Predictors of vaso-occlusive crisis-related emergency department encounters in children with sickle cell disease in Western French Guiana

Abstract

Vaso-occlusive crises (VOCs) are a major cause of emergency department (ED) utilization in children with sickle cell disease (SCD). In Western French Guiana, recurrent VOC-related ED encounters may reflect both disease severity and barriers to longitudinal preventive care. We conducted a retrospective observational study using routine-care data from children with SCD who had at least one VOC-related ED encounter during a standardized 12-month observation period. The dataset contained 235 VOC-related ED encounters corresponding to 105 unique patient identifiers. To address repeated records, descriptive analyses are presented both at encounter level and at patient level. The primary multivariable analysis used unique patients as the unit of analysis; the outcome was the number of VOC-related ED encounters per unique patient identifier. Negative binomial regression was used for the count outcome. Recurrent ED use was defined as ≥ 4 VOC-related ED encounters per year. The 235 ED encounters corresponded to 105 unique patient identifiers. At the patient level, median age was 8 years [IQR: 6.5–11], n = 56 (53.3%) were female, and n = 80 (76.2%) had SS genotype. The median number of VOC-related ED encounters per unique patient was 2 [IQR: 1–3], and n = 19 (18.1%) had ≥ 4 VOC-related ED encounters/year. In the patient-level multivariable model, ferritin was not independently associated with ED encounter burden (adjusted incidence rate ratio [IRR]: 0.96; 95% confidence interval [CI]: 0.84–1.08; p = 0.474). Rural residence, irregular hydroxyurea observance, regular follow-up, asthma, and dry season showed associations with encounter burden; however, these findings should be interpreted cautiously because the study included only children with at least one ED encounter and did not include children with zero VOC-related ED visits. In this high-burden pediatric SCD cohort from Western French Guiana, no single biological or treatment variable reliably explained recurrent VOC-related ED utilization. The patient-level analysis no longer supports ferritin as an actionable independent predictor. Recurrent ED use likely reflects a combination of disease severity, rurality, healthcare access, and continuity-of-care factors. Prospective studies including the full followed SCD population, including children with zero ED visits, are needed. Not applicable.