Abstract
Objective
To assess domains of social determinants of health (SDoH) and their associations with cognition and quality of life.
Method
This investigation uses baseline data from individuals participating in the ACTIVE trial (n = 2505) to reproduce the SDoH domains described in Healthy People 2030 (economic stability, health care, education, neighborhood and built environment, and social and community context).
The Healthy People 2030 report emphasizes the importance of our social and physical environments and their associations with overall health (Office of Disease Prevention and Health Promotion [ODPHP], n.d.). The places where we live, work, and play are important aspects of our social determinants of health (SDoH). SDoH are components of life such as housing, resources, education, neighborhoods, and communities that influence our health directly or indirectly. Healthy People 2030 presented five domains of SDoH: economic stability, education access and quality, health care access and quality, neighborhood and built environment, and social and community context. Race is a social construct that can have varying operational definitions depending on the observer and how it is assessed. There are similarities in the relationships of race and other SDoH with health outcomes. Therefore, exploring the link between SDoH, other than race, and participant outcomes is important for health disparities researchers as some SDoH domains are potentially mutable and can be targeted for intervention.
Each of the five domains of SDoH has been shown to be associated with cognition and/or quality of life in differing ways. In a longitudinal investigation of predictors of health and functioning, it was found that participants with sustained economic hardship were more likely to have poorer physical, psychological, and cognitive functioning compared to wealthy participants or those with no history of economic hardships (Lynch et al., 1997). Similar findings were found in a more recent study that suggest people with lower socioeconomic status (SES) experience an accelerated decline in cognition compared to those with higher SES (Steptoe & Zaninotto, 2020). We also know that education can be a protective factor for aging adults, and that cognitive impairment plays a large role in both the cost of health care, and health-related quality of life (HRQoL) even prior to a diagnosis of dementia (Pusswald et al., 2015). However, not everyone has the same access to education or the same quality of education. For example, a recent study found that better education in early life slowed age-related decline in cognition function in older adults (Chen et al., 2019). Additionally, the quality of education and academic rigor has a profound influence on the risk of cognitive impairment for older adults (Mantri et al., 2019).
With respect to neighborhood characteristics and the built environment, there have been relatively few studies which have examined the link between these community variables and cognition. Studies of community-level socioeconomic status and an individual’s perceived social and built environment have shown built environment factors to be associated with health behaviors such as physical activity, and mental health outcomes such as depression, and cognitive function in older adults (O. Ferdinand et al., 2012; Wu et al., 2015); as well as cognitive function in older age and dementia (Clarke et al., 2015). Potential reasons for the association between high built environment scores and good cognition in later life have included physical activity within one’s neighborhood if the environment is conducive (Hannon et al., 2012), more cognitive stimulation, and more opportunities for social engagement. Additionally, longitudinal studies are needed to explore these potential mechanisms in detail, and to determine whether community interventions are warranted on a large scale to help maintain cognitive function in older adults.
The current study is one of the few studies to examine all five domains of SDoH within a large sample of Black/African American and White older adults and to examine the associations of the SDoH composites with multiple domains of cognition and quality of life. This investigation uses baseline data from individuals participating in the Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) trial to reproduce the SDoH domains described in Healthy People 2030. In order to assess the predictive validity of the SDoH domains, their associations with cognitive function and quality of life were assessed. It is hypothesized that higher scores on the SDoH domains will be associated with higher levels of cognitive function and health-related quality of life.
Methods
Participants
The ACTIVE study included relatively healthy, community-dwelling adults aged 65 years and older recruited during the baseline visit from March 1998 to October 1999. Individuals were excluded from enrollment in the study if they showed Mini-Mental State Exam scores less than 23, reported functional impairment such as needing extensive assistance with dressing, personal hygiene, or bathing on Minimum Data Set Home Care (Morris & Morris, 1997), or showed visual acuity worse than 20/50. Individuals who self-reported diagnoses of Alzheimer’s disease, stroke, certain cancers; or who had communication difficulties were also excluded (Jobe et al., 2001). The sample size for the overall ACTIVE study was 2802. The possible options for self-reported race were White, Black, Asian, Native Hawaiian or Pacific Islander, American Indian of Alaskan Native, Biracial, and Other. Individuals who did not identify as Black/African American or White, or who had missing data on measures used to create the SDoH composites were not included in this investigation resulting in an analytic sample of 2505 individuals.
