Race-Based Differences in MCI And Dementia: A Propensity Score Matching Study

Introduction Mild cognitive impairment (MCI) and dementia are characterized by objective deficits in several cognitive domains. These cognitive disorders are prevalent in the fast-growing older population worldwide. There are several modifiable and non-modifiable risk factors for these cognitive disorders. Research has shown that the African American race is predictive of incident dementia. However, the African-American race is confounded by other risk factors which may explain much of the racial disparity in cognitive disorders, most notably education, socio-economic status and lack of access. The aim of this study was to compare the prevalence of MCI and Dementia between African-American and Caucasian community-dwelling older Veterans after matching of baseline characteristics. Methods We conducted a cross-sectional study using propensity score matching (PSM) among community-dwelling Veterans aged ≥50 years enrolled in VA primary care clinics from July 1, 2019 to May 31, 2020. Participants with baseline dementia diagnosis were excluded. Patients received mailed questionnaires including sociodemographic, information about exercise, education and an assessment of cognitive status using the validated Self-Administered Gerocognitive Examination (SAGE). We complemented the information with data from the electronic health records (EHR). To assess frailty, we used a 31-item VA Frailty Index (VA-FI) generated from claims-based data matched to the study date. The VA-FI categorizes patients into robust (<0.10), pre-frail (0.10-0.20) and frail (≥0.21). Geographical socio-economic conditions were determined by using the US Area Deprivation Index (ADI) datasets. The ADI score distribution was divided by tertiles (Low, Middle, High), with higher scores corresponding to more socio-economic deprivation. African American Veterans were matched with Caucasians using PSM with one-to-one nearest neighbor matching without replacement. Matching covariates used to calculate the propensity score included age, gender, marital status, BMI, ethnicity, years of education, frailty (VA-FI), Obstructive Sleep Apnea (OSA), and ADI with a tolerance level of .01. Using a Chi-Square test, we compared the proportion of patients with MCI and Dementia among Caucasians and African-Americans after matching for baseline characteristics. Results We obtained a response rate of 19.75% (n =1073) out of 5,432 mailed surveys. After propensity score matching of those Veterans who responded, 202 Caucasians and 202 African-Americans were selected and compared. Participants had a mean age of 68.57(SD=7.9) years, 50.2% were married, 90.6% non-Hispanic and 96% male. Almost all patients had achieved at least 10 years of education (91.8%), 39.9% were obese and 34.9% had OSA. The mean ADI score was 98.29 (SD=20.16) and 37.9% were found to be in the high ADI tertile. Regarding frailty status, 23.5%, 36.1%, and 40.3% were robust, pre-frail and frail respectively. After matching, all the baseline clinical characteristics were comparable between the two groups. Using a Chi-Square test after PSM, we found that the proportions of MCI (43.8%) and Dementia (40.3%) in Caucasian patients were not significantly different from that of African-Americans (MCI=56.3%, Dementia=59.7%), p=.087. Conclusions This study results suggest that race alone is not independently associated with MCI or Dementia. Even though some studies have shown that African-Americans are more likely to develop cognitive impairment, this association may be due to other socio-demographic factors that are more prevalent in the African-American population. Funding Bruce W. Carter VA Medical Center - Geriatric Research, Education and Clinical Center (GRECC)