Change in Views of Aging, Physical Activity, and Physical Health Over 8 Weeks: Results From a Randomized Study

Abstract

The AgingPLUS program targets motivational barriers, including negative views of aging, as mechanisms to increase adult physical activity. A pilot study was conducted to test the efficacy of this new program against a generic successful aging program. Fifty-six participants were randomly assigned to the AgingPLUS group, and 60 participants were assigned to the active control group. Repeated-measures multivariate analyses of variance assessed changes in views of aging, physical activity, blood pressure, and hand-grip strength from pretest (Week 0) to delayed posttest (Week 8). The Condition × Occasion interactions were nonsignificant; however, significant main effects for condition and occasion were found. Follow-up tests showed that views of aging were more positive, and physical activity had significantly increased at Week 8 for all participants. In addition, in the treatment group, elevated blood pressure had significantly decreased and hand-grip strength had significantly increased at Week 8. Despite the nonsignificant multivariate findings, the main effect findings provided partial support for the efficacy of the AgingPLUS program.

Physical activity (PA) engagement has been established as the single most effective nonpharmacological, noninvasive, and cost-effective method to promote healthy aging (Lachman et al., 2018). Despite the many benefits associated with PA, the majority of adults do not engage in the recommended amount of PA (Ashe et al., 2009; Clarke et al., 2017). A lack of motivation and/or poor self-regulation skills (Nielsen & Reiss, 2012), including negative views of aging (VoA), represent specific risk factors that may limit adult PA. The purpose of the current study was to assess the efficacy of the AgingPLUS intervention program, an innovative program that targets motivational barriers to increase PA among middle-aged and older adults.

Benefits of PA in Middle-Aged and Older Adults

Adults aged 50 years and older are considered the most sedentary age group (Harvey et al., 2013) with only 12.7% of older adults meeting the official exercise recommendations of engaging in at least 150 min of moderate-intensity PA per week (Clarke et al., 2017; World Health Organization, 2020). However, engagement in regular PA has consistently been shown to improve health outcomes, cognitive functioning, and psychological well-being in adults (Falck et al., 2019; Nelson et al., 2007; Piercy et al., 2018).

Physical activity has a notable impact on middle-aged and older adults due to the preventative role it plays in several chronic health conditions (Nelson et al., 2007; Piercy et al., 2018). Participation in regular moderate-intensity PA has been shown to decrease the risk of cardiovascular disease, Type 2 diabetes, obesity (Cleven et al., 2020, Nagamatsu et al., 2014; Piercy et al., 2018), stroke (Kramer et al., 2019), osteoporosis, mobility-related disabilities, and numerous cancers (Nelson et al., 2007; Piercy et al., 2018). Evidence from several studies also suggests that PA may serve as both a therapeutic and protective factor for decreasing symptom severity for many of the aforementioned illnesses (Falck et al., 2019; Nelson et al., 2007; Piercy et al., 2018) and reducing fall and injury risk in later life (Hamed et al., 2018; Nelson et al., 2007; Piercy et al., 2018). Although these health benefits are most likely when the minimum guidelines for a physically active lifestyle are met, individuals who do not reach the PA recommendations, but engage in some PA, are still likely to reap some of the benefits (Hupin et al., 2015; Kim et al., 2022; Spartano et al., 2019).

Beyond the physical health benefits, PA also leads to a variety of cognitive and psychological benefits. Engagement in regular PA may delay the onset of neurodegenerative diseases, such as dementia and Alzheimer’s disease (Spartano et al., 2019; Tan et al., 2017), support the maintenance of cognitive functioning (Kramer & Erickson, 2007; Piercy et al., 2018; Spartano et al., 2019), promote hippocampal neurogenesis (Erickson et al., 2011), and lead to better white matter health (Burzynska et al., 2014; Spartano et al., 2019) and increases in gray matter across various brain regions (Erickson et al., 2010; Halloway et al., 2019). In addition, regular PA has been associated with a reduction in depression and anxiety (Aguiñaga et al., 2018; Piercy et al., 2018; Schuch et al., 2016), higher perceived quality of life, and lower psychological distress (Awick et al., 2017; Kell & Rula, 2019; Piercy et al., 2018). Despite PA being associated with these positive health-related outcomes, the percentage of adults engaging in regular PA tends to be relatively low (Clarke et al., 2017).

