1. We identified 6,488 persons ≥ 3 SLE diagnoses and 1,201,999 persons with no diagnosis codes for SLE. The following SLICC attributes were identified from each group and identified at the following rates (No diagnosis vs. ≥3 diagnoses): Oral Ulcers: 2.1% vs. 3.5%, Alopecia: 2.5% vs. 5%, Neurological: 18% vs. 23%, Arthritis: 2.5% vs. 6.8%, Serositis: 7.1% vs. 17%, Acute Cutaneous: 0.17% vs. 19%, Chronic Cutaneous: 0.1% vs. 18%, Renal: 10% vs. 29%, Thrombocytopenia: 6.3% vs. 14%, Leukopenia: 67% vs. 77%, Hemolytic Anemia: 1.4% vs. 4.8, Antinuclear Antibodies: 2.2% vs. 26%, Anti-dsDNA Antibodies: 1.6% vs. 42%, Anti-Sm Antibodies: 0.078% vs. 6%, Antiphospholipid Antibodies: 0.26 % vs. 7.1%, Low complement: 0.56% vs. 37%, Direct Coombs Test: 0.057% vs. 1.4%. The number of patients satisfying the SLICC definition of “Definite SLE” was 7684 persons (0.64%) among those without a SLE diagnosis and 2487 persons (38%) among those with ≥3 SLE diagnoses.

Conclusion: The results demonstrate that we can identify all SLICC classification criteria attributes in the CAPriCORN data set. We observed an increased rate of attribute identification for all SLICC criteria and an increased rate of definite SLE classification via SLICC in patients with ≥3 SLE diagnoses when compared with patients without any SLE diagnoses, consistent with an expected higher occurrence rate for persons with SLE. This suggests that SLE presentation can be characterized in CDRN data.

« Back to ACR Convergence 2021

ACR Meeting Abstracts - https://acrabstracts.org/abstract/utilization-of-a-clinical-data-research-network-to-assess-systemic-lupus-international-collaborating-clinics-classification-criteria-attributes-in-patients-with-systemic-lupus-erythematosus/