https://orcid.org/0000-0002-1430-7463
Abstract
Emerging evidence suggests that financial and health literacy deteriorates in advanced age. By contrast, well-being promotes health in aging. This study tested the hypothesis that well-being is associated with slower aging-related literacy decline.
Participants were 1,099 community-based older adults without dementia at baseline. Financial and health literacy was assessed at baseline and annually thereafter via a 32-item measure. Well-being was assessed at baseline via the 18-item version of Ryff’s Scales of Psychological Well-Being.
During up to 12 years of annual follow-up, literacy declined about 1 percentage point per year on average (β = −0.91, standard error [SE] = 0.08, p < .001); however, there was considerable variation in change in literacy between participants (random slopes variance = 1.24, SE = 0.15, p < .001). In a linear mixed-effects model adjusted for age, sex, and education, higher well-being was associated with higher starting level of literacy (β = 2.31, SE = 0.67, p = .001) and, critically, slower literacy decline (β = 0.29, SE = 0.11, p = .01). The association of higher well-being with slower literacy decline persisted in models that additionally adjusted for income, medical conditions, depressive symptoms, and a robust measure of global cognition.
This study suggests that well-being helps stave off aging-related literacy decline.
Domain-specific financial and health literacy refers to the ability to acquire, manipulate, and utilize financial and health knowledge in order to successfully navigate society (Baker, 2006; Lusardi & Mitchell, 2014). While essential throughout adulthood, financial and health literacy is particularly critical in advanced age, as older adults face a slew of nuanced and consequential decisions that rely on their application of financial and health knowledge (e.g., appropriately drawing down retirement funds, determining transfers of wealth to loved ones, weighing end-of-life plans). Alarmingly, many older adults are poorly equipped to make these complex decisions due to low literacy (Gazmararian et al., 1999; Lusardi & Mitchell, 2009). For example, despite decades of experience managing finances, the National Financial Capability Study found that only roughly half of older adults can correctly answer three questions about simple monetary concepts that form the basis of many financial decisions (worth noting, this finding is unlikely to be due to mild cognitive impairment or dementia because older adults outperformed their younger-adult counterparts on the same three questions; Mitchell & Lusardi, 2021). In turn, low literacy undercuts older adults’ participation in popular retirement and healthcare programs (e.g., defined contribution retirement plans, Medicare Advantage plans) and contributes to a range of adverse outcomes in advanced age, including excessive debt, illness, and mortality (Baker et al., 2008; Braun et al., 2018; Dewalt et al., 2004; Fisch et al., 2020; Lusardi & Tufano, 2009; Sudore et al., 2006; Wolf et al., 2005).
The vast majority of research investigating low financial and health literacy in relation to adverse outcomes in aging has examined literacy at a single time point. However, emerging longitudinal evidence indicates that older adults’ low literacy is partly aging-related, that is, due to late-life decline in literacy over time (Angrisani et al., 2020; Kobayashi, Wardle, Wolf, et al., 2015; Yu et al., 2018, 2020). Much remains to be learned about why literacy declines in advanced age, but common aging-related neuropathologies are thought be a primary culprit. Support for this idea comes from prior work associating postmortem Alzheimer’s disease pathology and transactive response (TAR) DNA-binding protein 43 pathology with lower literacy during life (Kapasi et al., 2019; Yu et al., 2017, 2020). Because these and other neuropathologies accumulate slowly over many years prior to the manifestation of overt impairment (Jack et al., 2018), literacy likely degrades in a subtle but insidious manner, rendering even high-functioning older adults more susceptible to poor decisions and scams (Yu et al., 2021).
Importantly, while aging-related decline in financial and health literacy has been observed at the group level, there is substantial variation in literacy trajectories between older adults. For example, during up to 10 annual literacy assessments, we found that roughly half of older adults showed mild literacy decline (about 1 percentage point per year), while the other half of older adults showed much steeper literacy decline (>2 percentage points per year; Yu et al., 2020). Considerable variation in literacy decline between older adults was observed in a different longitudinal study as well (Kobayashi, Wardle, Wolf, et al., 2015). This person-to-person variation raises the possibility that certain factors might slow literacy decline in advanced age.
