Linking Mastery Across the Life Course to Mobility Device Use in Later Life

Abstract

Objectives

Mastery in older ages is shaped by earlier-life experiences. Prior research has demonstrated that mastery is associated with health-promoting behaviors; however, little research has examined whether mastery histories influence health behaviors such as mobility device use in later life.

Method

Using 25 years of data from the Americans’ Changing Lives Survey (N = 1,427), this research explores whether different trajectories of life course mastery influence the odds that an older adult will use a mobility device when experiencing functional impairment. We used growth mixture models with a distal outcome and examined the relationship between functional limitations and mobility device use as it varies across latent classes of life course mastery, controlling for social and health factors.

Results

The odds of device use in the face of functional limitations were significantly higher among those with a history of high life course mastery, relative to those with low life course mastery, all things being equal.

Discussion

Our findings suggest that mastery over the life course is a source of psychological human capital that is associated with health-promoting behaviors in later life among those with functional limitations.

Mastery refers to the perceptions individuals hold about their ability to control or manage their life circumstances (Pearlin & Schooler, 1978). Prior research has demonstrated that mastery is positively associated with health and well-being in part by attenuating the adverse effects of stressful life events and increasing participation in health-promoting behaviors (Thoits, 2010). Mastery in older ages is shaped by life experiences that unfold over the entire life course including past and present events (Pearlin, Nguyen, Schieman, & Milkie, 2007). Mastery declines with age and physical impairment (Mirowsky, 1995; Ross & Mirowsky, 2002; Schieman & Turner, 1998); therefore, it is important to understand the role of mastery in relation to changes in functional health or the loss of independence due to disabling health conditions in later life.

Among those who experience functional health declines, the decision to use mobility-related assistive technology (e.g., canes, walkers, or wheelchairs) is viewed as a health-promoting behavior and is influenced by numerous factors including attitudes, subjective norms, and perceived control (Roelands, Van Oost, Depoorter, & Buysse, 2002). Drawing from Mirowsky and Ross’ (1998) human capital hypothesis and an expanded version of Andersen’s (1968) behavioral model (see Bradley et al., 2002), we posit that the adoption of mobility devices among those with functional limitations in later life varies depending on past histories of mastery as they have evolved over adulthood. To our knowledge, no research has investigated whether mastery experienced over the entire adult life course is associated with health-promoting behaviors such as mobility device use in later life.

Literature Review

Mastery Over the Life Course as a Source of Psychological Human Capital

According to Mirowsky and Ross (1998), human capital is the “productive capacity developed, embodied, and stocked in human beings themselves” (p. 416). Mirowsky and Ross’ (1998) primary interest was to connect education and health using the theory of human capital; the authors speculated that one of the main pathways linking educational attainment to positive health outcomes was a healthy lifestyle derived from higher sense of control. Although Mirowsky and Ross (1998) focused on sense of control, the focus of this research is on mastery; however, previous research has noted a substantial theoretical and empirical overlap between Mirowsky and Ross’ (1991) personal control scale and Pearlin’s (1978) personal mastery scale (Hitlin & Long, 2009). Both concepts tap into a generalized, rather than situational, perception of one’s ability to shape his or her life (Hitlin & Long, 2009).

Previous research has observed important differences in mastery levels by social status including socioeconomic status, gender, race/ethnicity, and age, where individuals with marginalized statuses typically report lower personal mastery and sense of control (Cassidy & Davies, 2003; Mirowsky & Ross, 1998; Pearlin et al., 2007). Individuals who hold marginalized statuses often experience a greater burden of stressful life events over their adult lives, including discrimination, financial hardship (e.g., job loss), and health shocks (Pearlin et al., 2007; Turner & Avison, 2003), which may lower their self-perceptions of control. An individual’s sense of control maps onto objective experiences related to social status because it reflects exposure to structural opportunities or constraints throughout their individual biography (Mirowsky, 1995; Ross & Mirowsky, 2002). Given that mastery levels are thought to peak in midlife and decline into old age (Hitlin & Long, 2009), those with lower levels of mastery in midlife may face even larger health disparities in later life due to fewer psychological resources to manage health shocks and age-associated health declines.

There is substantial empirical evidence linking higher levels of mastery to better health and well-being. To illustrate, Bailis, Segall, Mahon, Chipperfield, and Dunn (2001) documented a positive association between higher mastery and better self-ratings of physical and mental health. It has been hypothesized that mastery protects health and well-being by operating as an important mediator and moderator (e.g., buffer) for the deleterious effects of stress on mental and physical health (Pearlin, 2010; Pearlin et al., 2007; Thoits, 2010). This line of research proposes that mastery is a psychological resource that individuals can draw upon to help cope with stressful life situations. Greater mastery is thought to promote problem solving, increase confidence, and encourage more optimistic and hopeful assessments of difficult situations (Pudrovska, Schieman, Pearlin, & Nguyen, 2005; Thoits, 2010).