Negative VoA as a Barrier to PA

Although the benefits of PA are widely known, examining negative VoA as a barrier to engagement in PA is a relatively new research focus. Negative VoA are defined as negative attitudes and beliefs toward the process of aging and older adults as a social group (Brothers & Diehl, 2017; Levy, 2003). These beliefs often revolve around a perceived inevitability and irreversibility of physical and cognitive decline in later life, despite evidence to the contrary (Diehl et al., 2020; Lindland et al., 2015). Nielsen and Reiss (2012), for example, suggested that for older adults, motivational and attitudinal factors, such as negative VoA, may represent key obstacles in adopting and maintaining regular exercise routines.

The harmful effects of negative VoA on adults’ behavior have been shown in both experimental and quasi-experimental research. Specifically, negative VoA have been associated with slower recovery from disabilities (Levy et al., 2012), poorer functional health (Sargent-Cox et al., 2012), decreased vitality (Emile et al., 2015), and shorter life expectancy (Kotter-Grühn & Hess, 2012; Levy et al., 2002; Wurm & Schäfer, 2022). Because older adults with negative VoA also tend to be less physically active, such negative views may indirectly lead to an increased risk for several chronic health conditions (Chalabaev et al., 2013). Although the mechanisms through which negative VoA lead to these adverse outcomes are still being investigated, it is hypothesized that older adults who hold these beliefs internalize and actualize them as self-fulfilling prophecies (Emile et al., 2015; Levy, 2009; Levy & Myers, 2004; Wurm et al., 2013). Emile et al. (2015) also suggested that the difficulty of coping with negative VoA in older adults may deplete mental resources required to engage in health-promoting behaviors (i.e., self-regulation), which, in turn, may lead to heightened susceptibility to illness.

In contrast to the deleterious effects of negative VoA, holding positive VoA has been shown to be associated with a variety of health benefits. The literature consistently notes the association between positive VoA and subsequent PA engagement such that those with positive VoA are more likely to follow a PA routine (Levy et al., 2014; Wolff et al., 2014; Wurm et al., 2010). It is also encouraging that these results emerged following experimental manipulations that made adults’ VoA more positive (i.e., Levy et al., 2014; Wolff et al., 2014). Beyond PA, positive VoA are also linked to increased longevity (Levy et al., 2002; Wurm & Schäfer, 2022), improved functional ability, and better health (Levy, 2009).

Not only are positive VoA associated with various health outcomes, but the previous literature also posits that positive VoA may act as a protective buffer against some health risks in adulthood. For example, positive VoA may lead to higher resiliency, which may then act as a protective factor against future physical functional decline (Sargent-Cox et al., 2012). As such, targeting negative VoA in interventions may be beneficial in dismantling prevalent attitudinal barriers that negatively affect regular PA engagement and in increasing positive VoA, thereby improving engagement in health-promoting behaviors.

Current Study

Based on the positive findings from a previous study showing the feasibility and acceptability of the AgingPLUS program (Brothers & Diehl, 2017), the objective of this study was to investigate the efficacy of the program in a more rigorous way. Specifically, using a randomized single-blind pre- to posttest active control group design with a larger sample of participants, the current study investigated whether participation in the AgingPLUS program resulted in more positive VoA, increased engagement in PA, and improved markers of physical health. It was hypothesized that compared with participants in the active control group, participants in the AgingPLUS program would (a) significantly improve their negative VoA from Weeks 0 to 8, (b) significantly increase their engagement in PA from Weeks 0 to 8, and (c) show significantly greater improvements in physical health indicators from Weeks 0 to 8. Some improvements were also expected in the active control group because participants took part in the exercise program and received a generic educational program. However, their improvements were expected to be significantly lower than the improvements in the AgingPLUS program.

Methods

Participants

Flyers and email announcements were used to recruit study participants from senior centers, local organizations, the university employee pool, and the larger community in a mid-sized town in the front range area of Colorado. Trained research assistants screened interested adults for eligibility. To be enrolled in the study, participants were required to be aged 50–85 years, mostly sedentary (i.e., engaging in less than 60min of PA per week), intent on starting regular PA, and without serious cognitive or other health problems. These criteria were important because the intervention aimed to increase PA in relatively healthy but mostly sedentary older adults. The study was approved by the Colorado State University Institutional Review Board, and all participants reviewed the consent form before providing written informed consent. Participation in the study was completely voluntary, and no monetary or other incentives were provided.

