Association of social determinants of health (SDoH) with clinic visit adherence, healthcare utilization and hydroxyurea (HU) acceptance and adherence in children with sickle cell disease (SCD)

Abstract Background: Children with chronic diseases living in deprived neighborhoods have lower medication adherence and poorer health outcomes. The knowledge of these associations in pediatric SCD is limited. To address this gap, we measured the association between area-level SDoH and adherence to clinic visits, healthcare utilization, and HU acceptance and adherence in patients receiving care at the University of Alabama at Birmingham (UAB) pediatric SCD clinics. Methods: We performed an IRB-approved retrospective cohort study of patients aged 2-19 years with SCD at UAB (Birmingham) or its satellite clinics (Montgomery, Tuscaloosa, and Opelika) from January-December 2024. Age during the study period, sex, race, SCD genotype, and residential address were collected from medical records. Addresses were geocoded and linked to the American Community Survey 5-year estimates (2019-2023) to develop sample-specific measures of Social Vulnerability Index (SVI) (overall score and four thematic sub-scores: socioeconomic status [SES], housing and transportation, minority status and language, and household composition and disability), and Low-income Low-access (LILA) measure of food access. The outcomes included adherence to clinic visits (≥1 clinic visit in the last 12 months), healthcare utilization (emergency department [ED] visit or hospitalization in the last 12 months), HU acceptance (parental HU agreement), and HU adherence (most recent fetal hemoglobin [HbF] level). The unadjusted associations between area-level SDoH and adherence to clinic visits, HU acceptance and healthcare utilization were examined with logistic regression, and between SDoH and HU adherence (among those taking HU) with multivariate log linear regression and adjusting for covariates (age, sex, and rurality). We also performed subgroup analysis among patients with sickle cell anemia (SCA: HbSS and HbSβo thalassemia). Results: Among 779 patients with SCD (median age 11 years; 53% with SCA), 605 (78%) were adherent to clinic visits, 313 (40%) had an ED visit, and 208 (27%) had a hospitalization. Among 415 patients with SCA, 304 (73%) accepted HU. Clinic adherence: Patients attending satellite clinics had higher clinic adherence than those attending UAB clinic (82% vs. 76%, P=0.042). In the full SCD cohort, there was no difference in clinic adherence by overall SVI or its thematic sub-scores, or by food access. In the subgroup analysis of patients with SCA, there was no difference in clinic adherence by overall SVI or its thematic sub-scores. However, a greater proportion of patients with SCA living in LILA neighborhoods were nonadherent to clinic visits than those in a non-LILA neighborhood (53% vs. 39%, P=0.05). Healthcare utilization: In the full SCD cohort, there was no difference in ED visit or hospitalization by overall SVI or thematic sub-scores, or with the subgroup analysis of patients with SCA. However, in the full SCD cohort, patients residing in a LILA neighborhood were more likely to require an ED visit (45% vs. 34%, P=0.013) and hospitalization (46% vs. 36%, P=0.028). In subgroup analyses, these associations were primarily limited to patients with SCA (48% vs. 35%, P=0.013 for ED; 48% vs. 37%, P=0.039 for hospitalization). HU acceptance and adherence: While HU acceptance was not associated with SVI or LILA, HU adherence was associated with several SDoH measures. Higher vulnerability on the SVI socioeconomic status sub-score and minority status sub-score were associated with decreased adherence. Among those with SCA and on HU, a 10% increase in SVI socioeconomic vulnerability was associated with a 4.14% decrease in HbF (P=0.022), whereas a 10% increase in minority vulnerability was associated with a 5.62% decrease in HbF (P=0.001). Adjusting for covariates and SVI socioeconomic status, a 10% increase in SVI minority vulnerability was associated with a 4.30% decrease in HbF level (P=0.049). SVI SES was not significant after adjusting for covariates and SVI minority vulnerability. Conclusion: Our results highlight that area-level socioeconomic vulnerability is associated with lower adherence to clinic visits, greater emergency healthcare utilization, and lower HbF levels among children with SCA. Future studies will explore mechanistic pathways of these associations so that targeted interventions could be designed and implemented to reduce social barriers and improve outcomes for this vulnerable population.