Need, Enabling, and Psychosocial Factors Related to Health Behaviors

Mastery can be conceptualized as a form of psychological human capital that is cultivated over the life course and drawn upon to foster health promotion by mitigating stressful experiences and increasing participation in positive health behaviors. In particular, we are interested in the linkages between mastery and health behaviors as it relates to adoption of assistive technology such as mobility devices. Andersen’s (1968) behavioral model highlights the role of enabling and need factors as they are related to health behaviors. Enabling factors refer to individual/family and community resources that influence one’s “ability to secure services” such as income, health insurance, geographic characteristics, and presence/absence of facilities (Andersen & Newman, 1973, p. 107), whereas need refers to perceived and evaluated illness including symptoms, impairment, general state, and diagnoses (Andersen & Newman, 1973).

Bradley and colleagues (2002) extend the original Andersen model to include the role of psychosocial factors. The expanded model continues to emphasize the importance of need and enabling factors for health behavior including use of health services; however, Bradley and colleagues (2002) add an additional, intermediate pathway linking enabling and need factors to use of health services via psychosocial factors. As stated previously, mastery empowers individuals to persevere in the face of adversity and to take advantage of opportunities such as modifying health behavior and using health services. However, health behavior reflects both need and enabling factors (Andersen, 1968, 1995; Bradley et al., 2002).

Consequently, the presence of need factors is consistently shown to be the primary predictor of mobility device use (Clarke & Colantonio, 2005; Gitlin, Schemm, Landsberg, & Burgh, 1996; Mathieson, Kronenfeld, & Keith, 2002; Verbrugge & Sevak, 2002). Although need factors may be the primary driver for mobility device use, it is only in the context of enabling and psychosocial factors that individuals with functional impairment are able to acquire and use devices. Previous research underscores important disparities in mobility device use among those with fewer enabling and psychosocial resources (Roelands et al., 2002); therefore, we apply the expanded model of Andersen’s behavioral model to mobility device use emphasizing the role of psychosocial factors, but situating this linkage within the milieu of need and enabling factors.

Mobility Device Use as a Health-Promoting Behavior

The risk of chronic conditions and functional limitations increases as individuals age, and accordingly, older age and disability are associated with lower levels of current mastery (Schieman & Turner, 1998). However, prior research has also noted the potential for mastery to “facilitate adaptation under stressful life situations” (Jang, Haley, Small, & Mortimer, 2002, p. 808) including preventing functional declines and facilitating recovery (Kempen et al., 1999). The choice to use assistive technology for mobility following the onset of functional limitations can be viewed as a health-promoting behavior because assistive technology increases independence and often improves psychological well-being and quality of life (Verbrugge & Sevak, 2002; Wressle & Samuelsson, 2004). For instance, the use of assistive technology among persons with disability is associated with less unmet need (Agree & Freedman, 2003) and fewer hours of personal assistance (Hoenig, Taylor, & Sloan, 2003). Moreover, Verbrugge and Sevak (2002) found that assistive technology use among older adults was a more effective adaption strategy than personal assistance because it was more likely to reduce difficulty completing activities of daily living (e.g., bathing or toileting).

Research exploring predictors of assistive technology use among people with disabilities documents important disparities based on socioeconomic status (i.e., education and income) and race/ethnicity (Kaye, Yeager, & Reed, 2008). Controlling for need factors (severity of the impairment), disadvantaged individuals are less likely to use assistive technology, which may perpetuate disparities in health and well-being among people with disabilities. These disparities in assistive technology use may, in part, be a consequence of lower levels of life course mastery. Individuals with a history of high life course mastery may be better equipped to manage health shocks such as a loss of physical functioning and actively pursue resources and options, including seeking out, obtaining, and using assistive technology, that permit the maintenance of independence even in the face of functional decline.

Summary of hypotheses

We employ Mirowsky and Ross’ (1998) human capital hypothesis and an expanded version of Andersen’s (1968) behavioral model to investigate whether mastery histories over the life course shape mobility device use among older adults. Because mastery in older ages reflects the culmination of life experiences (Pearlin et al., 2007), we hypothesize (H1) that older adults with marginalized statuses, representing exposure to structural constraints, such as those with low socioeconomic status (i.e., low education attainment and low incomes), racial and ethnic minorities, and women will belong to latent class trajectories representing lower mastery over their adult lives. We anticipate (H2) that older adults with persistently higher life course mastery will be less likely to use mobility devices because they have less need (i.e., fewer functional limitations). However, we also anticipate (H3) that among older adults with functional limitations, those with higher levels of life course mastery will be more likely to use mobility devices compared with their peers with low life course mastery, net of other social and health factors. In other words, we speculate that high life course mastery will act as a moderator and increase the probability of mobility device use among those with functional limitations (need factor), net of their current socioeconomic status (enabling factor) as well as their past social and health histories.