The sample size required for a properly powered pilot study was determined via an a priori power analysis for a repeated-measures multivariate analysis of variance (RM-MANOVA) with a within- and between-subjects interaction (i.e., Condition × Occasion interaction). The program G*POWER (Faul et al., 2007) was used to perform this power analysis. To estimate the proper sample size, the VoA and PA effect sizes obtained in the feasibility study (Brothers & Diehl, 2017) were used as a reasonable expectation. Thus, the input parameters included medium effect sizes of f = 0.30 for multivariate analyses of variance, a significance level of α = .05, and a statistical power of 1–β = 0.80. These assumptions resulted in a required total sample size of 90 participants, with 45 participants in each group. Because we also assumed a 20% attrition rate, a total sample size of 120 participants was determined, with 60 participants in each group.

Based on the inclusion criteria and the sample size calculation, 147 adults were invited to participate in the study, and 116 participants successfully completed the study. Participants ranged in age from 50 to 83 years (M = 63 years, SD = 8 years), were mostly women (79.3%), and the majority identified as White (91.4%). The sample was also fairly educated with an average of 17.5 years of schooling (SD = 2.4 years) and reported being in overall good health based on a scale from 1 to 6 (M = 4.82, SD = 0.84, with 1 = Very poor and 6 = Very Good). Random assignment resulted in 56 participants in the AgingPLUS group and 60 participants in the active control group. Additional sample characteristics are displayed in Table 1.

Table 1.

Sample Characteristics (N = 116)

	M (SD) or number (%)
Age (years)	63.29 (8.07)
Gender
Women	92 (79.3%)
Men	24 (20.7%)
Marital status
Single	4 (3.4%)
Married/partnered	78 (67.3%)
Separated/divorced	26 (22.4%)
Widowed	8 (6.9%)
Race/ethnicity
White	106 (91.4%)
African American	2 (1.7%)
Asian American	1 (0.9%)
Hispanic	6 (5.2%)
Other	1 (0.9%)
Employment status
Employed full or part time	67 (57.8%)
Pursuing a second career	3 (2.6%)
Retired	40 (34.5%)
Unemployed	6 (5.2%)
Education level (years)	17.5 (2.43)
Self-rated health (1 = very poor; 6 = very good)	4.82 (0.84)

Procedure

The study applied a randomized single-blind pre-to posttest active control group design to examine the efficacy of the AgingPLUS program against a generic successful aging program. The intervention was conducted during Weeks 1–4 of the program. Participants’ negative VoA and PA were assessed at Week 0 (baseline), Week 4 (immediate posttest), Week 8 (delayed posttest), and Week 12 (long-term follow-up), and markers of physical health were measured from Weeks 1 to 4 and at Week 8.

Baseline (Week 0)

Eligible adults were randomly assigned to either the AgingPLUS program (i.e., treatment group) or a generic successful aging program (i.e., active control group). After randomization, participants received a baseline questionnaire in the mail and returned the completed questionnaire to their first educational session (i.e., Week 1). At the first session, staff members measured participants’ resting blood pressure (BP) and hand-grip strength and distributed a calibrated pedometer and daily activity log for recording baseline PA for 7 days. Participants returned the pedometer and daily activity log at the start of the second educational session (i.e., Week 2).

Educational Sessions (Weeks 1–4)

The intervention required participants to attend one weekly, 2-hr small-group session for four consecutive weeks (i.e., Weeks 1–4). To avoid any contact between members of the two groups, the groups took place on different weekdays. The 2-hr sessions for both groups followed the same structure: During the first hour, participants took part in a physical exercise session, whereas during the second hour, the educational session took place. At the exercise session, all participants had their BP taken at the beginning and end of the session and participated in structured exercises that covered major areas of adult fitness (i.e., stretching and flexibility, aerobic exercise, strength training). The BP measurement at the end of the physical exercise session was taken to ensure that participants’ BPs had appropriately normalized and that they were safe to leave the exercise facility. After the exercise session, participants walked to a different building and began the 1-hr educational component of their respective programs. A detailed overview of content of the classes is provided in Table 2.

Table 2.