We recently proposed a conceptual model of financial and health\care decision making, termed “degraded rationality,” that has bearing on aging-related literacy decline (Boyle et al., 2022). Our conceptual model posits that decision making and closely related constructs like literacy depend on the dynamic interplay of diverse resources, including cognitive ability, brain health, and psychosocial factors. Psychosocial factors are of special interest to aging-related literacy decline in light of evidence that they augment or deplete brain reserve depending on their valence. For example, purpose in life, a positive psychosocial factor and subcomponent of well-being, has been associated with better cognitive outcomes in aging (e.g., slower rate of cognitive decline, lower risk of clinical dementia). Conversely, negative psychosocial factors, such as distress proneness and loneliness, have been associated with worse cognitive outcomes. Notably, the same psychosocial factors were unrelated to common neuropathologies, and in the case of purpose in life, the association of Alzheimer’s disease pathology with cognitive decline was weakened when purpose was high (Boyle et al., 2012; Wilson et al., 2006, 2014; Wilson, Krueger, et al., 2007; Wilson, Schneider, et al., 2007). These findings suggest that psychosocial factors do not affect neuropathological processes directly but rather amplify or dampen the brain’s capacity to support higher-order cognitive abilities in the face of neuropathologies. This work raises the possibility that psychosocial factors operate similarly on literacy, a higher-order ability that is not measured by traditional neuropsychological tests.
In the current study, we investigated whether well-being, a core aspect of psychological functioning, was related to aging-related decline in financial and health literacy. Encompassing much more than contentedness, well-being refers to one’s purpose in life, autonomy, personal growth, environmental mastery, positive social relationships, and self-acceptance. We were interested in well-being specifically given that positive psychosocial factors might buffer against the adverse impact of common aging-related neuropathologies and in light of a substantial literature associating well-being with a wide range of good outcomes, including financial stability, health-promoting behaviors, and health and longevity (Diener et al., 2017, 2018; Hill et al., 2016; Ryff et al., 2016). We therefore hypothesized an association of higher well-being with slower literacy decline. Our data came from 1,099 community-based older participants without dementia at analytic baseline who were enrolled in the Rush Memory and Aging Project (MAP), a prospective observational cohort study of aging (Bennett et al., 2018). Well-being was assessed at baseline, and literacy was assessed at baseline and annually thereafter for up to 12 years of follow-up. In a linear mixed-effects model, we examined the association of well-being with starting level of literacy and subsequent change in literacy, adjusting for basic demographics (age, sex, education). Next, to tease apart well-being from potential confounding variables, we repeated the linear mixed-effects model, this time further adjusting for income, medical conditions, depressive symptoms, and a robust measure of global cognition. In secondary analyses, we examined the relation of the six subcomponents of well-being (i.e., purpose in life, autonomy, personal growth, environmental mastery, positive relationships, and self-acceptance) with literacy decline.
Method
Participants
MAP is an ongoing, longitudinal, clinical–pathological study of aging and aging-related diseases in the Chicago area (Bennett et al., 2018). Participants are community-based older adults without known dementia upon study entry who agree to annual structured clinical evaluations, neuropsychological testing, and blood draws, in addition to donation of brain, spinal cord, nerve, and muscle at death. Recruitment occurs in a variety of settings, including continuous care facilities, retirement communities, Section 8 and Section 202 housing subsidized by the Department of Housing and Urban Development, local churches, and other social service agencies. Evaluations are performed as home visits. These design features help ensure good representation of common comorbidities and high follow-up participation (approximately 95%). Alzheimer’s dementia is documented annually using the criteria of the National Institute of Neurologic and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association, which require a history of cognitive decline and impairment in memory and one or more additional cognitive domains (McKhann et al., 1984). Written informed consent is obtained from participants following review of the benefits and risks of participation prior to study entry. MAP is approved by an Institutional Review Board of Rush University Medical Center (protocol number 07071201).
MAP started in 1997, and a substudy of financial and healthcare decision making, which includes an assessment of financial and health literacy, began in 2010. Of the 2,242 participants with an initial MAP evaluation, 475 passed away before the decision making substudy began and 128 withdrew before completing the baseline decision making evaluation. Another 203 participants moved or were ineligible due to severe comprehension or sensory limitations; 79 declined the decision making substudy; and 27 had not yet completed the baseline decision making evaluation. Among the remaining 1,330 participants who completed the baseline decision making evaluation, 188 had only completed the baseline decision making evaluation and 7 had missing literacy scores. The remaining 1,135 participants completed at least one follow-up literacy assessment, of which 1,099 were without dementia and available for the current analyses.
