Measures
ACS was defined by new chest radiograph infiltrate associated with fever, chest pain, increased work of breathing, or oxygen desaturation. Single ACS (Yes/No) was defined as only one admission for ACS throughout the lifetime. Recurrent ACS (Yes/No) was defined as at least 2 ACS incidents during the lifetime.Severe ACS was defined by admission to the pediatric intensive care or special care units.
Neighborhood deprivation was assessed with the 2015 Area Deprivation Index (ADI) for the State of Alabama. The ADI is an existing factor-based composite measure of socioeconomic deprivation, available in state-specific (1-10 scale) versions, with higher values indicating higher deprivation.18,19 The ADI is constructed from 17 variables in the domains of income, education, employment, and housing quality collected by the American Community Survey and aggregated to U.S. Census block groups. Neighborhood racial composition is not a component of the ADI. In this study, each patient was assigned a neighborhood ADI value according to the Census block group in which he or she resided based on the current address in their electronic health record. Because the ADI was not normally distributed, we dichotomized it as high (ADI 7-10) vs low (ADI 1-6) socioeconomic deprivation.
Neighborhood racial composition was calculated using the count of African American residents in each Census block group and converting it to a percentage of the total population count in that block group. The measure was not normally distributed and was treated as a categorical variable. As neighborhoods that were ≥90% African American were more prevalent in the sample than any other category, we dichotomized the measure as ≥90% vs <90% African American. Alternative categorization (tertiles and quartiles) produced similar results.
Rurality was assessed with the 2010 Rural-Urban Commuting Area (RUCA) codes, a classification that combines U.S. Census Bureau definitions with commuting information and ranks Census tracts on a scale of 1-10, with higher values indicating higher rurality.20 We dichotomized the measure as metro vs non-metro (RUCA scores 1-3 vs ≥4).
Covariates included sex (Male/Female), age, health insurance type (Public/Private), body mass index (BMI), chronic transfusions (Yes/No), hydroxyurea (Yes/No), asthma diagnosis (using ICD-10 code or active asthma treatment in the electronic medical records) (Yes/No), and SCD phenotype (SS, SB0, SB+, or SC), which was confirmed by hemoglobinopathy fractionation.