Association between elevated depressive symptoms and clinical ...

Association between Elevated Depressive Symptoms and Clinical Disease Severity in African-American Adults with Sickle Cell Disease Brian Laurence, DDS, PhD; David George, DDS; and Dexter Woods, DDS Baltimore, MD and Washington, DC Financial suppport: This study was partially supported by NIDCR/NIH grant DE14257, an Oral Health Disparities Center Grant. Objective: The purpose of this cross-sectional study was to examine the association between elevated depressive symptoms and the clinical severity of sickle cell disease (SCD) using African-American adults with and without SCD. Study Design: The population consisted of 102 African-American adults with SCD, diagnosed using hemoglobin electrophoresis, individually matched on age (±5 years), gender and recruitment location to 103 African-American adults without SCD (mean age of all subjects was 35.4 years, 55.6% female). Logistic regression was used to examine the association between SCD clinical severity and elevated depressive symptoms in bivariate and multivariable analyses. Results: The prevalence of elevated depressive symptoms as measured using the 10-item Center for Epidemiological Studies Depression Scale (CES-D) was 38.6% in those with SCD, compared to 27.5% in those without SCD; however, this difference was not statistically significant (p>0.05). Compared to African Amenrcans without SCD, Afrcan Americans with SCD were less educated, had lower incomes, and were more likely to be unemployed or disabled (p4, 39 of 40 depressed patients were correctly identified as having major depression.7 For this analysis, the standardized cutoff score of .4 was used to indicate the presence of elevated depressive symptoms. The two positively worded items on the 10-item version were reversed scored as in the validation study previously mentioned.7 Cronbach's alpha for the 10-item CES-D scale was 0.79 in the SCD subjects and 0.75 in the non-SCD subjects. Clinical severity was measured by utilizing two clinical indicators of severity of SCD.11'-3 First, patients were asked how many sickle cell crisis episodes requiring hospital admissions or emergency room visits occurred within the last 12 months. Second, use of hydroxyurea medication, which is prescribed to decrease the frequency of crisis episodes in patients with the most clinically severe forms of SCD, was included in the measure of severity by asking patients whether or not they were currently being treated with hydroxyurea. Severity of the disease was then classified into the following three groups: low severity = 0 hospital

admissions or emergency room visits for sickle cell crisis within the last 12 months and no current use of hydroxyurea; medium severity = one or two hospital admissions or emergency room visits for sickle cell crisis within the last 12 months and no current use of hydroxyurea; high severity = .3 hospital admissions or emergency room visits for sickle cell crisis within the last 12 months or current use of hydroxyurea. Subjects provided information on potential confounding variables, including educational attainment, current occupational status, yearly household income and current marital status, in addition to age and gender on which the SCD and non-SCD subjects were matched. Studies have suggested that women, those that are not married and those that have a poorer socioeconomic status are at risk for elevated depressive symptoms.'4'15 Educational level, income or occupational status are the most commonly used measures of socioeconomic position. 16,17

Statistical Analysis The Chi-squared test was used to detect sociodemographic differences between SCD subjects and non-SCD subjects. Odds ratios for the presence or absence of elevated depressive symptoms were calculated for the three SCD severity groups compared to subjects without SCD using unconditional logistic regression in bivariate and multivariable analyses. Sociodemographic variables that were associated with elevated depressive symptoms in bivariate analysis were included in the multivariable model. Statistical tests were two sided, 95% confidence intervals were reported and p values of