Session Content for the AgingPLUS Intervention and the Control Group

Session	AgingPLUS program	Successful aging program
1	• The nature and effects of negative views of aging • The nature and effects of age stereotypes • Misconceptions/myths about normal aging • Effects of negative self-stereotyping • Immunization against negative self-stereotyping • Homework: Stereotype watch	• Global population aging • Normal versus pathological versus successful aging • What is successful aging? • Reasons why it is meaningful to talk about successful aging? • Successful aging has not only to do with health
2	• What does research tell us about normal aging? • Aging and the plasticity of human behavior • Taking control: It is never too late to make a change • How can we take control? • Homework: Stereotype watch	• Successful aging in the physical domain • The role of lifestyle factors • The role of physical activity • The role of healthy eating • Psychosocial stress and stress management
3	• The benefits of physical exercise • Physical and mental health benefits • How much physical exercise is needed? • What kind of exercise is most beneficial? • Homework: Stereotype watch	• Successful aging in the cognitive domain • What are normal age-related changes in cognition? • The aging of human memory • The aging of human intelligence • Findings from intervention research on cognitive aging
4	• How to start being more physically active? • How to set a goal? • How to pursue and achieve a goal? • How to be physically active in the long run? • Graduation from the Aging PLUS program	• Successful aging and engagement with life • Leading an active, engaged, and meaningful life • The importance of meaningful social relationships • Giving is better than receiving: The many benefits of volunteering • Older adults as a “natural resource” in the community

At the end of the fourth and final exercise session, participants, again, had their BP and hand-grip strength measured and completed the same questionnaire as at Week 0 at the end of the educational session. Participants also were, again, given their calibrated pedometer and a new daily activity log to record their PA for the next week.

Experiential Component (Weeks 5–8)

After completion of the educational sessions, participants entered the experiential portion of the study (i.e., Weeks 5–8). During these weeks, treatment group participants were asked to pursue the individual PA goals they had set for themselves in Week 3. All participants were given daily activity logs to track their PA and received weekly phone calls inquiring about their recent PA. However, only treatment group participants were given the chance to discuss barriers and their perceived progress toward achieving their activity goals during the phone calls.

At Week 7, participants from both groups were asked to wear pedometers and complete a daily activity log for 7 days again. Participants returned at Week 8 with their pedometer and daily activity log, completed the same questionnaire from Weeks 0 to 4, and had their BP and hand-grip strength reassessed (i.e., Week 8 assessment). To conclude their participation in the study, participants completed a final questionnaire, which was mailed to them at Week 12.

Measures

Primary Outcome: VoA

Three different measures were used to assess participants’ VoA: The Awareness of Age-Related Change (AARC) questionnaire (Brothers et al., 2019), the age stereotypes scale (AS; Kornadt & Rothermund, 2011), and a shortened version of the Expectations Regarding Aging (ERA) questionnaire (i.e., Sarkisian et al., 2005). Given that VoA were a primary outcome, AARC, AS, and ERA effect sizes from the feasibility study (i.e., Brothers & Diehl, 2017) were used, in part, to estimate the total sample size for the current pilot study.

The AARC consisted of 50 items that assessed participants’ self-perceptions of aging in five domains: cognitive functioning, lifestyle and engagement, social–emotional and social–cognitive functioning, health and physical functioning, and interpersonal relations. This measure represented two latent factors: (a) awareness of positive age-related changes (i.e., AARC-Gains; “With my increasing age, I realize that … . I pay more attention to my health.”) and (b) awareness of negative age-related changes (i.e., AARC-Losses; “With my increasing age, I realize that … my mental capacity is declining.”). Thus, AARC-Gains and AARC-Losses were used as two separate outcome variables in the current study. Participants answered the items using a 5-point scale (1 = Not at all and 5 = Very much). Internal consistency reliability was good for both AARC-Gains (α = .82) and AARC-Losses (α = .78).

Participants also completed the AS, which assessed domain-specific age stereotypes (Kornadt & Rothermund, 2011). The AS had a total of 27 items with participants indicating their opinion between a negative and positive pole (i.e., “Older people … are lonely and alone” vs. “Older people … are secure and integrated”) across eight domains: family and partnership, physical and mental fitness, health and appearance, friends and acquaintances, work and employment, religion and spirituality, personality and way of living, leisure activities and social or civic commitment, and financial situation. A summary score across all eight domains was used to reflect a single AS score. Reliability of the summary score at Week 0 was excellent (α = .93).

