The Influence of Evidence-Based Exercise and Age Reattribution on Physical Function in Hispanic Older Adults: Results From the ¡Caminemos! Randomized Controlled Trial

Abstract

Older Hispanics routinely exhibit unhealthy beliefs about “normal” aging trajectories, particularly related to exercise and physical function. We evaluated the prospective effects of age reattribution on physical function in older Hispanics. Participants (n = 565, ≥60 years) were randomly assigned into (a) treatment group—attribution-retraining, or (b) control group—health education. Each group separately engaged in four weekly 1-hr group discussions and 1-hr exercise classes, followed by monthly maintenance sessions. The Short Physical Performance Battery (SPPB) measured physical function throughout the 24-month intervention. No significant difference in physical function between intervention arms was evident over time. However, both groups experienced significant improvements in physical function at 24 months (β = 0.43, 95% confidence interval [CI] = [0.16, 0.70]). Participating in the exercise intervention was associated with improvements in physical function, although no additional gains were apparent for age attribution-retraining. Future research should consider strengthening or modifying intervention content for age reattribution or dosage received.

Keywords

Hispanics exercise older adults physical function age reattribution

Introduction

Sitting has been labeled as the new smoking and for good reason (Slachta, 2017). Diaz and colleagues (2017) describe that total amounts of sedentary time (i.e., sitting or lying down) and extended periods of physical inactivity are associated with all-cause mortality. Chronic physical inactivity, characterized by less than 20 min of physical activity three times per week (U.S. Department of Health and Human Services, Centers for Disease Control and Prevention [CDC], 2016), also accelerates reductions in physical function in older adults (Booth et al., 2011). Impaired physical function elevates the risk for chronic illness incidence and medical comorbidities that increase dependence on others and reduce the quality of one’s life (Booth et al., 2011; Guralnik et al., 1994, 1995). According to Painter et al. (1999), physical function can be defined as basic activities (i.e., those which require mobility, strength, and endurance) that are essential to maintaining an individual’s independence. Unfortunately, many adults do not follow the recommended guidelines for physical activity, and far too many sit for prolonged periods of time (Dunstan et al., 2012).

Hispanic older adults tend to be among the least active population in the United States (Larsen et al., 2015). Investigators have found that among adults aged 50 years or older, 33% of Hispanics (compared with 26% of non-Hispanic Whites) did not meet the recommended physical activity guidelines that prescribe 150 min of moderate-intensity exercise per week (U.S. Department of Health and Human Services, CDC, 2016). These high levels of physical inactivity greatly increase the risk of developing chronic illnesses (Booth et al., 2011). Compared with Whites, Hispanics show a greater prevalence of major risk factors related to cardiovascular disease (Daviglus et al., 2012). They have high rates of diabetes (U.S. Department of Health and Human Services, CDC, 2011) and obesity (Flegal et al., 2016), which can markedly impede physical functioning (Angel et al., 2015). Over time, these conditions lead to diminished physical functioning and amplify the need to rely on others to perform the basic tasks of everyday life (Guralnik et al., 1995).

Although empirical evidence indicates that formal participation in exercise can deter age-related declines in physical function (Belza et al., 2006; Bird et al., 2011; Pahor et al., 2006; Shubert et al., 2018), many Hispanic older adults refrain from exercise due to multiple factors, such as neighborhood safety, gender norms, diminishing social networks, culturally appropriate exercise regimens, and cost of gym membership (Larsen et al., 2015). In addition, few exercise programs have been adapted for Hispanic older adults (Jurkowski et al., 2010; Larsen et al., 2015).

The ¡Caminemos! program is an exception when it comes to tailoring for Hispanic older adults. For this program, investigators culturally adapted an evidence-based exercise intervention (EnhanceFitness®; Belza et al., 2006) and examined its effects on 572 community-dwelling Hispanic older adults. They also evaluated the possible benefits of a training program that taught a randomized sample to modify unhealthy expectations of the aging process (i.e., age attribution-retraining). Futhermore, Hispanic older adults have been found to believe that low physical function and chronic disease are the unavoidable results of the “normal” aging process (Goodwin et al., 1999). However, this belief can diminish motivation to remain physically active and foster physical inactivity that exacerbates medical comorbidities and disabilities (Sarkisian et al., 2005).

According to Weiner’s (1985) attribution theory, people’s perceptions of successes and failures can be classified along three dimensions: (a) whether they experience an internal or external locus of causality, (b) whether they see their action as fixed or changeable, and (c) whether they believe they can exert change over the outcome. Attribution theory postulates that outcomes perceived as uncontrollable are specifically damaging to motivation. This notion is supported by findings of Courneya et al. (2004) in a post-analysis of a 10-week exercise program. Courneya and colleagues (2004) found that perceived success predicted more healthful quality of life and greater exercise adherence, and that perceived success interacted with personal control. Because attribution theory posits that behavior change is more likely to occur when people believe that actions related to the targeted outcomes are feasible and within their control (Sarkisian et al., 2005), the ¡Caminemos! investigators initially hypothesized that to reap the most benefits of an exercise program, one should change preexisting negative attributions of aging and physical activity.