Financial and Health Literacy
Financial and health literacy was assessed at baseline and annually thereafter via a 32-item measure requiring knowledge and utilization of basic financial- and health-related concepts (e.g., compound interest, stocks and bonds, treatment-associated risk, common causes of mortality). Many items were modeled off items from the Health and Retirement Survey (Mitchell & Lusardi, 2021). Question format was true/false or multiple choice. Percent correct was calculated separately for the financial literacy items (23 items) and the health literacy items (9 items). These two percentages were then averaged, yielding a measure of total literacy that equally weighs financial and health literacy at the domain level (range = 0%–100%). This measure has adequate internal consistency (Cronbach’s coefficient alpha = 0.83) and has been associated with health-promoting behaviors, health status, and longevity (Bennett et al., 2012; Stewart et al., 2020; Yu et al., 2017).
Well-Being
Well-being was assessed at baseline via the 18-item version of Ryff’s Scales of Psychological Well-Being (Ryff & Keyes, 1995). This measure consists of three items for each of the six subcomponents of well-being: (1) purpose in life (goal-directedness, finding meaning in current and future activities), (2) autonomy (having independent opinions and values), (3) personal growth (openness to new experiences, commitment to lifelong learning and development), (4) environmental mastery (agency over the demands and responsibilities of daily life), (5) positive relationships (having close, trusting social connections), and (6) self-acceptance (acknowledging and accepting personal strengths and shortcomings). For each item, participants rated their agreement or disagreement along a 7-point Likert rating scale. Ratings for all 18 items were averaged, yielding a measure of overall well-being (range = 1–7, with higher ratings indicating higher well-being). To calculate the six subcomponents of well-being, ratings of the three items comprising each subcomponent were averaged (range = 1–7, with higher ratings indicating greater endorsement of the subcomponent). Well-being and the six subcomponents have adequate internal consistency (Cronbach’s coefficient alpha for well-being = 0.73; Cronbach’s coefficient alphas for the six subcomponents ≥ 0.79) and established convergent, discriminant, and predictive validity (Keyes et al., 2002; Ryff, 2014; Wilson et al., 2013).
Covariates
Age was determined relative to analytic baseline (i.e., the date that participants completed the baseline assessments of literacy and well-being). Participants reported their sex (male or female) and education (years of schooling) at their initial MAP evaluation. Income was assessed at analytic baseline by having participants select among the following 10 categories: 1: $0–$4999, 2: $5000–$9999, 3: $10,000–$14,999, 4: $15,000–$19,999, 5: $20,000–$24,999, 6: $25,000–$29,999, 7: $30,000–$34,999, 8: $35,000–$49,999, 9: $50,000–$74,999, 10: >$75,000. Chronic medical conditions were the sum of self-reported medical conditions, including hypertension, diabetes, and cancer, at analytic baseline. Depressive symptoms was measured at analytic baseline via a 10-item version of the Center for Epidemiological Studies—Depression scale (Radloff, 1977; Wilson et al., 2014). Global cognition at analytic baseline was measured via 19 individual, performance-based tests, as previously described (Bennett et al., 2018). Raw scores on individual tests were converted to z-scores using the baseline mean and standard deviation (SD) of the full MAP cohort (from which the current group is drawn). The z-scores of individual tests were averaged, providing a measure of global cognition.
Statistical Analysis
Bivariate associations of well-being with starting level of financial and health literacy and covariates were initially examined via Pearson correlations, Spearman correlations, or t tests, as appropriate. Associations of well-being with starting level of literacy and subsequent change in literacy were examined via linear mixed-effects models with annual literacy scores as the longitudinal continuous outcome. The core model was comprised of terms for well-being, age, sex, education, and time in years since baseline, as well as the interaction of well-being, age, sex, and education with time. The interactions with time measured the association of the terms with change in literacy. Next, building on the core model, we performed separate linear mixed-effects models that additionally included terms for one of the following: income, medical conditions, depressive symptoms, and global cognition. To illustrate, the model with global cognition included all the terms in the core model plus terms for global cognition and the interaction of global cognition with time. In secondary analyses, we examined the six subcomponents of well-being in relation to change in literacy by repeating the core model, this time with overall well-being replaced by one of the subcomponents of well-being. This was done separately for each of the six subcomponents.