The 12-item shortened ERA questionnaire (Sarkisian et al., 2005) queried participants on how much they expected to experience positive or negative age-related changes (i.e., “When people get older, they need to lower their expectations of how healthy they can be”). Items were answered on a 4-point scale (1 = Definitely true and 4 = Definitely false). The shortened ERA showed good reliability (α = .84).

Primary Outcome: PA

Participants’ PA, which was a primary outcome, was measured in terms of weekly total steps walked, total kilocalorie expenditure, and total distance walked (in miles) using the OMRON HJ-323U pedometer (Omron Corp.). This model was chosen because previous versions of the OMRON HJ pedometer have provided data with high reliability and predictive validity (Giannakidou et al., 2012; Holbrook et al., 2009; Steeves et al., 2011). Participants wore a pedometer on their hip each day for seven consecutive days, with five complete days required for inclusion in the analyses. To obtain accurate measurements of the recorded variables, the pedometer was calibrated using the participant’s stride length, height, and weight. The three indicators of PA had significant test–retest reliability over an 8-week interval: total number of steps (r = .75, p < .001), total kilocalorie expenditure (r = .74, p < .001), and total distance walked (r = .77, p < .001).

Given that PA was a primary outcome, the PA effect size from the feasibility study (i.e., Brothers & Diehl, 2017) was also utilized to determine the pilot study sample size. It is important to note, however, that the feasibility study used a self-report PA measure, whereas the current pilot study utilized an objective measure of PA.

Secondary Outcome: BP

Resting systolic BP (SBP) and diastolic BP (DBP) values from before the exercise session were included in the study as secondary outcomes. SBP and DBP values were assessed as indicators of physical health with American Diagnostic Corporation Prosphyg Aneroid Sphygmomanometer BP cuffs. Based on 2017 American College of Cardiology and American Health Association standards (Greenland & Peterson, 2017), four categories were created to reflect participants’ hypertensive status. Participants were categorized as having either (a) normal BP (SBP < 120 mmHg and DBP < 80 mmHg), (b) elevated BP (SBP 120–129 mmHg and DBP < 80 mmHg), (c) Stage 1 hypertension (SBP 130–139 mmHg or DBP 80–89 mmHg), or (d) Stage 2 hypertension (SBP ≥ 140 mmHg or DBP ≥ 90 mmHg). To accurately assess change in hypertensive status over time, participants with normal BP at baseline were not included in any analyses in which BP was the outcome. DBP had low test–retest reliability over the 8-week interval (r = .18, p = .05), whereas SBP had modest reliability over the 8-week interval (r = .41, p < .001).

Secondary Outcome: Hand-Grip Strength

Participants’ hand-grip strength, which was included as a secondary outcome, was an additional marker of physical health/performance and was measured using the Jamar Plus hydraulic hand dynamometer. Prior to the assessment, experimenters demonstrated the proper way to hold and squeeze the device (i.e., in a standing position with the arm at a 90° elbow flexion) and then allowed participants to practice. Participants were instructed to comfortably grip the device in whichever hand they preferred to start with and then squeeze as hard as possible. After getting the measurements from one hand, participants then switched hands, and measurements were taken from the other hand. The maximum force this device could record was 200 pounds. The instrument recorded the highest force exerted from the participant. Participants completed three trials, and an average score was calculated for the left and right hand-grip strength. Although hand-grip strength is commonly assessed in a person’s dominant hand, and differences in grip strength between dominant and nondominant hand can be as much as 10% (Abe & Loenneke, 2015), participants’ hand dominance was not recorded. This was done because the investigators had decided a priori to assess and analyze grip strength separately for the left and right hand. Thus, differences in grip strength by hand dominance were not a focus of the current study. Both left hand-grip strength (r = .91, p < .001) and right hand-grip strength (r = .92, p < .001) had high test-retest reliability over the 8-week period and were highly correlated (r = .93, p < .001).¹

Statistical Analyses

All analyses were performed in SPSS (version 26). Data were first checked for normality and missingness. Skew and kurtosis levels were examined for all variables, which led to the finding that the PA variables (i.e., weekly total steps, total kilocalorie expenditure, and total distance walked) for Weeks 0 and 8 slightly deviated from normality. Because the skewness values were not extreme and given that RM-MANOVA is robust against deviations from multivariate normality in terms of Type 1 error (Stevens, 2009), we used the nontransformed data for the analyses. Although missing data were uncommon, missing data points were excluded listwise.