However, previous studies using the ¡Caminemos! data reveal the complex nature of how attribution might influence behavior. The ¡Caminemos! study exposed a sample of older adults to a structured exercise program. Half the group was randomly assigned to a treatment condition which taught that aging does not necessarily coincide with being sedentary and physical declines; instead, such outcomes can be forestalled by being more active (Hernandez et al., 2019; Piedra et al., 2017, 2018). The control group received general health information. Following the 4-week intervention, participants in the treatment group displayed significant improvements (p = .05) in their age expectation scores at 12 months (as per the Expectations Regarding Aging [ERA-12] survey; Piedra et al., 2018). Further analysis of the ¡Caminemos! data, however, did not support theorized benefits of age reattribution as no differences were found in physical activity levels after 2 years. When investigators evaluated cognitive function, they found no significant intergroup difference at either the 1- or 2-year mark; both groups showed significant improvements compared with baseline (Hernandez et al., 2019; Piedra et al., 2017, 2018). Despite mixed results, investigators observed a consistent, albeit nonsignificant, trend; the treatment group performed slightly better than controls across all measures over time. In this current study, we examined how the intervention affected physical function and whether we would observe a similar pattern. Given the earlier findings from the ¡Caminemos! trial, we hypothesized that both arms of the intervention would exhibit an improvement in physical functioning compared with baseline and an attenuated (possibly nonsignificant) improvement for the treatment group (Hernandez et al., 2019; Piedra et al., 2017, 2018).

Method

Study Population and Data Source

The data set for the current study derives from a randomized controlled trial, conducted between August 2005 and August 2009, which enrolled a total of 572 older Hispanic adults (clinicaltrials.gov identifier: NCT00183014; Piedra et al., 2018). Specifics of the ¡Caminemos! intervention have been published elsewhere (Hernandez et al., 2019; Piedra et al., 2017, 2018). Briefly, participants were recruited and enrolled from 27 community-based senior centers throughout the greater Los Angeles area. Potential participants had to meet the following criteria to be eligible: (a) ≥60 years of age; (b) self-identify as Hispanic; (c) be able to communicate verbally in English or Spanish; (d) answer four of six items correct on the cognitive impairment six-item screener (Callahan et al., 2002); (e) be physically able to walk (with or without the use of assistive devices); (f) report being physically inactive, which was defined as exercising less than 20 min three times per week; and (g) be able to attend weekly instructions and exercise classes held at their respective senior centers. Physician clearance was obtained for each potential participant to ensure they did not have any medical contraindications to participate in this study.

Eligible participants were randomly assigned to one of two arms, as follows: (a) those who received an attribution-retraining curriculum (treatment group), and (b) those who received general health education (control group). Participants of the treatment and control groups received separate 1-hr group discussion sessions for 4 consecutive weeks. The bilingual health educator delivered either the attribution-retraining (treatment group) curricula or generic health education, according to group assignment. The generic health education group received topics on senior wellness.

Participants assigned to the treatment group received an attribution-retraining curriculum developed by a multidisciplinary team of investigators (see Table 1 for details). The curriculum has been described in detail elsewhere (Piedra et al., 2018). Briefly, participants were taught that becoming sedentary should not be viewed as an inevitable part of aging and that health problems associated with aging should be attributed to mutable factors rather than “old age.” The trained health educator taught participants how to change their attributions of old age from those that are immutable to those that are mutable.

Table 1.

Attribution-Retraining Intervention—Curriculum Objectives.

Session	Objectives
1	Introduce the idea that being physically active should be part of normal aging and should continue at any age.
	Identify barriers of being less physically active.
	Differentiate between modifiable and unmodifiable causes.
	Teach that aging itself does not cause decreased physical activity.
	Make individual promises to increase walking and physical activity.
2	Review promises and problem-solve on barriers to completion.
	Generate feedback about the promises.
	Reinforce that being physically active should be part of normal aging and should continue at any age.
	Reinforce the difference between modifiable and unmodifiable contributors to physical inactivity.
	Identify common changes associated with aging and teach that with modifications one can resume an active life.
	Make new promises.
3	Review promises and problem-solve on barriers to completion.
	Reflect on whether expectations and beliefs about aging have changed.
	Reinforce key concepts.
	Teach that being unable to learn a new habit is not caused by aging.
	Problem-solve on how to maintain an exercise or walking plan.
	Make new promises.
4	Review promises and problem-solve on barriers to completion.
	Reflect on whether expectations and beliefs about aging have changed.
	Reinforce key concepts.
	Identify good things about getting older.
	Problem-solve (further) on how to maintain an exercise or walking plan.

Participants of both arms of the intervention also participated in group exercise classes, taught by a certified instructor, which met once per week for 1 hr over the 4-week period. The training targeted domains of muscle strength, endurance, balance, and flexibility. The exercise classes were a modified version of the EnhanceFitness® Program (previously called the Lifetime Fitness Program©) administered by Senior Services (Seattle) and designed to be safe for seniors with a wide range of abilities, so that the exercises could be done either standing or in a seated (chair) position (Belza et al., 2006).

After the 4-week core program, participants were asked to follow up monthly for 11 additional months of maintenance sessions, and then every 2 months for the following 12 months (total of 24 months of the intervention). The maintenance sessions included both 1-hr group discussion (either attribution-retraining or generic health education) and 1-hr exercise classes. Approval for the study was obtained through the University of California Los Angeles Institutional Review Board with all enrolled participants providing written informed consent.

Figure 1 presents a consort diagram of participant recruitment and study design. The final analytic sample for this study included a total of 565 participants after the exclusion of seven participants with missing data at baseline, specifically on marital status (n = 2), physical functioning (n = 2), and body mass index (BMI; n = 3).

Figure 1.

CONSORT flow diagram.

Measures

Self-report survey data collected included sociodemographic factors, behavioral practices, psychosocial states and traits, and health-related information. Measurement instruments that demonstrated reliability and validity in their published forms were used in the Spanish language appropriate for Hispanics of Mexican and Central American origin. Instruments were translated by a certified translator if there was no previously tested Spanish version of the instruments.