Results
At baseline, participants’ mean age was 81.0 years (SD = 7.5, range: 58.8–100.2), and their mean financial and heath literacy was 69.3% (SD = 14.1, range: 24.2–100). Please refer to Table 1 for additional descriptive characteristics of the participants. Higher well-being was associated with higher starting level of literacy (r = 0.25, p < .001), younger age (r = −0.21, p < .001), more years of education (r = 0.26, p < .001), higher income (r = 0.24, p < .001), fewer medical conditions (r = −0.09, p = .01), fewer depressive symptoms (r = −0.36, p < .001), and higher global cognition (r = 0.27, p < .001). Well-being did not differ by sex (Mfemales = 5.6, SDfemales = 0.6; Mmales = 5.6, SDmales = 0.6; t(1079) = −0.64, p = .52).
Descriptive Characteristics of Participants at Baseline
| Characteristic | M (SD) or percent |
|---|---|
| Literacy | 69.3% (14.1) |
| Well-being | 5.6 (0.6) |
| Age | 81.0 (7.5) |
| Female | 76.0% |
| Education | 15.6 (3.0) |
| Income | 7.53 (2.4) |
| Medical conditions | 1.61 (1.1) |
| Depressive symptomsa | 0 (0–1) |
| Global cognition | 0.21 (0.5) |
| MMSE | 28.3 (1.7) |
| Characteristic | M (SD) or percent |
|---|---|
| Literacy | 69.3% (14.1) |
| Well-being | 5.6 (0.6) |
| Age | 81.0 (7.5) |
| Female | 76.0% |
| Education | 15.6 (3.0) |
| Income | 7.53 (2.4) |
| Medical conditions | 1.61 (1.1) |
| Depressive symptomsa | 0 (0–1) |
| Global cognition | 0.21 (0.5) |
| MMSE | 28.3 (1.7) |
Notes: Values are mean (SD) unless otherwise indicated. MMSE = Mini-Mental State Examination; SD = standard deviation.
aMedian (interquartile range).
Descriptive Characteristics of Participants at Baseline
| Characteristic | M (SD) or percent |
|---|---|
| Literacy | 69.3% (14.1) |
| Well-being | 5.6 (0.6) |
| Age | 81.0 (7.5) |
| Female | 76.0% |
| Education | 15.6 (3.0) |
| Income | 7.53 (2.4) |
| Medical conditions | 1.61 (1.1) |
| Depressive symptomsa | 0 (0–1) |
| Global cognition | 0.21 (0.5) |
| MMSE | 28.3 (1.7) |
| Characteristic | M (SD) or percent |
|---|---|
| Literacy | 69.3% (14.1) |
| Well-being | 5.6 (0.6) |
| Age | 81.0 (7.5) |
| Female | 76.0% |
| Education | 15.6 (3.0) |
| Income | 7.53 (2.4) |
| Medical conditions | 1.61 (1.1) |
| Depressive symptomsa | 0 (0–1) |
| Global cognition | 0.21 (0.5) |
| MMSE | 28.3 (1.7) |
Notes: Values are mean (SD) unless otherwise indicated. MMSE = Mini-Mental State Examination; SD = standard deviation.
aMedian (interquartile range).
Associations of Well-Being With Change in Literacy
During up to 12 years of annual follow-up (M = 6.1, SD = 3.3), financial and health literacy declined about 1 percentage point per year on average (β = −0.91, standard error [SE] = 0.08, p < .001); however, there was considerable person-to-person variation in change in literacy (random slopes variance = 1.24, SE = 0.15, p < .001). In the core model that adjusted for age, sex, and education, higher well-being was associated with higher starting level of literacy and, critically, slower literacy decline (starting level of literacy: β = 2.31, SE = 0.67, p = .001; change in literacy: β = 0.29, SE = 0.11, p = .01; Table 2). Contextualizing the latter finding, the model-predicted decline in literacy for participants with low well-being (10th percentile) was 1.25 percentage points per year. By contrast, the model-predicted decline for participants with high well-being (90th percentile) was only 0.83 percentage points per year (Figure 1). In models that included terms for other covariates, the association of well-being with starting level of literacy remained significant after adjusting for medical conditions and depression symptoms but was not significant after adjusting for income and global cognition. Importantly, however, the association of higher well-being with slower literacy decline persisted in all models (Table 2). In secondary analyses examining the six subcomponents of well-being in relation to starting level and change in literacy, personal growth and environmental mastery were associated with higher starting literacy and slower literacy decline (Supplementary Table 1). Self-acceptance was not associated with starting literacy but was associated with slower literacy decline. Purpose in life was associated with higher starting literacy but not change in literacy. Autonomy and positive relationships were not associated with starting literacy or change in literacy.
