Alzheimer’s Disease and Related Dementias Referral Outcomes among Diverse Rural Residents

Abstract

AbstractBackgroundThis study to increase dementia detection in rural settings was conducted in a medically underserved area of southcentral Florida referred to as “the Glades.” Residents face multiple disparities related to Alzheimer’s disease and related dementias, including less than a 60% high school graduation rate, diabetes and hypertension as the primary health diagnosis, and increased exposure to ambient air particles from sugar cane burning.MethodLocations for testing included eight churches and a virtual “location,” for a total of ten sites. Individuals (N = 217) were cluster sampled within these sites. Participants were administered a brief educational intervention and brain health was screened using Nasreddine’s telephone “Mini‐MoCA” (2019), in which a score < 12 indicates cognitive risk. Referrals were determined for 133 participants.ResultsThe sample consisted of 59 virtual and 74 in‐person participants, with a range of 1 to 14 participants per church. Seventy percent were female, 57% lived alone, 13% had a previous memory test, and less than 1% had been previously diagnosed with dementia. The sample consisted primarily of four ethnicities, n = 44 (33%) African Americans, n = 21 (16%) Afro‐Caribbeans, n = 19 (14%) Hispanics, and n = 45 (34%) Whites.There was substantial variance in outcomes across sites, with an ICC = .179 (n = 133), indicating that 17.9% of the variance in referrals was between sites. Therefore, a mixed linear model approach was used for analysis with participants nested within sites (Feaster et al., 2011) The covariate of whether they had a previous memory test was not statistically significant and only reduced the ICC by .002. Neither gender, nor whether the person lived alone were statistically significant predictors of referrals, (p > .05), however adding ethnic group reduced the ICC to .102, a 47% reduction in intercept variance (N = 131). In particular, African American participants were significantly more likely to have referrals than White participants, g20 = 1.29, p<.001, OR = 3.66. Hispanic and Afro‐Caribbean participants were not significantly more likely to have referrals than White participants, p = .52 and p = .06 respectively.ConclusionFurther investigation is needed to determine place‐based factors that uniquely increase risk for African Americans in an 89% racially/ethnically diverse community.