Outcome Variable

Physical function

The primary outcome was physical function as measured by the Short Physical Performance Battery (SPPB). This measure tests lower extremity function, and it has been well-validated as a measure of leg strength, which is strongly associated with mobility in older adults (Mijnarends et al., 2013). The SPPB consists of three assessments: (a) repeated chair stands (sit-to-stands), (b) gait speed, and (c) standing balance. For the repeated chair stands, participants were asked to quickly stand up and sit down, consecutively, up to five times. Time to completion was recorded to the nearest second. Gait speed was assessed by timing participants as they walked an 8-ft course at the usual speed on two separate occasions. They were allowed the use of assistive devices if needed; the faster of two walks was used. To test standing balance, participants were asked to stand in one of three separate foot positions (feet together, semi-tandem, and full tandem) for up to 10 s. If a participant moved their feet or grasped for support prior to 10 s, the test was concluded and that time was recorded. If the participant was able to balance for 10 s in a given foot position, they proceeded to the next position. Each participant began in the feet-together position followed by the semi-tandem (heel of one foot placed by the big toe of the other foot) and, finally, full tandem (feet directly in front of each other; Guralnik et al., 1994).

Each assessment is scored from 0 (poor) to 4 (best) and the sum yields a total SPPB score that ranges from 0 (poor) to 12 (best; Guralnik et al., 1994). Lower SPPB total scores are associated with greater disability (Guralnik et al., 1994).

Covariates

Sociodemographic factors

The sociodemographic factors consisted of age, sex (male vs. female), education (no school, ≤eighth grade, or some high school or greater), income (less than US$20,000, greater than US$20,000, or not reported), and marital status (never married, married, separated/divorced, or widowed).

Health-related measures

Measured BMI was calculated as weight (kg)/height (m)² and classified as underweight (<18.5), normal (18.5–24.9), overweight (25–29.9), or obese (30 or higher). Due to the very few underweight participants (n = 4), we combined them with the normal weight group in the analyses. Participants self-identified the presence of 16 medical conditions, using a revised version of the Charlson comorbidity index (Katz et al., 1996) to capture the prevalence of medical comorbidities. These included any of the following conditions: (a) high blood pressure; (b) heart attack; (c) congestive heart failure; (d) stroke; (e) diabetes; (f) arthritis; (g) hip fracture; (h) fracture of wrist, arm, or spine; (i) lung disease; (j) liver disease; (k) cancer; (l) Parkinson’s disease; (m) coronary artery bypass surgery; (n) Alzheimer’s disease or dementia; (o) depression; or (p) anxiety. An additive score was created, indicating the number of medical morbidities experienced. The total score was treated as a continuous measure, with scores ranging from 0 to 16.

Data Analysis

We calculated descriptive statistics for sociodemographic factors, health status, acculturation levels, and physical function data at baseline for those who were in the treatment and control groups. Results for the continuous variables were reported as the mean and standard deviation, whereas results for categorical variables considered the count and percentages in each category. Group differences between the intervention and control groups were tested using chi-square tests (categorical variables) and t tests (continuous variables). Differences in retention between the treatment and control groups were calculated using a log-rank test. Two sample t tests were used to compare the continuous variables over time between the treatment and control groups.

To test the primary hypothesis concerning the effects on physical function by intervention condition over time, we conducted a repeated mixed-effects linear regression (McCulloch & Searle, 2000; Verbeke, 1997) to determine intervention effects, that is, control versus intervention condition, on longitudinal changes in physical function. The multivariate mixed-effects models included the following variables: group (general health education or age reattribution), time in months (Time point; Baseline, 1, 12, 24 months), interaction of intervention group and time (Group × Time point), and terms capturing baseline age, sex, education, annual household income, marital status, BMI, and total number of medical comorbidities. Random effects of the intercept were included to allow subjects to vary in their scores for physical function from baseline. We tested whether to include random slopes to account for variability in the rate of change in physical function, but results based on R² and the Bayesian and Akaike information criteria indicated no improvement in model fit. For the regression models, we present the regression coefficient, confidence intervals (CIs), and p values. To facilitate the interpretation of regression results, particularly the interaction effects, we examined the linear predictions obtained with the “margins” command and the contrasts, involving factor variables and their interactions, using the “contrast” command. We use the “marginsplot” command to graph the influence of the intervention on physical function over time.

Finally, we conducted numerous sensitivity tests. A dichotomous variable was added to indicate those with complete data on all waves versus those with missing data for at least one follow-up. Second, we conducted mixed-effects regressions, which included this dummy variable for incomplete versus complete data. Finally, we performed a sensitivity analysis using multiple imputation procedures to further evaluate the impact of missing data. All data analyses were conducted in 2018–2019, using statistical software (Stata, SE 15.1).

Results

Baseline Characteristics of the Study Sample

A total of 565 subjects were included in the analyses, with 275 randomly assigned to the attribution-retraining condition and 290 to the generic health education arm. Table 2 displays the baseline distributions of sociodemographic variables and health status for the entire sample and stratified by trial arm, that is, intervention versus control condition. Baseline characteristics of the entire sample were as follows: M_age = 73 years (SD = 6.80), 77% were females, 59% had less than an eighth-grade education, 76% had income below US$20,000, 13% were never married, and 58% were either separated/divorced or widowed; 84% had a BMI indicating they were either overweight or obese and, on average, sample participants reported more than two medical comorbidities (M = 2.6, SD = 1.64). The results showed no significant group differences for baseline sociodemographic variables between the two trial arms.

Table 2.

Baseline Descriptive Statistics for Sociodemographic Factors.