Associations of Well-Being With Starting Level and Change in Literacy
| Model | Model term | β (SE) | p Value |
|---|---|---|---|
| A (core) | Well-being | 2.31 (0.67) | .001 |
| Time | −0.91 (0.08) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| B (adjusted for income) | Well-being | 1.14 (0.66) | .09 |
| Time | −0.93 (0.08) | <.001 | |
| Well-being × time | 0.27 (0.11) | .01 | |
| C (adjusted for medical conditions) | Well-being | 2.35 (0.67) | <.001 |
| Time | −0.82 (0.12) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| D (adjusted for depressive symptoms) | Well-being | 1.57 (0.71) | .03 |
| Time | −0.89 (0.08) | <.001 | |
| Well-being × time | 0.26 (0.11) | .03 | |
| E (adjusted for global cognition) | Well-being | 0.29 (0.56) | .60 |
| Time | −1.11 (0.09) | <.001 | |
| Well-being × time | 0.24 (0.11) | .04 |
| Model | Model term | β (SE) | p Value |
|---|---|---|---|
| A (core) | Well-being | 2.31 (0.67) | .001 |
| Time | −0.91 (0.08) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| B (adjusted for income) | Well-being | 1.14 (0.66) | .09 |
| Time | −0.93 (0.08) | <.001 | |
| Well-being × time | 0.27 (0.11) | .01 | |
| C (adjusted for medical conditions) | Well-being | 2.35 (0.67) | <.001 |
| Time | −0.82 (0.12) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| D (adjusted for depressive symptoms) | Well-being | 1.57 (0.71) | .03 |
| Time | −0.89 (0.08) | <.001 | |
| Well-being × time | 0.26 (0.11) | .03 | |
| E (adjusted for global cognition) | Well-being | 0.29 (0.56) | .60 |
| Time | −1.11 (0.09) | <.001 | |
| Well-being × time | 0.24 (0.11) | .04 |
Notes: Model A (core model) adjusted for age, sex, and education and their interactions with time. Building on the core model, Model B additionally adjusted for income and the interaction of income with time; Model C additionally adjusted for medical conditions and the interaction of medical conditions with time; Model D additionally adjusted for depressive symptoms and the interaction of depressive symptoms with time; and Model E additionally adjusted for global cognition and the interaction of global cognition with time. SE = standard error.
Associations of Well-Being With Starting Level and Change in Literacy
| Model | Model term | β (SE) | p Value |
|---|---|---|---|
| A (core) | Well-being | 2.31 (0.67) | .001 |
| Time | −0.91 (0.08) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| B (adjusted for income) | Well-being | 1.14 (0.66) | .09 |
| Time | −0.93 (0.08) | <.001 | |
| Well-being × time | 0.27 (0.11) | .01 | |
| C (adjusted for medical conditions) | Well-being | 2.35 (0.67) | <.001 |
| Time | −0.82 (0.12) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| D (adjusted for depressive symptoms) | Well-being | 1.57 (0.71) | .03 |
| Time | −0.89 (0.08) | <.001 | |
| Well-being × time | 0.26 (0.11) | .03 | |
| E (adjusted for global cognition) | Well-being | 0.29 (0.56) | .60 |
| Time | −1.11 (0.09) | <.001 | |
| Well-being × time | 0.24 (0.11) | .04 |
| Model | Model term | β (SE) | p Value |
|---|---|---|---|
| A (core) | Well-being | 2.31 (0.67) | .001 |
| Time | −0.91 (0.08) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| B (adjusted for income) | Well-being | 1.14 (0.66) | .09 |
| Time | −0.93 (0.08) | <.001 | |
| Well-being × time | 0.27 (0.11) | .01 | |
| C (adjusted for medical conditions) | Well-being | 2.35 (0.67) | <.001 |
| Time | −0.82 (0.12) | <.001 | |
| Well-being × time | 0.29 (0.11) | .01 | |
| D (adjusted for depressive symptoms) | Well-being | 1.57 (0.71) | .03 |
| Time | −0.89 (0.08) | <.001 | |
| Well-being × time | 0.26 (0.11) | .03 | |
| E (adjusted for global cognition) | Well-being | 0.29 (0.56) | .60 |
| Time | −1.11 (0.09) | <.001 | |
| Well-being × time | 0.24 (0.11) | .04 |
Notes: Model A (core model) adjusted for age, sex, and education and their interactions with time. Building on the core model, Model B additionally adjusted for income and the interaction of income with time; Model C additionally adjusted for medical conditions and the interaction of medical conditions with time; Model D additionally adjusted for depressive symptoms and the interaction of depressive symptoms with time; and Model E additionally adjusted for global cognition and the interaction of global cognition with time. SE = standard error.