Measures
Social Determinants of Health
Measures of SDoH were obtained from four sources: (1) baseline data from the ACTIVE trial, (2) US Census data, (3) North American Industry Classification System (NAICS) data, and (4) self-reported participant occupations were converted to occupational codes and subsequently verified by two raters to create measures of occupational status. The Dictionary of Occupational Titles (DOT) was first published in 1939 and has been a valuable source of occupational data in the United States over the subsequent decades. The revised fourth edition, published in 1991, contains descriptions of 12,762 occupations (U.S. Department of Labor, 1991). Each occupation was methodically analyzed by a rater and contains a unique nine-digit code with embedded occupational complexity characteristics that included complexity of work with data, people, and things. Each domain was organized by decreasing complexity ranging from 0–6 (synthesizing to comparing) for work with data, 0–8 (mentoring to taking instructions) for work with people, and 0–7 (setting up to feeding/off bearing) for work with things. For the current project, these items were then reverse coded so that higher scores corresponded to higher complexity in that occupational domain.
Cognition. Baseline cognitive measures included the domains of speed of processing, reasoning, and memory. In addition, the Mini-Mental State Exam, a common dementia screening tool, was used to assess global mental status (Folstein et al., 1975) and to screen for enrollment purposes.
Cognitive speed of visual processing and attention was assessed with the Useful Field Of View® (UFOV) test using standard procedures (Edwards et al., 2005). This assessment includes four increasingly difficult computerized subtests in which the identification of objects in the participant’s central and peripheral view were required to meet a duration threshold value for 75% correct performance under varying cognitive demands. Scores for each subtest can range from 16.67 to 500 ms and were summed across subtests to obtain a composite score (with higher scores indicating longer presentation durations and thus poorer performance).
Episodic memory measures included the Hopkins Verbal Learning Test (HVLT), the Rey Auditory Verbal Learning test (AVLT) and the Rivermead Behavioral Memory test. The HVLT assessment requires immediate recall of 12 words from three semantic categories across three trials (Brandt, 1991). A recognition trial was then completed by asking participants to indicate whether or not a word had been presented. The AVLT (Jobe et al., 2001; Rey, 1941), requires participants to recall 15 semantically unrelated words across five trials. The total correct across these trials was used in analyses. The Rivermead stories subtest was used to assess prose memory (Wilson et al., 1985). Participants were read a short story and asked to write down as much of the story as possible within two-minutes. Higher scores across all memory measures indicate better performance.
Three measures of inductive reasoning were obtained. The Letter Series test presents a series of 10–15 letters, and participants must recognize a pattern and predict which letter would come next in the series from a set of five possible answers (Thurstone & Thurstone, 1949). Participants are given 6 minutes to answer 30 items. Similar to the Letter Series task, Word Series (Gonda & Schaie, 1985) requires participants to recognize patterns among words. Finally, for the Letter Sets Test, participants are presented with rows of letter sets consisting of four letters each (Ekstrom, 1976). Participants choose the set of letters within each row that does not fit the same pattern as the others. Higher scores across all reasoning measures indicate better performance.
The cognitive assessments within each domain were grouped into three composites using a Blom transformation. Scores for tests at each time point were standardized to the baseline mean and SD. If one or more tests of a composite were missing, the composite score was calculated as the average of the non-missing tests. The memory composite outcome included Hopkins Verbal Learning Test, Rey Auditory Verbal Learning Test, and Rivermead Behavioral Memory Test (immediate recall). The reasoning composite included Letter Series, Letter Sets, and Word Series. The speed composite included the four subtests of the UFOV® test.
Health-Related Quality of Life
The SF-36 was used to assess physical health and mental health-related quality of life (Brazier et al., 1992; Ware & Sherbourne, 1992). Scores on the general health and mental health domains have a potential range of 0–100 with higher scores corresponding to higher levels of health-related quality of life.