Method

Data

Data were obtained from the Americans’ Changing Lives (ACL) survey (House, Kessler, & Herzog, 1990; House, Lantz, & Herd, 2005; House et al., 1994), a cohort longitudinal study based on a stratified, multistage area probability sample of noninstitutionalized adults aged 25 and older, living in the coterminous United States, and followed over a 25-year period. African Americans and adults aged 60 and older each were oversampled at twice the rate of non-African Americans and those aged 25–59, respectively. The first wave of the survey was conducted in 1986 with 3,617 adults (68% sample response rate for individuals or 70% for households). Surviving respondents were reinterviewed in 1989 (Wave 2; N = 2,867, 83% of survivors), in 1994 (Wave 3; N = 2,562, 83% of survivors including 164 proxy respondents), in 2001/2002 (Wave 4; N = 1,787, 74% of survivors including 95 proxies,), and again in 2011/2012 (Wave 5; N = 1,427, 81% of survivors including 108 proxies). Sampling weights for nonresponse as well as a poststratification adjustment to the 1986 Census estimates of the U.S. population aged 25 years and older, make the ACL sample representative of the age, gender, and race distribution of the U.S. population living in the United States in 1986, and except for differences due to post-1986 immigration and out-migration, representative of this cohort of Americans as they aged over this 25-year period (House et al., 2005). Because questions on the use of assistive technology for mobility were only asked in the fifth wave of the survey, we restrict our analyses to the 1,427 respondents in Wave 5. However, we utilize data from these respondents that were collected over the entire 25 years of the ACL study.

Measures

The key variable of interest, mastery, was measured using three-items from the Pearlin Mastery Scale, which captures the extent to which an individual feels control over life outcomes (Pearlin & Schooler, 1978). At each wave respondents were asked to indicate the extent to which they agree with the following statements: “There is really no way I can solve some of the problems I have”; “I can do just about anything I really set my mind to do” (reverse scored); and “Sometimes I feel that I am being pushed around in life.” Response options were “agree strongly,” “agree somewhat,” disagree somewhat,” and “disagree strongly,” coded from 1 to 4, and averaged across all three items so that a high mean score represents high mastery.

In terms of questions about the use of mobility devices, respondents were asked if, in the last month, they had used equipment or devices, such as a cane, walker, or wheelchair to get around more easily, safely, or on their own. We created a dichotomous variable to indicate whether respondents did or did not use assistive technology for mobility in the past month. (Due to small cell sizes, it was not possible to examine the use of different types of mobility devices.)

The key independent variable relates to the need for using a mobility device, which we capture using a measure of functional limitations. Functional status was assessed with five questions about difficulty bathing, climbing stairs, walking several blocks, doing heavy work around the house, and whether the respondent was in a bed or a chair for most of the day because of his/her health. ACL derived a four-level functional health index (range 1–4) at each wave to capture severe limitations (in bed/chair most of the day/has a lot of difficulty or cannot bathe self), moderate limitations (has a lot of difficulty or cannot climb a few flights of stairs or walk several blocks), minor limitations (has a lot of difficulty or cannot do heavy work around the house), and no limitations (House et al., 1994).

Because the ACL functional health index takes into account levels of severity, even minor limitations (i.e., a lot of difficulty or cannot do heavy work around the house) reflect relatively serious impairment that would be indicative of need for assistance. Therefore, for analytic purposes, we created a binary indicator contrasting those reporting any limitations in physical functioning at Wave 5 with those reporting no limitations, as the key independent variable predicting device use at Wave 5. In addition, because we cannot account for a history of device use prior to Wave 5, which is likely to affect both life course mastery and functional limitations, we also control for an individual’s functional health history by creating a mean score on the functional health index (range 1–4) across Waves 1–4.

In order to better isolate the effects of life course mastery on mobility device use in the face of later-life functional limitations, we adjust for key covariates that could either account for any observed relationship between mastery and device use (i.e., age, gender, race/ethnicity, socioeconomic status, and physical health conditions) or could be operating in parallel with life course mastery (i.e., mental health histories, histories of social support) rendering any observed association spurious (see conceptual model in Supplementary Figure 1). Age at baseline was measured in years. Gender was modeled using a dummy variable coded 1 for female and 0 for male; similarly, race/ethnicity was coded as 1 for racial/ethnic minority (non-Hispanic Black, Hispanic, and other race/ethnicity [e.g., Asian and Native American]) and 0 for White respondents.