Predicted starting level and 5-year paths of change in literacy for participants with high well-being (dashed line, 90th percentile) and low well-being (solid line, 10th percentile), with 95% confidence bands, adjusted for age, sex, education, and their interactions with time.
Discussion
Emerging longitudinal evidence indicates that older adults’ financial and health literacy declines over time (Angrisani et al., 2020; Kobayashi, Wardle, Wolf, et al., 2015; Yu et al., 2018, 2020). This likely presents unique challenges in aging. While lifelong low literacy is problematic in its own right, it at least affords a broad window of opportunity to develop and implement compensatory strategies. By contrast, aging-related literacy decline involves the insidious degradation of an ability that older adults have grown accustomed to using throughout adulthood. Even if literacy decline is detected, older adults may be left to compensate in the context of other degraded resources (e.g., poorer health, cognition, and social support) and thus considerably disadvantaged. Worse yet, if literacy decline goes undetected, older adults might continue to behave as if their literacy is fully intact, leaving them vulnerable to poor financial and healthcare decision making or financial exploitation (Yu et al., 2021).
Despite the potentially dire consequences, almost nothing is known about whether specific factors alter the trajectory of literacy in aging. Addressing this gap, the current study leveraged up to 13 years of annual financial and health literacy assessments among more than 1,000 older adults without dementia at baseline and examined the association of well-being with subsequent change in literacy. We found that higher well-being was associated with slower literacy decline and that this association was relatively independent of income, medical conditions, depressive symptoms, and a robust measure of global cognition. This suggests that well-being has therapeutic potential in attenuating literacy decline and extending independence in advanced age.
To our knowledge, the only study similar to the current one found that older adults who used the internet or participated in civic and cultural activities were less likely to show decline on a health literacy screener measured at two time points (Kobayashi, Wardle, & von Wagner, 2015). The current study builds on this earlier finding by tracking literacy via a much more extensive measure across many waves of data and by explicitly disentangling the associations of well-being with starting level of literacy and change in literacy. The specificity and robustness of our finding was further reinforced by analyses investigating potential confounding factors. In particular, by adjusting for depressive symptoms, we demonstrated that psychological thriving, and not merely a lack of distress, was the key ingredient driving the protective effect on literacy decline. Additionally, given that literacy and cognition are related but distinct higher-order abilities, it was important to show that the association of well-being with slower literacy decline persisted above and beyond global cognition. This study also expands the scope of factors that might slow literacy decline. Whereas accessibility might be a barrier for specific behaviors such as using the internet or participating in civic and cultural activities, especially for those with low literacy, well-being is a core aspect of psychological functioning that is essential to all older adults. More broadly, the current finding adds to the burgeoning literature linking positive psychosocial factors with a wide range of good financial and health outcomes (Diener et al., 2017, 2018; Hill et al., 2016).