Characteristics	Total (N = 565)	Control (N = 290)	Intervention (N = 275)	p value
Age, M (SD)	73.1 (6.8)	73.2 (6.8)	73.1 (6.7)	.83
Female, n (%)	435 (77.0)	233 (80.3)	202 (73.5)	.05
Education, n (%)				.22
No schooling completed	83 (14.7)	49 (16.9)	34 (12.4)
≤Eighth grade	251 (44.4)	130 (44.8)	121 (44.0)
Some high school or greater	231 (40.9)	111 (38.3)	120 (43.6)
Income, n (%)				.91
Less than US$20,000	428 (75.8)	219 (75.5)	209 (76.0)
US$20,000 or greater	89 (15.8)	45 (15.5)	44 (16.0)
Not reported	48 (8.5)	26 (9.0)	22 (8.0)
Marital status, n (%)				.28
Never married	72 (12.7)	41 (14.1)	31 (11.3)
Married	163 (28.9)	76 (26.2)	87 (31.6)
Separated/divorced	126 (22.3)	61 (21.0)	65 (23.6)
Widowed	204 (36.1)	112 (38.6)	92 (33.5)
BMI (kg/m²), n (%)				.43
Underweight and normal	92 (16.3)	46 (15.9)	46 (16.7)
Overweight	212 (37.5)	105 (36.2)	107 (38.9)
Obese	261 (46.2)	139 (47.9)	122 (44.4)
Number of medical comorbidities, M (SD)	2.6 (1.6)	2.7 (1.8)	2.6 (1.5)	.34

Note. BMI = body mass index.

Boldface indicates statistical significance: p < .05.

Change in Physical Function Over Time

Figure 2 displays how the average SPPB total scores improved throughout the study for the control and treatment groups. At baseline, both groups had an average SPPB total score below 10 (treatment 8.05, control 8.12), which indicates moderate functional impairment in the lower extremities (Guralnik et al., 1995). As the study progressed, participants in the control group steadily improved, whereas those in the intervention group improved after 1 month and then slightly declined over the next 11 months. However, by 24 months, they had rebounded and ended with higher average SPPB total scores (8.85) than the control group (8.65). Both groups showed significant improvements at 12 and 24 months, relative to baseline. However, we found no significant differences between groups.

Figure 2.

Predictive margins of SPPB total scores with 95% confidence interval.

Table 3 documents the longitudinal changes in total SPPB, along with associated subscales (chair stand, gait speed, and balance), for both the treatment and control groups. Similar to the total SPPB scores, we found no statistically significant difference between the groups across subscales. However, we observed a trend in which those in the intervention group displayed greater scores in total SPPB, chair stand, and gait speed than the control group. Balance, however, remained stable over time.

Table 3.

Mean Scores for Physical Function During the 24-Month Trial.

Component	Baseline		1 month		12 months		24 months
Component	Control	Intervention	Control	Intervention	Control	Intervention	Control	Intervention
Chair stands	1.85 (1.09)	1.76 (1.04)	1.90 (1.16)	2.00 (1.14)	2.04 (1.23)	1.96 (1.16)	2.30 (1.27)	2.40 (1.32)
Gait speed	2.75 (0.97)	2.72 (1.00)	2.80 (0.98)	2.78 (1.04)	2.77 (0.98)	2.85 (0.95)	2.80 (0.95)	2.91 (0.91)
Balance	3.52 (0.87)	3.57 (0.81)	3.60 (0.83)	3.61 (0.90)	3.63 (0.82)	3.47 (0.98)	3.49 (0.99)	3.48 (0.93)
SPPB total	8.12 (2.03)	8.05 (2.04)	8.33 (2.17)	8.43 (2.28)	8.43 (2.24)	8.33 (2.26)	8.65 (2.42)	8.85 (2.37)

Note. Value: Mean ± (SD). SPPB = Short Physical Performance Battery.

Comparison of Trial Conditions Over Time

Table 4 displays the results of a mixed-effects linear regression, which tests the effects of the exercise program and the age reattribution intervention on longitudinal changes in physical function. After controlling for age, sex, education level, marital status, income, BMI, and number of medical comorbidities, we found that both groups experienced improvements throughout the entirety of the intervention, although there were no group differences. Specifically, participants across both arms of the intervention showed significant improvements from baseline to 12 and 24 months in the chair stand (12 months; β = 0.18, 95% CI = [0.02, 0.33] and 24 months; β = 0.42, 95% CI = [0.26, 0.58]) and their SPPB totals (12 months; β = 0.27, 95% CI = [0.01, 0.53] and 24 months; β = 0.43, 95% CI = [0.16, 0.70]).

Table 4.

Mixed-Effects Linear Regression on Physical Function.

