Table 2.
Studies which examined the association between community and cognitive function
| Reference | Setting | Method | Participants | Age | Measures | Community definition | Community measurements | Individual controlsa | Mediator/modifier | Statistical analysis | Resultsd |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Boardman et al. [23] |
US, Chicago, Chicago health and ageing project |
Cross-sectional (1993) | 1,665 people living in Chicago | 65+ | MMSE, memory, perceptual speed | Boundaries created by the local authority |
Compositional: Neighborhood disadvantage, social disorders |
D, education | Modifier: disadvantage genotype | Multilevel analysis | Inter (APOE genotype) |
| Clarke et al. [24] | US, city of Chicago, Chicago community adult health study | Cross-sectional (2002) | 949 adults living in Chicago city | 50+ | TICS | 2000 US census tracts |
Compositional: Neighborhood disadvantage, neighborhood affluence Contextual: Neighborhood resources: recreational centers, institutions; neighborhood disorder |
D, SES, HS |
Mediators: social integration, civic engagement, physical activities; |
Multilevel analysis |
C-SES (−, β = 0.28) Contextual (resource: +, β = 1.81) |
| Sisco and Marsiske [31] | US, 6 areas, Advanced cognitive training for independent and vital elderly | Cross-sectional (1997–2000) | 2,802 people | 65–94 | Multiple cognitive functionsb | 2000 US census tracts |
Compositional: Neighborhood socioeconomic position |
D, quadratic age, education | Multilevel analysis | C-SES (−, β = 0.07) | |
| Wee et al. [32] | Singapore | Cross-sectional (2012) | 909 people in two residential sites | 60+ | Cognitive impairment (MMSE) | Blocks |
Compositional: Area socioeconomic status |
D, SES, HS | Multilevel analysis | C-SES (+, OR = 3.8) | |
| Al Hazzouri et al. [20] | US, California; Sacramento Area Latino study on aging | Longitudinal, 7 waves (1998–2008) | 1,789 Mexican American | 60–101 | Cognitive decline (3MSE) | 2000 US census tracts |
Compositional: Neighborhood socioeconomic context |
D, SES, HS | Multilevel analysis (3 levels) | C-SES (−, β = 0.01) | |
| Aneshensel et al. [21] | Heath and retirement survey | Cross-sectional at time 3 (1996) | 4,525 nationally representative | 70+ | TICS | 1990 US census tracts |
Compositional: Socioeconomic disadvantage, racial segregation |
D, SES, HS, social integration | Modifiers: ethnicity (African American and Hispanic) | Multilevel analysis |
C-SES (−, β = 0.04) Inter (ethnicity) |
| Lee et al. [28] |
US, Baltimore The Baltimore memory study |
Cross-sectional (2001) | 1,124 people from 65 neighborhoods of Baltimore | 50–70 | Multiple cognitive functionsb | 2000 US census tracts |
Compositional: Social disorganization, economic deprivation Contextual: Public Safety, physical disorder |
D, SES, HB, numbers of siblings and children, retirement, job control | Modifier: disadvantage genotype | Multilevel logistic analysis (2 levels) | Inter (APOE genotype) |
| Shih et al. [30] | US, Women’s health initiative memory study | Cross-sectional at baseline (1996) | 1,342 women | 65–81 | 3MSE | Combined 1990/2000 US census tracts |
Compositional: Neighborhood socioeconomic status |
D, SES, HB, HS | Multilevel analysis | C-SES (−, β = 0.02) | |
| Wen and Gu [33] |
China, 22 provinces, Chinese longitudinal healthy longevity survey |
Cross-sectionald (baseline 2002) Longitudinal (health outcomes 2005) |
8,099 people | 65–79 | Cognitive impairment (MMSE) | County or city district |
Compositional: Average years of schooling, labor force participation rate, proportion of urban population Contextual: Per capital GDP, Number of hospital beds per 1,000 persons |
D, HB, SES, pollution index, Childhood SESc | Multilevel analysis | C-SES (+; OR: 1.4) | |
| Sheffield and Peek et al. [29] |
US, 5 southwestern states, Hispanic established population for epidemiologic studies of the elderly |
Longitudinal, 3 waves (1993–1999) | 1,980 Hispanic people | 65+ | Cognitive decline (MMSE) | 1990 US census tracts |
Compositional: Economic advantage, social disadvantage |
D, SES, HS | Multilevel analysis (2 and 3 levels) | C-SES (+, OR = 1.8) | |
| Basta et al. [22] |
UK, 5 centers MCR cognitive function and ageing study |
Cross-sectional (1992) | 13,004 people | 65+ | Cognitive impairment (MMSE), | Census 1991 Postcodes mapped to enumeration district |
Compositional: Townsend deprivation score |
Sex, SES, centers | Multilevel analysis | C-SES (+, OR = 2.3) | |
| Lang et al. [27] |
UK, England, The English longitudinal study of ageing |
Cross-sectional (2002) | 8,102 urban-based adults | 50+ | MMSE | Census 2001; super output area |
Compositional and contextual: Index of Multiple Deprivation 2004 |
D, SES, HB, HS | Ordinary least square regression | C-SES (−, β = 0.18) | |
| Wight et al. [34] | US, Study of assets and health dynamic among the oldest old | Cross-sectional (1993) | 3,442 people in urban setting | 70+ | TICS | 1990 US census tracts |
Compositional: Neighborhood education and median household income |
D, SES, HS | Multilevel analysis | C-SES (−, β = 1.76) | |
| Deeg and Thomese [25] |
The Netherlands, Longitudinal aging study Amsterdam |
Cross-sectional (1992) | 2,981 people | 55–85 | MMSE | Postcode |
Contextual: Neighborhood income status: rental price of rented houses, purchase price of owner-occupied houses, the monthly household income of a sample family |
D, SES, living periods |
Modifier: individual income status |
Linear regression | Inter (individual income) |
| Espino et al. [26] |
US, San Antonio longitudinal study of aging |
Cross-sectional (1992) | 452 Mexican and 375 European Americans | 65+ | Cognitive impairment (MMSE) | Neighborhood (Barrio, transitional, suburban) |
Compositional: Barrio (low-income, Mexican–American) Transitional (middle-income, mixed ethnicity) Suburbs (high-income, adapted culture) |
D, SES, HS |
Modifier: ethnicity (Mexican and European Americans) |
Logistic regression | Inter (ethnicity,) |
Results: C-SES Community-level socioeconomic status (more vs less deprived), Inter interaction with individual factors, + positive association, − negative association, OR odds ratio, β effect size of regression coefficient
aIndividual controls: D demographics, including age, sex, marital status, ethnicity, nativity; SES individual socioeconomic status, including education, occupation, social class, income; HS health status, including hypertension, diabetes, stroke, depression, arthritis, heart attack, fall in the past year, hearing impairment; HB health behaviors, including smoking, alcohol drinking, physical activity
b Multiple cognitive functions language, processing speed, eye-hand coordination, executive function, verbal memory and learning, visual memory, visuoconstruction (Lee et al. [28]); memory, reasoning, processing speed, everyday cognition, vocabulary (Sisco and Marsiske [31])
c Childhood SES measured by urban-born, father white-collar job, parents alive at age 10, access to health care, hungry, arm length
dMore detailed results of cross-sectional associations at baseline 2002 were reported in Zeng et al. [47]