While providing compelling evidence of a relationship between well-being and slower literacy decline, this study does not directly address the mechanisms underlying this relationship. We suspect that late-life literacy decline is primarily driven by common aging-related neuropathologies, such as Alzheimer’s disease pathology and TAR DNA-binding protein 43 pathology, which build up slowly in the years preceding overt impairment (Jack et al., 2018; Kapasi et al., 2019; Yu et al., 2017, 2020). At the same time, psychosocial factors appear to modulate the brain’s capacity to support higher-order abilities as neuropathologies accumulate (i.e., psychosocial factors seem to increase or decrease brain reserve depending on their valance; Boyle et al., 2012; Wilson et al., 2006, 2014; Wilson, Krueger, et al., 2007; Wilson, Schneider, et al., 2007). Integrating these lines of evidence with the current finding, well-being might buffer the brain’s capacity to support literacy in the face of neuropathologies. While the neurobiological basis of this buffering effect remains unknown, a growing body of evidence suggests that well-being promotes cardiovascular, immune, and endocrine health (e.g., by lowering blood pressure, inflammation, and cortisol; Diener et al., 2017), and these or other as-yet unidentified physiological benefits of well-being might boost brain resilience (e.g., by enhancing neural efficiency or neural recruitment). As an initial step toward elucidating mechanisms, in future work, we plan on investigating whether well-being influences literacy decline independent of neuropathologies or whether well-being moderates the relationship between neuropathologies and literacy decline.
In addition to biological mechanisms, well-being might operate on contextual and behavioral levels to bolster literacy in aging. For example, given that older persons with higher well-being tend to be financial- and health-minded (Hill et al., 2016; Ryff et al., 2016), these individuals might find themselves in environments or actively pursue opportunities to continually develop their financial and health knowledge, and this might counter aging-related literacy decline. In particular, our secondary analyses examining the subcomponents of well-being keyed in on personal growth, environmental mastery, and self-acceptance. Thus, lifelong learning, a feeling of control over daily demands, and awareness of personal limitations might be especially important in staving off literacy decline.
Regarding therapeutic implications, a major problem of interventions that target literacy directly (e.g., via educational activities) is that many people put off opportunities to learn about finances and health/healthcare (Duflo & Saez, 2003). Assessing and, when indicated, intervening on well-being might be key in motivating older adults to participate in direct literacy interventions (Ryan & Deci, 2000). For example, older adults with low environmental mastery might adopt an external locus of causality in relation to their finances and health and therefore perceive opportunities to learn about these topics as not worth their time. However, if environmental mastery is increased (e.g., by eliciting personal examples of how literacy tangibly affects their finances and health), then older adults should be more energized to engage with interventions that directly target literacy. Beyond addressing motivational barriers, well-being is attractive from a therapeutic perspective because it is universally valued, can be intervened on in a flexible manner, confers broad health benefits, and is inherently satisfying (therefore, well-being-related benefits might be self-reinforcing and long-lasting).
This study has strengths and limitations. A major strength is our well-characterized cohort of older adults. Confidence in our finding is increased because we tracked literacy in high fidelity over many years, disentangled associations of well-being with starting level and change in literacy, and ruled out potential confounding variables (e.g., depressive symptoms, global cognition). Also worth noting, because literacy decline likely occurred prior to participants’ first literacy assessment, the current finding probably underestimate well-being’s full potential in attenuating aging-related literacy decline. A study limitation is that our group of older adults was mostly White and highly educated. We are actively collecting longitudinal literacy data among older Black adults and will examine generalization of the current finding to this group once we accrue enough data. Another limitation is that this study does not address mechanisms underlying the relationship of well-being with literacy decline. As alluded to above, in future studies, we plan on examining whether well-being and other psychosocial factors influence the association of postmortem neuropathologies with literacy and related abilities (e.g., scam susceptibility). This is an exciting area of research because it will illuminate how psychosocial factors affect the brain’s capacity to support critical higher-order abilities and cope with common neuropathologies that cannot be directly intervened on at present.
Funding
This work was supported by National Institute on Aging (R01 AG017917 to D. A. Bennett and R01 AG033678, R01 AG034374, and R01 AG060376 to P. A. Boyle).
Conflict of Interest
None declared.
Acknowledgments
We express our hearty thanks to the thousands of participants in the Rush Memory and Aging Project (MAP) and the investigators and staff of the Rush Alzheimer’s Disease Center (RADC). This research would not be possible without their steadfast dedication. Please visit the RADC’s Research Resource Sharing Hub at www.radc.rush.edu to request data from MAP for research purposes.