Fixed effects	Balance		Gait speed		Chair stands		SPPB total
Fixed effects	β	95% CI	β	95% CI	β	95% CI	β	95% CI
Intervention (ref = control)	0.03	[−0.12, 0.17]	−0.08	[−0.23, 0.07]	−0.12	[−0.31, 0.07]	−0.18	[−0.52, 0.16]
Month (ref = baseline)
1	0.09	[−0.03, 0.20]	0.05	[−0.07, 0.17]	0.05	[−0.09, 0.19]	0.21	[−0.03, 0.45]
12	0.09	[−0.04, 0.22]	0.01	[−0.11, 0.14]	0.18 *	[0.02, 0.33]	0.27 *	[0.01, 0.53]
24	−0.06	[−0.19, 0.07]	0.03	[−0.11, 0.16]	0.42 ***	[0.26, 0.58]	0.43 **	[0.16, 0.70]
Group × Month
Intervention × 1	−0.05	[−0.22, 0.12]	0.01	[−0.16, 0.18]	0.19	[−0.02, 0.39]	0.16	[−0.19, 0.51]
Intervention × 12	−0.18	[−0.36, 0.01]	0.11	[−0.07, 0.30]	0.03	[−0.19, 0.25]	0.01	[−0.36, 0.39]
Intervention × 24	−0.05	[−0.24, 0.14]	0.15	[−0.04, 0.34]	0.19	[−0.04, 0.41]	0.3	[−0.08, 0.69]
Age	−0.03 ***	[−0.04, −0.02]	−0.03 ***	[−0.04, −0.02]	−0.02 ***	[−0.03, −0.01]	−0.08 ***	[−0.10, −0.06]
Female (ref = male)	0.06	[−0.08, 0.19]	−0.19 *	[−0.33, −0.04]	−0.01	[−0.20, 0.18]	−0.15	[−0.51, 0.20]
Education (ref = no schooling)
≤ ighth grade	0.12	[−0.04, 0.27]	0.18 *	[0.00, 0.35]	−0.04	[−0.27, 0.18]	0.27	[−0.16, 0.69]
Some high school or greater	0.18 *	[0.01, 0.34]	0.36 ***	[0.17, 0.54]	0.1	[−0.13, 0.33]	0.64 **	[0.20, 1.08]
Marital status (ref = never married)
Married	0.03	[−0.15, 0.20]	0.06	[−0.14, 0.26]	0	[−0.26, 0.25]	0.07	[−0.41, 0.56]
Separated/divorced	−0.01	[−0.19, 0.17]	0.15	[−0.05, 0.36]	0.1	[−0.16, 0.37]	0.24	[−0.25, 0.73]
Widowed	−0.08	[−0.24, 0.09]	0.07	[−0.13, 0.26]	−0.04	[−0.29, 0.20]	−0.07	[−0.53, 0.39]
Income (ref = below US$20,000)
US$20,000 or more	0.09	[−0.06, 0.24]	0.19 *	[0.02, 0.36]	−0.01	[−0.22, 0.21]	0.25	[−0.15, 0.66]
Not reported	0.14	[−0.04, 0.33]	−0.04	[−0.25, 0.18]	−0.15	[−0.42, 0.13]	−0.03	[−0.54, 0.48]
BMI (ref = under or normal weight)
Overweight	−0.05	[−0.18, 0.09]	−0.1	[−0.24, 0.05]	−0.20 *	[−0.38, −0.02]	−0.3	[−0.62, 0.02]
Obese	−0.12	[−0.25, 0.02]	−0.13	[−0.28, 0.01]	−0.31 **	[−0.49, −0.12]	−0.52 **	[−0.85, −0.19]
Medical comorbidities	−0.05 **	[−0.08, −0.02]	−0.09 ***	[−0.12, −0.05]	−0.10 ***	[−0.15, −0.05]	−0.23 ***	[−0.32, −0.14]
Intercept	5.55 ***	[4.89, 6.20]	5.24 ***	[4.50, 5.99]	4.01 ***	[3.07,4.95]	14.65 ***	[12.88, 16.42]
Random effects
Intercept	−0.76 ***	[−0.87, −0.66]	−0.56 ***	[−0.64, −0.47]	−0.29 ***	[−0.37, −0.20]	0.38 ***	[0.30, 0.46]
Residual	−0.33 ***	[−0.37, −0.30]	−0.34 ***	[−0.38, −0.30]	−0.17 ***	[−0.21, −0.13]	0.36 ***	[0.32, 0.40]

Note. SPPB = Short Physical Performance Battery; CI = confidence interval; BMI = body mass index.

Boldface indicates statistical significance: *p < .05. **p < .01. ***p < .001.

As expected, we found that increasing age and a greater number of comorbid conditions were associated with diminished functioning across all SPPB subsets. Furthermore, overweight and obese participants had lower scores for chair stands; obese participants showed reduced overall functioning. Higher levels of education were associated with higher levels of balance, gait speed, and overall physical function. We found no difference by marital status (see Table 4 for details).

Sensitivity Analysis Considering Missing Data

Given the presence of missing data at follow-up, we explored the nonresponse patterns. We compared cases with complete data (n = 332, 58%) across all waves with cases in which participants had missed at least one follow-up visit (n = 233, 32%). We found that those who had missing data were more likely to be older (odds ratio [OR] = 1.02, 95% CI = [1.00, 1.06]; p = .032). However, there were no differences by group assignment, sex, education, income, marital status, BMI, or chronic conditions.

Results from mixed-effects regressions that included a dummy variable for complete versus incomplete data indicated that those who had missing data did not differ from those who had complete data on gait speed, but they had lower levels of balance (β = −0.14, 95% CI = [–0.25, –0.04]; p = .009), chair stand (β = −0.18, 95% CI −0.34, –0.03; p = .021), and overall physical functioning (β = −0.35, 95% CI = [–0.64, –0.06]; p = .017). Although gains in overall physical functioning were limited to 24 months, the remaining statistical inferences were unchanged.

To further evaluate the impact of missing data, we performed a sensitivity analysis using multiple imputation. Based on the imputed data, results indicate similar gains in overall physical functioning at 24 months (β = 0.42, 95% CI = [0.09, 0.75], p = .013) than those obtained with no imputation, but differences at 12 months were no longer statistically significant (β = 0.29, 95% CI = [–0.2, 0.59], p = .060). The findings on multiple imputation also confirmed that participants showed improvements from baseline to 24 months for chair stand (12 months; β = 0.19, 95% CI = [0.01, 0.36], p = .035 and 24 months; β = 0.36, 95% CI = [0.18, 0.53], p < .001). The data on multiple imputation confirmed no group differences over time. All other statistical inferences remained the same.