Socioeconomic status was captured through measures of education and income. Education, which tends to be completed by early adulthood, is a time-invariant variable categorized in 1986 as less than high school (0–11 years of completed education), high school diploma (12–15 years of education), and college degree or higher (16 or more years of education). Household income was assessed at each wave. Due to item nonresponse on the income questions, we used imputed income values provided in the ACL data that were generated using the sequential regression imputation method in IVEware (Raghunathan, Solenberger, & Van Hoewyk, 2002). We capture income at Wave 5 (as a covariate) and at Wave 1 (to predict latent class membership, see Supplementary Figure 1) using three income categories: less than $10,000 per year, $10,000–$29,999 per year, and $30,000 or higher (inflation adjusted to 1986 dollars across waves). We also controlled for the type of health insurance respondents reported at Wave 5 (no health insurance, public insurance only [e.g., Medicare, Medicaid, VA, or military health care], and any private insurance) and their physical health status at Wave 5 based on the sum of the number of medically diagnosed chronic health conditions (e.g., heart disease, diabetes, cancer, arthritis, hypertension, stroke, emphysema).

We account for histories of depressive symptoms to disentangle trajectories of mastery from trajectories of mental health over adulthood. Depressive symptoms were assessed with a short form (seven items) of the Center for Epidemiologic Studies Depression Scale (CES-D; Kohout, Berkman, Evans & Cornoni-Huntley, 1993; Radloff, 1977). Three items measure depressive affect (felt depressed, sad, lonely) and four items tap somatic symptoms (everything was an effort, sleep was restless, did not feel like eating, could not get going). For each item, respondents were asked to indicate how often they experienced each symptom during the past week, using a three-category set of response options (hardly ever, some of the time, and most of the time). Responses were averaged to produce an index of depressive symptoms ranging from 1 to 3 for each wave. Mental health histories were captured using a mean of the depression scores across Waves 1–4.

High levels of mastery over adulthood could also be confounded with social support over adulthood. We therefore account for social support histories using an index of social integration based on two questions asked at each wave: “In a typical week, how often do you get together with friends, neighbors or relatives and do things like go out together or visit in each other’s homes?” and “About how many times do you talk on the telephone with friends, neighbors or relatives?”, each assessed with a six-category set of response options (never, less than once a month, about once a month, two to three times per month, once a week, more than once a week). A mean social integration index was calculated based on an average of the two items standardized. We created a measure of social integration histories based on a mean of the social integration index across Waves 1–4. We also account for marital status at Wave 5 (married/partnered vs separated/divorced, widowed, or never married).

Statistical Analyses

We used generalized growth mixture modeling to identify latent classes of individuals according to their trajectories of mastery over adulthood. Growth mixture modeling is an extension of conventional growth modeling that relaxes the assumption of a single population trajectory. By using latent trajectory classes (categorical latent variables), the growth mixture model allows different classes of individuals to vary around different mean growth curves (Muthén and Muthén, 2000). We expect that some individuals will experience consistently high mastery of their adult lives, whereas others (particularly those who experience social and economic disadvantage) will experience persistently low mastery over adulthood. Rather than considering only a single population trajectory of mastery over adulthood, the growth mixture model allows us to explicitly identify these multiple underlying trajectories of mastery, which we expect will have very different effects on device use in later life.

The measurement part of the model captures the growth factors (intercept and slope) as measured by the four indicators of mastery over time (Supplementary Figure 1). Age was used as the indicator of time creating a synthetic cohort from ages 25 through 100. In order to facilitate parameter interpretation, we centered age at the initial point of data collection (setting age 25–0). The structural part of the model incorporates the growth model within a larger latent variable model by relating the growth factors to other observed and latent variables (Supplementary Figure 1). Of particular interest is the latent mastery trajectory class variable, which represents the unobserved subpopulation of life course mastery for respondents. This allows a separate mastery trajectory for each of the latent classes. Time-invariant sociodemographic covariates predict class membership in a logistic regression. As a general extension of the growth mixture model, we include a distal outcome (Muthén, 2004), use of mobility devices at Wave 5 (age ≥50), and examine the relationship between Wave 5 functional limitations and mobility device use as it varies across the latent classes of life course mastery, controlling for sociodemographic characteristics (age, gender, race/ethnicity, education, income, health insurance), as well as functional health histories, mental health histories, and histories of social integration.

Model building proceeded in a sequential process by first specifying the growth model for mastery (using data from Waves 1–4) and then incrementally increasing the number of latent classes. Although substantively based theory is used as the primary means to determine the optimal number of classes, good fitting models are also characterized by (a) a low value for the Bayesian information criterion (BIC) and Akaike information criterion (AIC); (b) a statistically significant (low p value) Lo–Mendell–Rubin (LMR) likelihood ratio test; and (c) distinct posterior probabilities for individual class membership. Mobility device use at Wave 5 is then regressed on functional limitations at Wave 5 (including covariates) within each latent trajectory class of mastery. This essentially captures the interaction effect between life course mastery and Wave 5 functional limitations as it relates to device use at Wave 5.