Descriptive statistics and bivariate correlations were calculated to provide an initial representation of the data. Four 2 (Condition) × 2 (Occasion) mixed-factors RM-MANOVAs were conducted to assess group differences and change over time. Significant Condition × Occasion interactions were expected in support of the study hypotheses. VoA (AARC-Gains, AARC-Losses, AS, and ERA), PA (total steps, total kilocalories, and total distance), BP (SBP and DBP), and hand-grip strength (left and right) data from Weeks 0 to 8 assessments were used as the outcome variables. Condition (i.e., treatment or active control group) was a between-subjects factor, whereas occasion (i.e., Week 0 pretest and Week 8 posttest) was a within-subjects factor. Along with reporting the F-statistic for the RM-MANOVAs, Wilks’ lambda (Λ), which provides information on group mean differences for a subset of dependent variables (i.e., Λ values range from 0 to 1, with values closer to 0 indicating more variance explained by the independent variables), and partial eta-squared (${\eta }_{\text{p}}^2$), which measures the effect size of independent variables and is the analysis of variance analog to R² in multiple regression (i.e., ${\eta }_{\text{p}}^2$ values range from 0 to 1, with values closer to 1 indicative of higher predictive ability), are provided as well. Follow-up independent samples t-tests were performed if the RM-MANOVA resulted in significant condition multivariate main effects, and paired samples t-tests were conducted to examine significant occasion main effects.

Results

Descriptive Statistics

Bivariate correlations among the main variables are displayed in Table 3. Focusing on associations between condition and the key outcome variables, point-biserial correlations show that Aging-PLUS treatment group participation was positively associated with (a) higher Week 0, but not Week 8, BP and (b) both Weeks 0 and 8 PA. However, condition was not significantly associated with VoA or hand-grip strength.

Table 3.

Correlation Matrix Estimates Between the Primary Study Variables (N = 116)

	Condition	AS	ERA	Gain	Loss	Steps	Distance	Kcals	SBP	DBP	LH grip	RH grip
Condition	—	−.08	.14	−.08	.06	.21*	.26**	.22*	.13	.13	.07	.10
AS	.00	.55 ***	.50***	.24*	−.49***	.14	.13	.11	.04	−.13	.10	.12
ERA	.19*	.49***	.65 ***	.15	−.50***	.14	.12	.11	.01	−.10	.01	.05
Gain	−.08	.19*	.34***	.53 **	−.54***	.07	.05	−.04	−.20*	−.10	−.10	−.06
Loss	−.13	−.29**	−.55***	−.19*	.52 ***	−.12	−.17	−.07	.12	.07	−.13	−.15
Steps	.23*	−.03	.04	−.00	−.23*	.75 ***	.97***	.89***	−.06	−.05	.22*	.22*
Distance	.29**	−.01	.06	−.07	−.26**	.96***	.77 ***	.88***	−.04	−.02	.32***	.32***
Kcals	.20*	−.03	−.06	−.09	−.14	.89***	.88***	.74 ***	.02	.02	.33***	.31***
SBP	.28**	.02	.00	−.12	−.02	−.01	.00	.07	.41 ***	.58***	.05	.08
DBP	.25**	.15	.14	.00	−.06	−.06	−.02	−.01	.47***	.18	.07	.10
LH grip	.02	.01	−.03	−.29**	−.16	.14	.22*	.20*	.08	.20*	.91 ***	.93***
RH grip	.05	.03	−.03	−.28**	−.17	.15	.25*	.23*	.12	.16	.93***	.92 ***

Note. Condition (0 = active control group; 1 = treatment group). Intercorrelations for Week 0 variables are shown in the lower half of the table, whereas intercorrelations for Week 8 variables are shown in the upper half of the table. The main diagonal shown in bold font displays the test–retest reliabilities between Weeks 0 and 8. AS = age stereotypes; ERA = Expectations Regarding Aging; Kcals = kilocalories; SBP = systolic blood pressure; DBP = diastolic blood pressure; LH grip = left hand-grip strength; RH grip = right hand-grip strength.

^*p < .05.

^**p < .01.

^***p < .001.