Discussion

This study adds to a growing body of evidence, based on a substantially large sample of community-dwelling urban Hispanic older adults, that structured exercise programs may serve as a health promotion strategy for Hispanic older adults (Hernandez et al., 2019; Piedra et al., 2017, 2018). By completing just 1 hr per week for 4 weeks with additional maintenance sessions (once a month for 11 months followed by once every 2 months for 12 months), participants in both groups had sustained benefits in physical function for up to 24 months. As expected, we found no added benefits for those who received the age attribution-retraining.

These findings resonate with the existing literature that documents improvements in physical function from structured exercise programs (Belza et al., 2006; Bird et al., 2011; Pahor et al., 2006; Shubert et al., 2018). In addition, programs with multiple domains of training (e.g., muscular strengthening, flexibility, balance, and cardiovascular endurance) have been found to significantly improve physical function in older adults (Bird et al., 2011; Pahor et al., 2006)—as evidenced when tested using randomized trial designs. Furthermore, investigators have observed that incorporating maintenance sessions, to facilitate participation, promotes long-term adherence, which further improves physical functioning (Müller-Riemenschneider et al., 2008).

Despite the lack of a control group, our findings are particularly remarkable, given the relative disadvantage of our sample. The majority of the total sample reported low socioeconomic status (75.8% reported an annual income of less than US$20,000) and low education levels (59.1% had less than an eighth-grade education). In addition, our sample was burdened with poor health. On average, participants reported two or more chronic diseases, and nearly all exhibited high BMI levels (84% were categorized as either overweight or obese). Yet, despite levels of impaired health, participants in both arms of the study significantly improved their physical function at 12 and 24 months compared with baseline. And although SPPB score improvements from both arms of the intervention were modest (0.80 points for the treatment group and 0.54 points for the controls), they remain noteworthy. These increases in SPPB fall within a small and a clinically significant improvement (SPPB score improvement of 0.5 and 1.0 point, respectively) in activities of daily living (ADLs; Perera et al., 2006).

Whereas both arms of the intervention improved their physical function from baseline to 24 months, the treatment group had a reduction in SPPB total scores at 12 months. This conflicts with our previous ¡Caminemos! Findings, showing that physical activity (total number of steps) increased during this period in the treatment group (Piedra et al., 2018). Increased physical activity levels have been shown to be indicative of better SPPB performance (Pahor et al., 2006). However, findings by Patel et al. (2014) suggest that improvements in muscular strength have a larger effect on SPPB scores than muscular endurance. Moreover, the initial increases in SPPB scores after 1 month may have been due to neural adaptations, which can increase muscular strength in the early stages of exercise training (Raastad et al., 2010). We speculate that although participants of the treatment group increased their physical activity (total number of steps), these improvements may not have elicited increases in lower body muscular strength, which did not translate into increases in SPPB total scores.

Although our data lack statistical significance between intervention arms, we observed a trend that displayed the intervention group increasing in overall physical function at 24 months over that of the control group. Similar trends were reported by other investigators using the ¡Caminemos! data (Hernandez et al., 2019; Piedra et al., 2017), although these findings contradict what Courneya et al. (2004) found in a post-analysis of a 10-week exercise program, which showed that expectations for success coincided with enhanced quality of life and postexercise adherence. In addition, a study by Wolff and colleagues (2014) found that, compared with exercise alone, older adults exposed to a combination of exercise and “views-on-aging” had a more positive outlook on the aging process, thus increasing their physical activity.

It is possible that the benefits of the exercise program overshadowed the benefits of the intervention. The act of participating in an exercise program may have reinforced the idea that one can remain active as one ages, thus motivating future behavior (Resnick & Nigg, 2003). Cuddy (2015) cites a body of evidence that supports the idea that action precedes cognitive change. However, more research is needed to determine whether these findings remain under differing modes of delivery (e.g., text messages or phone calls).

This intervention has certain limitations. Participants were from the greater Los Angeles area and may not reflect the entire Hispanic population in the United States. This study also only included Hispanic older adults who were initially physically inactive—limiting generalizability among older adults with moderate engagement in exercise. Also, the SPPB has previously displayed ceiling effects and may have prevented further improvement for high functioning participants (Sayers et al., 2006). Finally, we did not conduct any follow-up testing upon the termination of the intervention. Therefore, we cannot definitively conclude whether these physical function improvements were sustained past the 24-month time point.

However, the strengths of this study are noteworthy. Most prominent is the role of a low-cost, practical, and applicable exercise program to promote health. This study demonstrated that an evidence-based exercise program with minimal supervision and timely maintenance sessions was associated with improvements in physical function for up to 24 months. In addition, such a program improved physical function even after controlling for socioeconomic status, BMI status, or presence of preexisting chronic diseases. These findings suggest that, at least among Hispanic older adults, physical function can improve when exercise programs are made available. Future research should include a more diverse Hispanic population. Such studies should also examine whether increases in SPPB total score correlates with improvements in ADLs and quality of life.

Conclusion

In conclusion, this study demonstrated that participation in an evidence-based exercise program is associated with sustained improvements in physical function over a 24-month period for community-dwelling Hispanic older adults. Although trained exercise professionals administered the exercise program, it did not require prolonged supervision. Yet, with minimal supervision, improvements in physical function were evident. Findings from this study may be relevant for senior centers, even those with limited resources, as a way to maintain or improve the physical function and quality of life of their residents. However, more research is needed to fully understand and determine the impact of age reattribution on physical health outcomes.