All models were estimated in Mplus Version 7.4 using full-information maximum likelihood with robust standard errors. Multiple random starts were used to minimize local optima in the likelihood. Respondent-level weights were used to adjust for unequal selection probabilities in the ACL study. Our statistical model allows for respondents with as little as one observation to enter the model. Additionally, by including variables related to attrition (age, education, income, health status), maximum likelihood produces unbiased coefficients under the assumption that the attrition process is conditional on observed variables in our models (Cnaan, Laird, & Slasor, 1997; Feng, Silverstein, Giarrusso, McArdle, & Bengtson, 2006; McArdle & Hamagami, 1992).

Results

Table 1 describes the characteristics of the sample, weighted to account for the sampling design and attrition over time. Roughly half of the respondents were female and one-fifth racial/ethnic minorities. Three quarters had less than a college education and half had annual household incomes less than $30,000 at baseline (1986). Average levels of mastery were quite high over the 25 years of the study (scores more than 3 on the 4-point scale). At Wave 5 (2011; when all subjects were aged 50 and older), the prevalence of functional limitations was 25%, and 14% of respondents reported using an assistive device for mobility. On average, respondents had at least one chronic health condition, and the vast majority of respondents had some form of health insurance (public or private) at this stage of adulthood (Wave 5). Mean depression history scores were close to 0 and mean functional health history scores were close to 4, indicating that most of these study respondents were free of depressive symptoms and functional limitations in the earlier waves of the study.

Table 1.

Weighted Percents and Means for Study Sample Characteristics (N = 1,427)

Variable	Weighted mean (SD) or percent
Gender
Male	46.14
Female	53.86
Race/ethnicity
Hispanic or non-Hispanic Black	20.01
White	79.99
Education
Less than high school	14.41
High school diploma	60.61
College degree	24.98
Income in Wave 1 (1986)
Less than $10,000	11.20
$10,000–29,999	38.85
$30,000 or more	49.95
Mastery
Baseline/Wave 1 (1986)	3.26 (.36)
Wave 2 (1989)	3.51 (.39)
Wave 3 (1994)	3.30 (.41)
Wave 4 (2001)	3.22 (.51)
Wave 5 (2011)	3.09 (.63)
Functional limitations in 2011
None	74.51
Any	25.49
Use of a mobility device in 2011
Yes	13.67
No	86.33
Income in 2011a
Less than $10,000	18.06
$10,000–29,999	40.41
$30,000 or more	41.53
Health insurance in 2011
Only public	19.78
Private	74.95
No insurance	5.27
Number of chronic health conditions in 2011	1.43 (1.21)
Marital status in 2011
Married/partnered	63.16
Not married	36.84
Mental health history (mean CES-D Waves 1–4)	0.39 (0.30)
Social integration history (mean Waves 1–4)	0.00 (0.71)
Functional health history (mean Waves 1–4)	3.87 (0.36)

Note: CES-D = Center for Epidemiologic Studies Depression Scale.

^aInflation adjusted to 1986 dollars.

Table 2 reports the results and fit statistics for a systematic progression of linear growth mixture models. The first column presents the results for the single-class model. At age 25 (intercept), the predicted mastery score is 3.3, reflecting moderately high levels of mastery at this age. However, levels of mastery decline steadily with age (slope = −0.022, p < .05). For each 10-year increase in age, predicted levels of mastery decline by 0.2 on a scale from 1 to 4. A quadratic term was introduced in the model but did not result in an improvement in model fit.

Table 2.

Generalized Growth Mixture Model Regression Coefficients for Trajectories of Life Course Mastery: Americans’ Changing Lives Study (1986–2001)

	Single-class model	Two-class model
Variable	Overall	Class 1 (77%)	Class 2 (23%)
		High life course mastery	Low life course mastery
Intercept	3.300***	3.490***	2.789***
Slope (age)	−0.022*	−0.028*	−0.026
Goodness of fit statistics	BIC = 10,712.74	BIC = 10,678.78
	AIC = 10,654.85	AIC = 10,594.57

Note: AIC = Akaike information criterion; BIC = Bayesian information criterion.

Trajectories are based on mastery scores from ACL Waves 1–4 (1986–2001).

*p < .05. ***p < .001 (two-tailed tests).

