Background: While literature is relatively abundant on systematic erythematosus lupus (SLE) and lupus nephritis (LN), less evidence is available on specific patient phenotypes with a refractory form of SLE/LN and it mainly comes from case reports and case series with varying definitions. Generic datasets like insurance claims and electronic medical records can generate real-world evidence on this particular population as they offer large patient populations and a relatively good longitudinal track of patients. However, as they are not disease-specific, they often lack disease-specific variables as well as confirmed diagnoses or information on disease severity. Robust patient identification strategies are therefore needed to detect the targeted patient population.
Objectives: To develop and indirectly validate a patient identification algorithm for refractory lupus using US insurance claims (Optum).
Methods: The algorithm was built by combining diagnostic codes (International Classification of Diseases ICD-10 M32.1x, M32.8 and M32.9 for SLE), treatment episodes and rheumatologist visits. Refractory lupus was defined as adult patients who received glucocorticoids (GC) and who were exposed to at least 2 sequential treatment regimens (incl. immunosuppressants, biologics and calcineurin inhibitors), 3 months each during the post index period. Study period was 01 January 2016 – 31 March 2023. Index date was the first SLE code recorded in the identification period. To indirectly validate our algorithm, a complete description of the patient population obtained by applying the selected algorithm was performed and their key characteristics (age distribution, gender ratio, comorbidities) were compared against trial populations and published literature.
Results: Among the 30,969 adult patients with a confirmed diagnosis of SLE/LN, 85.2% were ever exposed to GC during the follow-up period. A total of 2,152 (8.2%) were identified as refractory. Focusing on those with 1 year pre-index and 2 years post-index data, mean age was 52.6 (SD 14.6) years and 88.7% were women. African American, Hispanic and Asian ethnicity was reported for 18.6, 13.9 and 5.5% respectively. A high proportion (73.7%) reported organ damage in the post-index period. Several comorbidities were frequently observed: anemia (24.6%), depression (15.0%), dyslipidemia (26.6%), hypertension (44.8%), obesity (14.5%), osteoporosis (11.4%) and thyroid disease (26.3%).
Conclusion: The patient identification algorithm using US claims was able to distinguish a subset of patients with refractory lupus. Organ damage was very high and a considerable burden of comorbidities was also observed in this particular population.
