Emergency Department Leave Against Medical Advice Among Adults with Sickle Cell Disease: An Equity-Sensitive Indicator of System Performance
Kenneth A Rivlin, Anna Zhilkova, Jenai Jackson, Antonia EyssallenneAbstract
Background
Prior work has demonstrated that hospitalizations for sickle cell disease (SCD) are disproportionately associated with leave against medical advice (LAMA) compared with other adult hospitalizations. The magnitude of variation observed across hospitals suggests that this outcome reflects system-level failures leading to unmet care needs rather than people not following physician directives or advice. However, prior analyses focused on inpatient settings and did not examine the emergency department (ED). This early phase of acute SCD care is highly dependent on timely and responsive system processes. In quality improvement frameworks, unwarranted variation in outcomes across sites caring for similar populations is interpreted as a signal of differences in system performance rather than individual behavior. In equity-relevant conditions such as SCD, these outcomes may serve as equity-sensitive indicators of care quality. This study aimed to characterize inpatient and ED LAMA among adults with SCD across New York City hospitals in 2022-2024 and examine hospital-level variation as a signal of system performance.
Methods
We conducted a descriptive analysis of 2022-2024 inpatient and ED treat-and-release encounters for adults with SCD across New York City hospitals using Statewide Planning and Research Cooperative System (SPARCS). SPARCS is an all-payer administrative data system provided by the New York State Department of Health. LAMA was identified based on patient disposition status of “left against medical advice” or code ‘07’. Hospital-level rates for SCD were compared with rates among all adult patients. Hospitals with fewer than 50 SCD encounters during the three-year period in each setting were excluded to ensure stability of estimates. Gender-stratified analyses were not performed because of the study’s focus on hospital-system performance. Analyses emphasized the magnitude and range of variation across hospitals.
Results
Across New York City hospitals in 2022-2024, inpatient LAMA occurred in approximately 14% of hospitalizations for adults with SCD, compared with approximately 4% among all adult hospitalizations, with hospital-level SCD rates ranging from under 10% to 50%. ED LAMA occurred in approximately 7% of ED encounters for adults with SCD, compared with approximately 3% among all adult ED visits, with SCD rates ranging from under 5% to over 20% (Figure 1). Patterns of hospital-level variation were similar for ED and inpatient LAMA, with hospitals that demonstrated higher rates in one setting often showed higher rates in the other. Marked hospital-level variation was observed for both outcomes, including within a single integrated public health system, where SCD inpatient LAMA varied by more than fivefold, and ED LAMA showed substantial heterogeneity. Similar patterns of variability were observed across academic and community hospitals citywide, including academic centers with comprehensive sickle cell programs, indicating that the presence of specialized services alone does not ensure consistent performance in acute care delivery. These findings highlight equity-relevant differences in care delivery that are not explained solely by diagnosis.
Conclusions
Both inpatient and ED LAMA are common among adults with sickle cell disease and exhibit substantial unwarranted variation across hospitals, including within the same health system. The similarity in hospital-level patterns across ED and inpatient settings suggests linked but distinct failure points along the acute care pathway, rather than isolated events confined to a single care setting. Given their disproportionate burden in a population historically marginalized in health care and their sensitivity to local care-delivery structures, these outcomes serve as equity-sensitive indicators of system performance. In other domains of patient safety, variation of this magnitude would routinely prompt focused investigation and quality improvement; applying comparable standards to equity-relevant outcomes in sickle cell disease is both warranted and necessary. Incorporating LAMA metrics into quality dashboards may enable health systems to identify inconsistent acute care delivery, prioritize targeted interventions across ED and inpatient settings, and advance equitable, high-quality care for adults with sickle cell disease. Finally, we believe these findings likely underestimate the true proportion of early ED departures because administrative data may not capture patients who eloped or left without being seen (LWBS). Future studies can utilize EHRs or other data sources to better understand the true burden and magnitude of disparities.