The second and third columns of Table 2 report the results for the two-class solution. The change in the BIC and AIC values, coupled with a significant LMR likelihood ratio test (p < .001, not shown), suggest that a two-class solution is preferable to a single-class model. Membership in each class showed good classification quality with individuals most likely to belong to their predicted class (posterior probability is markedly higher [>.90] than for the other class). Adding a third class (model not shown) did not result in any improvement in model fit, and the posterior probabilities did not differentiate class membership well.

Supplementary Figure 2 illustrates the estimated growth curves of life course mastery according to the two-class solution. The two curves represent distinct trajectories of mastery over adulthood: Class 1 (with 77% of the sample) represents the majority of the respondents, which we term the “high life course mastery” class, with a high mastery score at age 25 (intercept = 3.49) and a gradual decline in mastery with increasing age (slope = −0.03). Conversely, individuals in Class 2 (23% of the sample, termed the “low life course mastery” class) have considerably lower mastery scores in early adulthood (2.80 at age 25) and remain persistently low over adulthood (slope not significantly different from zero).

The next step in the modeling process adds the baseline covariates to the model to identify who is most likely to be in each of the two latent mastery classes. Table 3 reports the results from the logistic regression for class membership (adjusted odds ratios and 95% confidence intervals [CI]) using the low mastery class as the reference class. Compared with individuals with low life course mastery, individuals in the high life course mastery class had a higher socioeconomic position at baseline and were less likely to be female. Compared with men, women had almost 50% lower odds of being in the high life course mastery class (odds ratio = 0.55, 95% CI = 0.34, 0.90), all other things being equal (Table 3). Adults with a college degree had more than double the odds of being in the high mastery group (odds ratio = 2.43, 95% CI = 1.10, 5.36) compared with those with less than a high school education. Similarly, higher income levels at baseline were associated with greater odds of being in the high mastery class. Compared with low income adults, adults with moderate income had a twofold higher odds of being in the high mastery class (odds ratio = 2.29, 95% CI = 1.19, 4.40), and those with high income were over fourfold more likely to be in the high life course mastery class (odds ratio = 4.32, 95% CI = 2.03, 9.19).

Table 3.

Logistic Regression Predicting Latent Class Membership for Trajectories of Life Course Mastery: Americans’ Changing Lives Study (1986–2011; N = 1,427)

	Latent Class 1a (High life course mastery)
Variable	Coefficient	OR (95% CI)
Baseline age cohortb	−0.027	0.97 (0.92, 1.04)
Femalec	−0.594*	0.55 (0.34, 0.90)
Hispanic or non- Hispanic Blackd	−0.563	0.57 (0.28, 1.14)
High school diplomae	0.388	1.47 (0.80, 2.70)
College degreee	0.888*	2.43 (1.10, 5.36)
Moderate income ($10,000–30,000)f	0.829*	2.29 (1.19, 4.40)
High income (>$30,000)f	1.464***	4.32 (2.03, 9.19)

Note: CI = confidence interval; OR = adjusted odds ratio.

^aLatent Class 2 (low life course mastery) is the reference class. ^bRefers to age in 1986. ^cReference group is male. ^dReference group is White. ^eReference group is less than high school education. ^fReference group is low income (<$10,000) in Wave 1 (1986).

*p < .05. ***p < .001 (two-tailed tests).

The final step in the modeling process adds the key outcome variable of interest, mobility device use, as a distal outcome in the growth mixture model and examines how the odds of device use vary by prior histories of life course mastery (reflected by latent class membership). Table 4 presents results for logistic models regressing mobility device use on class membership, adjusting for health, socioeconomic, and sociodemographic controls. Odds ratios and 95% CI are presented. The unadjusted Model A indicates that adults with a history of high life course mastery have a lower odds of using mobility devices after the age of 50 (odds ratio = 0.10), but after controlling for functional limitations (Model B), there is no significant difference in the odds of device use according to histories of life course mastery. Functional limitations accounts for the unadjusted negative relationship between high mastery and device use, suggesting that individuals with high levels of mastery are less likely to use devices because they tend to be free of functional limitations.

Table 4.

Logistic Regression Odds Ratios for Mobility Device Use in 2011 (1986–2011): Americans’ Changing Lives Study

