Matthew Harris and Evan Thacker, Public Health
Introduction:
5.4 million people in the United States are classified as cognitively impaired, with 11.7% progressing to dementia annually (Plassman 2008). Alzheimer’s Disease has risen to the 6th leading cause of death in the United States, an 89% increase from the year 2000 (Alzheimer’s Association 2017). Cognitive decline, and differences in cognitive decline across racial or ethnic groups, are also related to life course social determinants of health in the United States, such as socioeconomic position, financial resources, occupational and educational opportunities, discrimination, segregation, migration, and group resources (Glymour & Manly 2008). However, one social determinant of health that has not been investigated much is urban vs rural living. As such there may be a disparity in the occurrence of cognitive impairment or dementia across rural vs urban settings, but prior research has shown mixed results. We are aware of three epidemiologic cohort studies that have assessed rural vs urban living in relation to cognitive impairment or dementia. Two studies conducted in Israel (Merims 2015) and Spain (Contador 2015) found that rural residence was associated with a higher prevalence of dementia compared to urban residence, while a study conducted in Canada (St John 2016), noted no difference between rural/urban residence and dementia outcome. However, no good study of the same has been conducted in the United States. As such the purpose of our study is to address this knowledge gap surrounding the implications of rural/urban residence on incident cognitive impairment (ICI), especially given our historical context of racism, segregation, discrimination, and other social determinants of cognitive aging, which vary geographically. To address this gap, we have used data from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study to determine whether there is a rural/urban disparity in cognitive impairment or cognitive decline in the United States as well as analyzed some of the reasons for this rural/urban disparity.
Methods:
REGARDS is a cohort study of 30,239 adults aged 45+ in the 48 contiguous states that contains both rural and urban residents. We analyzed 20,592 participants who at baseline (2003-2007) were cognitively intact with no history of stroke and had cognition assessed an average of 7.1 years later. We used Rural-Urban Commuting Area (RUCA) codes to classify participants as urban (n = 16,436), large city/town (n = 2,420), or small/isolated rural (n = 1,736) at baseline. We defined ICI as falling ≥1.5 SD below the mean on at least two of three cognitive tests administered during follow-up: word list learning, word list delayed recall, and animal naming. Using urban as the referent, we estimated odds ratios of ICI for rural and for large city/town via logistic regression analysis while accounting for confounding, mediation, and effect heterogeneity. Potential confounders that we considered included age, sex, race, and census region. Potential mediators included social factors, clinical factors, and incident stroke (stroke occurring after baseline but before follow-up cognitive assessment). Effect heterogeneity was determined by rural/urban status on the associations of race, sex, income, education, and census region with ICI.
Results:
ICI occurred in 1,291 participants (6.3%). Rural residents had 49% higher odds of ICI adjusted for confounding by demographics (Model 2 in Table, OR = 1.49 [95% CI: 1.19, 1.85]). After further adjusting for potential mediators (Models 3-6), odds of ICI remained 25% higher for rural vs urban (Model 6, OR = 1.25 [0.99, 1.56]). In assessing effect heterogeneity, we found synergism of rural dwelling with black race, physical inactivity, and low self-rated health (all P < 0.1; see ORs in Table), but not for other ICI risk factors. We found no difference in ICI for large city/town vs urban (demographics-adjusted OR = 1.08 [0.88, 1.33]; fully adjusted OR = 0.95 [0.77, 1.18]), and no effect heterogeneity of ICI risk factors by large city/town (all P > 0.2).
Discussion:
Our results show that rural residents are almost twice as likely to develop ICI than their urban counterpart, which is partially explained by potential mediators. Furthermore, odds of ICI were highest for rural residents combined with black race, physical inactivity, or low self-rated health. This is a substantial discovery because no study with this depth or scale has been done in the United States. As such, we expect that it will be of great value to other researchers who seek to lessen the impact that ICI has on our communities. Further research could be a longitudinal approach to ICI to better understand when ICI occurs over time.
Conclusion:
Rural living is an important social determinant of cognitive health in the United States.