Repeated-Measures Multivariate Analyses of Variance

Changes in VoA

The RM-MANOVA with VoA variables as dependent variables yielded a nonsignificant Condition × Occasion interaction effect, Wilks’ Λ = .93, multivariate F(4, 105) = 1.95, p = .11, ${\eta }_{\text{p}}^2=\mathrm{.07.}$ This finding indicated that participants perceived that age-related gains and losses, age stereotypes, and ERA changed in similar ways in the treatment and the active control group from pretest to posttest. However, the analysis revealed a significant multivariate main effect of occasion, Wilks’ Λ = .79, multivariate F(4, 105) = 6.86, p < .001, ${\eta }_{\text{p}}^2=.21$, indicating that participants in at least one of the groups experienced significant changes in VoA from pretest to posttest. The multivariate main effect of condition was not significant, Wilks’ Λ = .92, multivariate F(4, 105) = 2.21, p = .07, ${\eta }_{\text{p}}^2=.08$.

Follow-up analyses for the multivariate main effect of occasion indicated that age stereotypes, F(1, 108) = 16.34, p < .001, ${\eta }_{\text{p}}^2=.13$, and ERA, F(1, 108) = 20.26, p < .001, ${\eta }_{\text{p}}^2=.16$, significantly contributed to the effect. Paired samples t-tests performed for the treatment group showed that age stereotypes, t(54) = −2.41, p = .02, d = 0.30, and ERA, t(54) = −2.72, p = .01, d = 0.33, were significantly more positive at the Week 8 posttest compared with baseline. In the active control group, age stereotypes, t(54) = −3.28, p = .002, d = 0.434, and ERA, t(54) = −3.65, p < .001, d = 0.40, were also significantly more positive at the Week 8 posttest compared with baseline (Figure 1).

Figure 1 —

Group means for AS and ERA at baseline and Week 8 (N = 110). AS = age stereotypes; ERA = Expectations Regarding Aging. *Significant difference at .05 alpha level.

Changes in PA

Findings from the RM-MANOVA showed a nonsignificant group Condition × Occasion interaction, Wilks’ Λ = .98, F(3, 109) = 0.88, p = .46, ${\eta }_{\text{p}}^2=.02$, indicating that participants in the treatment group did not show significantly greater changes in PA from Weeks 0 to 8 follow-up compared with participants in the active control group. A significant multivariate main effect of condition, however, was found, Wilks’ Λ = .88, F(3, 109) = 4.97, p = .003, ${\eta }_{\text{p}}^2=.12$, as well as a significant multivariate main effect of occasion, Wilks’ Λ = .87, F(3, 109) = 5.56, p = .001, ${\eta }_{\text{p}}^2=.13$. The two significant multivariate main effects suggested (a) that there were significant differences in the measures of PA between participants in the treatment and the active control group and (b) that participants in at least one of the groups showed significant changes in PA from Weeks 0 to 8 (Figure 2).

Figure 2 —

Group means for total steps, kilocalories, and distance at baseline and Week 8 (N = 113). *Significant difference at .05 alpha level.

Follow-up analyses for the multivariate main effect of condition showed that total steps walked, univariate F(1, 111) = 6.50, p = .01, ${\eta }_{\text{p}}^2=.06$; total kilocalorie expenditure, univariate F(1, 111) = 6.35, p = .01, ${\eta }_{\text{p}}^2=.05$; and total distance walked, univariate F(1, 111) = 10.60, p = .001, ${\eta }_{\text{p}}^2=.09$, contributed to this effect. Independent samples t-tests indicated that participants in the active control group walked significantly fewer total steps at Week 0, t(111) = −2.58, p = .01, d = 0.48, and at the Week 8 posttest, t(111) = −2.22, p = .03, d = 0.42, compared with the participants in the treatment group. For total kilocalorie expenditure, results showed that participants in the active control group had significantly lower kilocalorie expenditure at Week 0, t(111) = −2.27, p = .03, d = 0.42, and at the Week 8 posttest, t(111) = −2.41, p = .02, d = 0.45, compared with participants in the treatment group. For total distance walked, findings indicated that participants in the active control group had walked a significantly shorter distance at Week 0, t(111) = −3.37, p = .001, d = 0.63, and at the Week 8 posttest, t(111) = −2.79, p = .01, d = 0.52, compared with participants in the treatment group.

