Background/Purpose: The BILD^1,2 was developed and validated as a measure of SLE organ damage for use in epidemiologic studies in which administration of the SLICC Damage Index (SDI) by physicians is not feasible. This analysis examines the criterion validity and other performance characteristics of the BILD in 4 racial/ethnic groups.

Methods: Data were from the California Lupus Epidemiology Study (CLUES), a population-based, multi-ethnic SLE cohort. Subjects participated in a research clinic visit during which the SDI was completed by a physician and completed a structured interview administered by a trained interviewer in which BILD was administered. Race and ethnicity were self-reported. Prevalence-adjusted bias-adjusted kappa (PABAK) coefficients were calculated to determine item-by-item agreement between physicians and subjects. Spearman correlation coefficients examined relationships between BILD and SDI scores. Construct validity was evaluated by examining relationship of sociodemographic, disease-specific, general health, and health care utilization factors with BILD scores. BILD scores are not normally distributed, so were divided into rough quartiles for these analyses and differences tested using analysis of variance or chi-square analysis. All analyses were performed separately for each of 4 racial/ethnic groups: white (W), Hispanic (H), African American (AA), and Asian (AS).

Results: The sample (n=281) was 29% W, 23% H, 11% AA, and 37% AS; 89% female; mean age 45 (±14) years; 22% with education ≤high school; 12% with poverty-level income; mean disease duration 16 (±10) years. 85% of interviews were completed in English. Correlations with SDI ranged from 0.45 for AA subjects to 0.80 for AS subjects (Table). All PABAK coefficients were >0.75, with the exception of 1 item for H subjects. For each racial/ethnic group, individuals with higher BILD scores had disease of longer duration; were less likely to be working; had poorer self-rated health, physical functioning and pain interference; reported more physician visits during the previous year; and were more likely to have been hospitalized in the previous year. No differences were seen by education or health literacy. Among H and AS groups, no differences were seen in BILD performance by language of administration.

Conclusion: The BILD functioned reasonably well in all 4 racial/ethnic groups, regardless of language of administration. Correspondence with physician-completed SDI was moderate to good; correlations were lower for AA and H patients where sample size was smaller. Future research will examine the source of discrepancies (patient or physician reports). Results provide further support the use of the BILD in observational and epidemiologic research, including among racial/ethnic minorities.

¹Yazdany, et al., Arthritis Care Res 2011; 63:1170

²Katz, et al., Arthritis Care Res 2014; 66:1057

« Back to 2017 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/performance-of-the-brief-index-of-lupus-damage-bild-in-a-multi-ethnic-population-based-systemic-lupus-erythematosus-sle-cohort/