	Life course mastery	+ Current functional limitations	+ Interaction	+ Controls
	Model A	Model B	Model C	Model D
Variable	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Intercept	0.67 (0.41, 1.10)	0.02 (0.01, 0.05)	0.06 (0.02, 0.14)	0.06 (0.04, 0.80)
Life course mastery
High life course masterya	0.10 (0.02, 0.17)	1.40 (0.74, 14.49)	0.36 (0.17, 2.20)	2.23 (0.72, 6.86)
Current functional limitations
Any limitations in function (FL) in 2011b		30.61 (18.55, 50.52)	8.66 (2.42, 31.04)	6.04 (1.90, 19.19)
FL × high life course mastery			5.84 (1.27, 26.73)	4.44 (1.10, 17.84)
Baseline sociodemographic controls
Baseline age cohortc				1.04 (1.02, 1.07)
Femaled				0.52 (0.30, 0.88)
Hispanic or non-Hispanic Blacke				0.81 (0.49, 1.34)
High school diplomaf				0.71 (0.41, 1.24)
College degreef				0.55 (0.23, 1.30)
Current (Wave 5) health and socioeconomic controls
Number of chronic health conditions				1.48 (1.23, 1.78)
Moderate income ($10,000–30,000)g				0.53 (0.29, 0.96)
High income (>$30,000)g				0.33 (0.14, 0.75)
No health insuranceh				1.36 (0.47, 3.98)
Public health insurance onlyh				0.93 (0.54, 1.62)
Marital status				0.75 (0.44, 1.26)
Controls for health and social support history Waves 1–4
Mean CES-D score Waves 1–4				0.70 (0.31, 1.60)
Mean social integration score Waves 1–4				0.92 (0.67, 1.26)
Mean functional health score Waves 1–4				0.81 (0.52, 1.26)

Note: HS = high school; CES-D = Center for Epidemiologic Studies Depression Scale; CI = confidence interval; OR = odds ratio. Cells in bold indicate statistically significant effects (p < .05).

^aReference group is low life course mastery. ^bReference group is no functional limitations in 2011 (Wave 5). ^cRefers to age in 1986. ^dReference group is male. ^eReference group is White. ^fReference group is less than high school education. ^gReference group is low income (<$10,000) in 2011 (inflation adjusted to 1986 dollars). ^hReference group is private health insurance in 2011.

However, the averaged effect of functional limitations in Model B masks important differences in the effect of functional limitations on device use across adults with different histories of life course mastery. The interaction effect tested in Model C indicates that in the presence of functional limitations older adults with a history of high life course mastery have a fivefold higher odds of using assistive technology for mobility than those with a history of low life course mastery (odds ratio = 5.84, 95% CI = 1.27, 26.73, Model C, Table 4). This moderating effect remains significant after controlling for individual covariates, including physical health conditions, mental health and functional health histories, social integration histories, sociodemographic characteristics, current income, and health insurance (Model D, Table 4). (Sensitivity analyses [not shown] adjusting for concurrent levels of mastery at Wave 5 also had no impact on the significance of this interaction effect.) Predicted probabilities of device use (Supplementary Figure 3) were generated using the coefficients from Model D (Table 4), for a 60-year-old White male, with less than a high school education, with public health insurance, low income, no depression history, average social integration history, and no history of functional limitations. In the face of later-life functional limitations, respondents with a history of high life course mastery had a predicted probability of .70 of using a mobility device, whereas those respondents with a history of low life course mastery had a probability of .48.

Discussion

Drawing from Mirowsky and Ross’ (1998) psychological human capital hypothesis, we hypothesized (H1) that older adults with marginalized statuses, often reflecting increased exposure to structural constraints over the life course, would be more likely to belong to latent class trajectories representing lower mastery over their adult lives. We found general support for this hypothesis. Women had decreased odds of membership in the high life course mastery class, whereas respondents with college degrees and higher incomes had increased odds of membership in the high life course mastery class; however, we did not see significant differences in class membership for race and ethnicity. We further hypothesized (H2) that because of the positive association between mastery and health, a history of high life course mastery would lower the odds of mobility device because there would be less need (i.e., fewer functional limitations) among those with high mastery. Again, we found support for this hypothesis. Respondents in the high life course mastery class had a 90% reduction in the odds of using mobility device in 2011.

Finally, we hypothesized (H3) that among older adults with functional limitations, those with higher levels of life course mastery will be more likely to use mobility devices compared with their peers with low life course mastery. We documented support for this hypothesis; we observed a significant interaction between functional limitations and life course mastery. Even after controlling for current socioeconomic status (enabling factors) and histories of health and social support, high life course mastery continued to be a significant moderator. Using predicted probabilities, we observed stark differences in mobility device use between the low mastery and high mastery classes.

Using generalized growth mixture modeling, our results provide evidence of distinct experiences of mastery over adulthood. Most individuals report high levels of mastery in early adulthood, which declines steadily with age. However, a less common but empirically distinct trajectory captures individuals who have persistently lower levels of mastery throughout all of adulthood. Women and adults with fewer socioeconomic resources were more likely to be in this lower mastery class, consistent with the notion that these are the groups experiencing a greater burden of stressful life events over their lives (Pearlin et al., 2007). These findings highlight the heterogeneity in mastery that is masked when considering only a single population trajectory.