Demographics
Information was collected on participant age in years, gender (male vs. female), race (Black/African American vs. White), and marital status (currently married vs. all other response options). Participants were enrolled at performance sites at the University of Alabama in Birmingham, Alabama (UAB); Wayne State University in Detroit, Michigan; Indiana University in Indianapolis, Indiana; Johns Hopkins University in Baltimore, Maryland; Penn State University from State College, Pennsylvania; and Boston, Massachusetts.
Analyses
All analyses were conducted using SAS Version 9.4. T-tests for continuous variables and chi-square for nominal variables were used to compare White and Black participants at baseline on variables of interest. Factor analysis was performed to test our conceptualization of the SDoH composites as well as to reduce the variables to a small number of factors to be examined with respect to impact on cognition and quality of life. A principal components extraction method with varimax rotation was used to generate orthogonal, uncorrelated factors. A criterion of eigenvalues greater than one along with an examination of the scree plot were used to determine the number of factors to retain. The uncorrelated factor scores derived from the factor analysis for the five SDoH composites were subsequently used as independent variables in multiple regression models adjusted for age, gender, marital status, and study site to examine the ability of the SDoH composites to predict cognitive performance, mental health-related quality of life, and physical health-related quality of life.
Results
Descriptive Statistics on Measures of Interest for ACTIVE Cognitive Trial Participants With SDoH Domain Scores by Race (n = 2505).
Note: SDoH = Social determinants of health.
Factor Loadings From Social Determinants of Health Factors With Varimax Rotation.
Note: Bold text identifies variables that load on a factor.
Covariate-Adjusted Associations of Social Determinants of Health Factors With ACTIVE Cognitive Composites and Quality of Life.
For speed of processing, higher values represent longer processing times (worse functioning); MH = mental health-related quality of life domain score from SF-36; GH = general health-related quality of life domain score from SF-36. Higher scores correspond to higher levels of quality of life.
Notes: Models adjusted for age, gender, marital status, and site, with standardized betas reported. Separate covariate-adjusted models conducted for association of each social determinant of health composite with each outcome.
Physician’s offices do not include mental health specialists.
*p < .10, *p < .05, **p < .01, ***p < .0001.
Discussion
Results from this investigation showed that data collected from the ACTIVE trial and administrative sources can be used to evaluate the five SDoH domains as presented in Healthy People 2030, and the ability of higher scores on the composites to predict higher levels of baseline performance on measures of cognition and self-reported, health-related quality of life within a sample of older adults. This serves as an extension of the evaluation of SDoH as shown in previous literature as all five domains were assessed using composite scores within a sample of Black/African American and White older adults. The ACTIVE trial oversampled Black/African American individuals to study cognitive aging across six cites (Jobe et al., 2001). Findings to highlight include higher scores on Health Care Access and Quality being significantly associated with better functioning on all outcome measures. Individuals with better health care access and quality, as it is assessed within this investigation, are able travel shorter distances to receive services and have more options located close to them. Additionally, the strongest associations of SDoH domains with outcome measures were with Reasoning, specifically the associations of Economic Stability and Access to Healthcare with Reasoning. The nature of these associations are less clear and require future investigation.
It was interesting to note that in relation to the social and community context domain, the percent of individuals categorized as White served as a measure of social and community context that loaded on the factor similarly to the presence of golf courses and country clubs and does not constitute a comparison of neighborhoods that are predominately White with neighborhoods that are more diverse in nature. Additionally, the presence of supermarkets and other grocery stores loaded negatively on this factor. A potential explanation is that individuals may have been living in areas that were more exurban, farther outside the city with less housing density and the appropriate amount of land necessary for gold courses and country clubs.
Complementary results were found with health data from the REGARDS (Reasons for Geographic and Racial Differences in Stroke) study. Individuals with one or more SDoH risk factors had an increased risk of 90-day mortality following heart failure hospitalization compared to individuals with no SDoH risk factors (Sterling et al., 2020). Additionally, our results align with a previous investigation examining longitudinal data from multiple sources ranging from 1935–2016 to assess SDoH and health inequalities, which revealed SDoH disparities on outcome measures ranging from infant mortality to life expectancy that still remain despite improvements in overall health during this time span (Singh et al., 2017). Conclusions from those studies, and ours, suggest that SDoH domains are contributing factors to the racial disparities that we continue to see and are potential areas to be targeted by interventions.