Footnotes

Acknowledgements

The authors are very grateful to the leadership and staff of the 27 senior centers that collaborated with us on activities of the ¡Caminemos! trial.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research reported in this publication was supported, in part, by the National Institute on Aging of the National Institutes of Health [grant number R01 AG024460-05 to C. Sarkisian]; the National Center for Advanced Translational Sciences [grant number UL1TR001881 to C. Sarkisian]; the National Institute on Aging of the National Institutes of Health [grant number P30AG0215684-16 to C. Sarkisian]; the Midcareer Award in Patient-Oriented Community-Academic Partnered Aging Research [grant number 1K24AG047899-02 to C. Sarkisian]; the National Heart, Lung, and Blood Institute [grant number 1K01HL130712 to R. Hernandez]; and the National Institute of Minority Health and Health Disparities of the National Institutes of Health [grant number U54MD012523 to R. Hernandez]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Research Ethics

This study was approved by the University of California, Los Angeles Institutional Review Board (protocol #12-001599).

ORCID iDs

Brett Burrows

Lissette Piedra

References

Angel

R. J.

Angel

J. L.

Hill

T. D.

(2015). Longer lives, sicker lives? Increased longevity and extended disability among Mexican-origin elders. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 70(4), 639–649. https://doi.org/10.1093/geronb/gbu158

Belza

Shumway-Cook

Phelan

E. A.

(2006). The effects of a community-based exercise program on function and health in older adults: The Enhance Fitness Program. Journal of Applied Gerontology, 25(4), 291–306. https://doi.org/10.1177/0733464806290934

Bird

Hill

Ball

Hetherington

Williams

(2011). The long-term benefits of a multi-component exercise intervention to balance and mobility in healthy older adults. Archives of Gerontology and Geriatrics, 52(2), 211–216. https://doi.org/10.1016/j.archger.2010.03.021

Booth

F. W.

Laye

M. J.

Roberts

M. D.

(2011). Lifetime sedentary living accelerates some aspects of secondary aging. Journal of Applied Physiology, 111, 1497–1504. https://doi.org/10.1152/japplphysiol.00420.2011

Callahan

C. M.

Unverzagt

F. W.

Hui

S. L.

Perkins

A. J.

Hendrie

H. C.

(2002). Six-item screener to identify cognitive impairment among potential subjects for clinical research. Medical Care, 40(9), 771–781. https://doi.org/10.1097/00005650200209000-00007

Courneya

K. S.

Friedenreich

C. M.

Sela

R. A.

Quinney

H. A.

Rhodes

R. E.

Jones

L. W.

(2004). Exercise motivation and adherence in cancer survivors after participation in a randomized controlled trial: An attribution theory perspective. International Journal of Behavioral Medicine, 11(1), 8–17.

Cuddy

(2015). Presence: Bringing your boldest self to your biggest challenges. Little, Brown; Hachette Book.

Daviglus

M. L.

Talavera

G. A.

Avilés-Santa

M. L.

Allison

Cai

Criqui

M. H.

. . . Stamler

(2012). Prevalence of major cardiovascular risk factors and cardiovascular diseases among Hispanic/Latino individuals of diverse backgrounds in the United States. Journal of the American Medical Association, 308(17), 1775–1784. https://doi.org/10.1001/jama.2012.14517

Diaz

K. M.

Howard

V. J.

Hutto

Colabianchi

Vena

J. E.

Safford

M. M.

. . . Hooker

S. P.

(2017). Patterns of sedentary behavior and mortality in U.S. middle-aged and older adults: A national cohort study. Annals of Internal Medicine, 167(7), 465–475. https://doi.org/10.7326/M17-0212

10.

Dunstan

D. W.

Howard

Healy

G. N.

Owen

(2012). Too much sitting—A health hazard. Diabetes Research and Clinical Practice, 97(3), 368–376. https://doi.org/10.1016/j.diabres.2012.05.020

11.

Flegal

K. M.

Kruszon-Moran

Carroll

M. D.

Fryar

C. D.

Ogden

C. L.

(2016). Trends in obesity among adults in the United States, 2005 to 2014. Journal of the American Medical Association, 315(21), 2284–2291. https://doi.org/10.1001/jama.2016.6458

12.

Goodwin

J. S.

Black

S. A.

Satish

(1999). Aging versus disease: The opinions of older Black, Hispanic, and Non-Hispanic White Americans about the causes and treatment of common medical conditions. Journal of the American Geriatrics Society, 47(8), 973–979. https://doi.org/10.1111/j.1532-5415.1999.tb01293.x

13.

Guralnik

J. M.

Ferrucci

Simonsick

E. M.

Salice

M. E.

Wallace

R. B.

(1995). Lower-extremity function in persons over the age of 70 years as a predictor of subsequent disability. The New England Journal of Medicine, 332(9), 556–561. https://doi.org/10.1056/NEJM199503023320902

14.

Guralnik

J. M.

Simonsick

E. M.

Ferrucci

Glynn

R. J.

Berkman

L. F.

Blazer

D. G.

. . . Wallace

R. B.

(1994). A short physical performance battery assessing lower extremity function: Association with self-reported disability and prediction of mortality and nursing home admission. Journal of Gerontology, 49(2), 85–94. https://doi.org/10.1093/geronj/49.2.M85

15.

Hernandez

Andrade

F. C.

Piedra

L. M.

Tabb

K. M.

Sarkisian

(2019). The impact of exercise on depressive symptoms in older Hispanic/Latino adults: Results from the “¡Caminemos!” study. Aging and Mental Health, 23(6), 680–685. https://doi.org/10.1080/13607863.2018.1450833

16.

Jurkowski

J. M.

Mosquera

Ramos

(2010). Selected cultural factors associated with physical activity among Latino women. Women’s Health Issues, 20(3), 219–226. https://doi.org/10.1016/j.whi.2010.01.004

17.

Katz

J. N.

Chang

L. C.

Sangha

Fossel

A. H.

Bates

D. W.

(1996). Can comorbidity be measured by questionnaire rather than medical record review? Medical Care, 34(1), 73–84. https://doi.org/10.1097/00005650-199601000-00006

18.