We also found that these distinct mastery histories had important consequences for the use of assistive technology for mobility in the face of later-life functional limitations. Thus, histories of self-concepts and sense of control have long-term consequences for empowering the decisions to manage functional declines in later life. Compared with older adults with a life history of lower mastery, respondents with a history of high life course mastery had fourfold higher odds of using mobility-related assistive technology in the face of functional limitations. This association was robust against key social, economic, and health factors.

These results suggest that individuals with a history of high life course mastery may be better equipped to consider using assistive technology as an adaptation strategy in the face of functional health declines. A history of high life course mastery may represent a deeper well of psychological human capital that individuals can draw from despite declining health. For adults without persistent histories of high life course mastery, there may be a tipping point or some threshold where individuals become fatalistic about their health situation and may be less able to consider resources and adaptive strategies to cope with declining function. On the other hand, high levels of life course mastery may enable older adults to preserve feelings of control and confidence in their ability to manage their own health, including obtaining and using assistive technology for mobility.

Although these findings draw attention to the importance of life course mastery for mobility device use, we also observed important disparities in mobility device use even after considering mastery histories. All things being equal, woman and those with low incomes had lower odds of mobility device use. These disparities highlight the need to contextualize sources of psychological human capital such as mastery within larger structural conditions. As featured in the expanded Anderson model (Bradley et al., 2002), psychosocial factors are a key pathway linking need and enabling factors to use of health services; however, we must recognize that the health advantages conferred by psychosocial factors are contingent on need and, in particular, enabling factors such as socioeconomic status and social support.

Strengths and Limitations

This research uniquely contributes to the current literature by empirically assessing life course mastery using 25 years of prospective data collected over the entire adult life course and employing growth mixture models to identify different latent trajectories of mastery over adulthood. To date, the concept of life course mastery has been used in a limited fashion with nearly no empirical research with the exception of Pearlin and colleagues (2007) and, to our knowledge, no research using prospective (versus retrospective) measures of life course mastery. Furthermore, Pearlin and colleagues’ (2007) seminal paper examines the predictors of life course mastery, whereas we empirically test whether life course mastery has consequences for health-promoting behaviors (i.e., mobility device use) in mid to later life.

Although this research has key strengths, it is also subjected to limitations. For example, mobility-related assistive technology use was only assessed in the last wave. The prior waves did not have any measures on the use of mobility devices. The use of assistive technology over adulthood may influence both health and mastery levels in later life. The strongest predictor of mobility device use is physical health and functioning (i.e., need factors); therefore, we included an index of functional health histories over adulthood in an attempt to capture the impact of these unmeasured factors in our models. Although the ACL was well suited for these analyses because of the long observation period over the full adult life course, and consistent mastery measures across all waves, the generalizability of our results is limited to those in the original sampling frame who survived to the fifth wave. There have been important demographic shifts—especially related to immigration since 1986—that are not captured in this population. Although sampling weights adjust for attrition over the course of the study, individuals surviving to Wave 5 are likely to be in better physical health with greater socioeconomic resources than those who died before then. As a consequence, our results may underestimate the impact of functional limitations on device use and may also underestimate the prevalence of the low life course mastery trajectory class if these were more typical of the attrited subjects. Nonetheless, our findings provide evidence of the salience of life course mastery for the health and well-being of older adults as it shapes their decision or ability to use assistive technologies in the face of declining function.

Conclusion

Recognizing that mastery in older ages reflects individual histories and biographies has broad policy implications. Our research highlights the need for a life course approach to health promotion that includes policies that encourage high levels of mastery from early ages. Exposing children and young adults to environments that embolden feelings of confidence and perceptions of control may have long-lasting impact on health and health behaviors in older ages. Policymakers should recognize that life course mastery is a source of psychological human capital that can be fostered and molded over a lifetime. Giving individuals in early life the opportunities to achieve status attainment in home, school, or work environments and exposing them to mastery experiences may promote high levels of mastery throughout their lives. Early investments in fostering life course mastery, particularly in socially disadvantaged groups, may have larger gains for population health in the long term.

This should not take away from policies and interventions that seek to enhance current mastery or sense of control among older adults, but it is a push to acknowledge that psychological human capital like other types of capital may accumulate over time. Yet, even after considering mastery as a source of psychological human capital, important disparities in mobility device use persisted. Low life course mastery women and older adults with low incomes may be doubly disadvantage when it comes to adopting mobility-related assistive technology. Interventions aimed at addressing both enabling factors and psychosocial factors among older adults with functional limitation may be the most effective. No amount of perceived mastery will overcome a lack of access to mobility devices and health services, but increasing current mastery and access may improve the health and well-being of older adults with functional limitation.

Supplementary Material

Supplementary data is available at The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences online.

Funding

This research was supported by Grant #R01AG018418 from the National Institute on Aging. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the NIA or NIH.

Conflict of Interest

The authors have no conflict of interests to declare.