Some domains of SDoH are mutable and interventions targeted to improve these domains can potentially help reduce disparities. In fact, previous studies that compared Black/African American and White participants living in similar social environments with similar SES, found that racial disparities in outcomes often disappeared (LaVeist et al., 2008, 2011; Thorpe Jr. et al., 2015). This provides evidence that more of a focus should be placed on the SDoH domains and not race alone. An example of the effectiveness of interventions targeting SDoH domains is a recent literature review which examined six interventions utilized within older minority groups targeted to increase social participation, decrease social isolation, and/or decrease loneliness which are all aspects of social and community context (Pool et al., 2017). The investigations were group interventions that provided volunteering activity, educational activity, physical activity, educational activity combined with physical activity. Five of the six interventions showed improvements on the social and community context constructs assessed providing evidence that SDoH domains can be effectively intervened upon. These improvements on SDoH measures could potentially lead to better health outcomes over time. Other SDOH, including those representing the built and neighborhood environments, may require structural or policy interventions to address.
Strengths of using data from the ACTIVE cognitive trial include an oversampling of Black/African American older adults recruited from six sites across the United States and data on multiple domains of cognition. Due to the inclusion criteria for the ACTIVE cognitive trial, the analytic sample overall was more cognitively intact and functionally healthy than the general population of adults aged 65 and older. This potentially could lead to an underestimation of disparities seen within the general population. Additionally, the analytic sample for this investigation was composed of only Black/African American and White individuals and methodological challenges with cognitive testing including racial and education-based bias may threaten the validity of study findings. These factors must be considered when attempting to generalize findings to larger populations. There may have also been bias in terms of social desirability and recall bias for the self-report measures. Finally, the analyses presented are cross-sectional in nature, and therefore, statements about cause and effect cannot be justified. The ACTIVE Cognitive trial collects data from participants up to 10 years past baseline and merges it with data up to 20 years past baseline from administrative sources such as Medicare claims, death records, and driving accident reports. Future investigations will look at the associations of SDoH with cognition over time and determine if the effectiveness of interventions delivered in ACTIVE differed by varying levels of SDoH.
Conclusion
The findings from this investigation are in line with a recent review article that summarizes the associations of the five SDoH domains with cognition and ADRD risk (Majoka & Schimming, 2021). Future work from the ACTIVE trial will be able to (a) utilize these factor scores to assess the associations of SDoH with cognition and health over time, (b) evaluate the effectiveness of the ACTIVE interventions over time and determine whether intervention effects are moderated by the SDoH factors, and (c) inform future interventions targeted to reduce SDoH-related disparities. These findings point to the importance of access to healthcare and the nature of the associations of SDoH with Reasoning requires additional investigation.
Footnotes
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article:Karlene Ball is one of the inventors of the UFOV® test and speed of processing training program. As such she is a stockholder in the Visual Awareness Research Group, Inc. Dr. Ball serves on the Scientific Advisory Board of Posit Science, and is a paid consultant for the Visual Awareness Resource Group, Inc.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The ACTIVE Cognitive Training Trial was supported by grants from the National Institutes of Health to six field sites and the coordinating center, including Hebrew Senior-Life, Boston (NR04507), the Indiana University School of Medicine (NR04508), the Johns Hopkins University (AG014260), the New England Research Institutes (AG014282), the Pennsylvania State University (AG14263), the University of Alabama at Birmingham (AG14289), and the University of Florida (AG014276). Subsequently, this work was supported by a grant from the National Institutes of Health for a 20 year follow-up of the ACTIVE Trial (R01 AG056486) as well as an Administrative Supplement to this grant to examine SDoH for some outcome measures. Dr. Clay is also supported by the University of Alabama at Birmingham Alzheimer’s Disease Research Center [P20AG068024]. Joshua Owens is supported by the McKnight doctoral fellowship. The opinions expressed here are those of the authors and do not necessarily reflect those of the funding agencies, academic, research, governmental institutions, or corporations involved.