Larsen

B. A.

Noble

M. L.

Murray

K. E.

Marcus

B. H.

(2015). Physical activity in Latino men and women: Facilitators, barriers, and interventions. American Journal of Lifestyle Medicine, 9(1), 4–30. https://doi.org/10.1177/1559827614521758

19.

McCulloch

C. E.

Searle

S. R.

(2000). Linear mixed models (LMMs). In Shewhart

W. A.

Wilks

S. S.

(Eds.), Generalized, linear, and mixed models (pp. 156–186). Wiley.

20.

Mijnarends

D. M.

Meijers

J. M.

Halfens

R. J.

ter Borg

Luiking

Y. C.

Verlaan

. . . Schols

J. M.

(2013). Validity and reliability of tools to measure muscle mass, strength, and physical performance in community-dwelling older people: A systematic review. Journal of the American Medical Directors Association, 14(3), 170–178. https://doi.org/10.1016/j.jamda.2012.10.009

21.

Müller-Riemenschneider

Reinhold

Nocon

Willich

S. N.

(2008). Long term effectiveness of interventions promoting physical activity: A systematic review. Preventive Medicine, 47(4), 354–368. https://doi.org/10.1016/j.ypmed.2008.07.006

22.

Pahor

Blair

S. N.

Espeland

Fielding

Gill

T. M.

Guralnik

J. M.

. . . Studenski

(2006). Effects of a physical activity intervention on measures of physical performance: Results of the Lifestyle Interventions and Independence for Elders Pilot (LIFE-P) study. Journal of Gerontology: Medical Sciences, 61(11), 1157–1165. https://doi.org/10.1093/gerona/61.11.1157

23.

Painter

Stewart

A. L.

Carey

(1999). Physical functioning: Definitions, measurement, and expectations. Advances in Renal Replacement Therapy, 6(2), 110–123. https://doi.org/10.1016/S1073-4449(99)70028-2

24.

Patel

M. S.

Mohan

Andersson

Y. M.

Baz

Kon

S. S. C.

Canavan

J. L.

. . . Polkey

M. I.

(2014). Phenotypic characteristics associated with reduced short physical performance battery score in chronic obstructive pulmonary disease. Chest, 145(5), 1016–1024. https://doi.org/10.1378/chest.13-1398

25.

Perera

Mody

S. H.

Woodman

R. C.

Studenski

S. A.

(2006). Meaningful change and responsiveness in common physical performance measures in older adults. Journal of the American Geriatrics Society, 54(5), 743–749. https://doi.org/10.1111/j.1532-5415.2006.00701.x

26.

Piedra

L. M.

Andrade

F. C.

Hernandez

Boughton

S. W.

Trejo

Sarkisian

C. A.

(2017). The influence of exercise on cognitive function in older Hispanic/Latino adults: Results from the ¡Caminemos! study. The Gerontologist, 57(6), 1072–1083. https://doi.org/10.1093/geront/gnw256

27.

Piedra

L. M.

Andrade

F. C.

Hernandez

Trejo

Prohaska

T. R.

Sarkisian

C. A.

(2018). Let’s walk! Age reattribution and physical activity among older Hispanic/Latino adults: Results from the ¡Caminemos! randomized trial. BMC Public Health, 18, 964–984. https://doi.org/10.1186/s12889-018-5850-6

28.

Raastad

Refsnes

P. E.

Paulsen

Rønnestad

Wisnes

A. R.

(2010). Strength training: In theory and practice (1st ed.). Gyldendal Akademisk.

29.

Resnick

Nigg

(2003). Testing a theoretical model of exercise behavior for older adults. Nursing Research, 52(2), 80–88.

30.

Sarkisian

C. A.

Prohaska

T. R.

Wong

M. D.

Hirsch

Mangione

C. M.

(2005). The relationship between expectations for aging and physical activity among older adults. Journal of General Internal Medicine, 20(10), 911–915. https://doi.org/10.1111/j.1525-1497.2005.0204.x

31.

Sayers

S. P.

Guralnik

J. M.

Newman

A. B.

Brach

J. S.

Fielding

R. A.

(2006). Concordance and discordance between two measures of lower extremity function: 400 meter self-paced walk and SPPB. Aging Clinical and Experimental Research, 18, 100–106. https://doi.org/10.1007/BF03327424

32.

Shubert

T. E.

Smith

M. L.

Jiang

Ory

M. G.

(2018). Disseminating the Otago Exercise Program in the United States: Perceived and actual physical performance improvements from participants. Journal of Applied Gerontology, 37(1), 79–98. https://doi.org/10.1177/0733464816675422

33.

Slachta

(2017, September 13). “Sitting is the new smoking”: Sedentary behavior linked to increased all-cause mortality. Cardiovascular Business, Lipids & Metabolic. https://www.cardiovascularbusiness.com/topics/lipids-metabolic/sitting-new-smoking-sedentary-behavior-linked-increased-all-cause-mortality

34.

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. (2011). National diabetes fact sheet: National estimates and general information on diabetes and prediabetes in the United States, 2011.

35.

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. (2016). 2014 Behavioral risk factor surveillance system.

36.

Verbeke

(1997). Linear mixed models for longitudinal data. In Verbeke

Molenberghs

(Eds.), Linear mixed models in practice: A SAS-oriented approach (pp. 63–153). Springer.

37.

Weiner

(1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92, 548–573. https://doi.org/10.1037/0033-295X.92.4.548

38.

Wolff

J. K.

Warner

L. M.

Ziegelman

J. P.

Wurm

(2014). What do targeting positive views on ageing add to a physical activity intervention in older adults? Results from a randomised controlled trial. Psychology & Health, 29(8), 915–932.