Follow-up tests for the multivariate main effect of occasion showed that total steps walked, univariate F(1, 111) = 13.44, ${\eta }_{\text{p}}^2=.11$; total kilocalorie expenditure, univariate F(1, 111) = 16.78, ${\eta }_{\text{p}}^2=.13$; and total distance walked, univariate F(1, 111) = 12.80, ${\eta }_{\text{p}}^2=.10$, all ps < .001, contributed to this effect. Paired samples t-tests performed for the treatment group showed that total steps walked, t(54) = −2.39, p = .02, d = 0.23, total kilocalorie expenditure, t(54) = −2.87, p = .01, d = 0.27, and total distance walked, t(54) = −2.18, p = .03, d = 0.19, were significantly higher at the Week 8 posttest compared with Week 0. Similarly, paired samples t-tests performed for the active control group showed that total steps walked, t(57) = −2.87, p = .01, d = 0.31, total kilocalorie expenditure, t(57) = −2.96, p = .004, d = 0.34, and total distance walked, t(57) = −3.02, p = .004, d = 0.33, were significantly higher at the Week 8 posttest compared with Week 0. These findings indicated that both groups improved their engagement in PA in similar ways from baseline to the Week 8 follow-up.

Changes in BP

This analysis included 90 of the 116 study participants because, by definition, participants with normal BP were excluded. Findings from the RM-MANOVA showed a nonsignificant Condition × Occasion interaction, Wilks’ Λ = .98, F(2, 87) = 0.82, p = .44, ${\eta }_{\text{p}}^2=.02$. This indicated that participants in the treatment group did not experience significantly greater changes in SBP or DBP from Weeks 0 to 8 follow-up compared with active control group participants. However, a significant multivariate main effect of condition emerged, Wilks’ Λ = .88, F(2, 87) = 5.90, p = .004, ${\eta }_{\text{p}}^2=.12$, along with a significant multivariate main effect of occasion, Wilks’ Λ = .92, F(2, 87) = 3.80, p = .03, ${\eta }_{\text{p}}^2=.08$.

Follow-up analyses for the multivariate main effect of condition showed that both SBP, univariate F(1, 88) = 7.82, ${\eta }_{\text{p}}^2=.08$, and DBP, univariate F(1, 88) = 8.31, ${\eta }_{\text{p}}^2=.09$, both p’s = .01, contributed to this effect. Despite random assignment independent samples t-tests showed that participants in the active control group had significantly lower Week 0 SBP, t(88) = −2.84, p = .01, d = 0.60, and Week 0 DBP, t(88) = −2.42, p = .02, d = 0.51, but not Week 8 SBP and DBP, compared with the participants in the treatment group.

Follow-up tests for the multivariate main effect of occasion showed that both SBP, univariate F(1, 88) = 5.41, p = .02, ${\eta }_{\text{p}}^2=.06$, and DBP, univariate F(1, 88) = 4.85, p = .03, ${\eta }_{\text{p}}^2=.05$, contributed to this effect. Paired samples t-tests for the treatment group showed that both SBP, t(45) = 2.41, p = .02, d = 0.43, and DBP, t(45) = 2.08, p = .04, d = 0.39, were significantly lower at the Week 8 posttest compared with baseline. In contrast, paired samples t-tests for the active control group showed that neither SPB, t(43) = .79, p = .43, nor DBP, t(43) = 1.15, p = .26, was significantly different at the Week 8 follow-up compared with baseline. Thus, as can be seen in Figure 3, a reduction in SBP and DBP from baseline to Week 8 had occurred in the treatment group but not in the active control group.

Figure 3 —

Group means for SBP and DBP at baseline and Week 8 (N = 90). SBP = systolic blood pressure; DBP = diastolic blood pressure. *Significant difference at .05 alpha level.

Changes in Hand-Grip Strength

Findings from the RM-MANOVA showed a nonsignificant Condition × Occasion interaction, Wilks’ Λ = .98, F(2, 113) = 1.07, p = .35, ${\eta }_{\text{p}}^2=.02$. This finding indicated that participants’ left and right hand-grip strength changed similarly in the treatment and active control group from baseline to Week 8. The analysis, however, revealed a significant multivariate main effect of occasion, Wilks’ Λ = .87, F(2, 113) = 8.13, p < .001, ${\eta }_{\text{p}}^2=.13$. This main effect indicated that participants in at least one of the groups experienced significant changes in hand-grip strength from baseline to Week 8. The multivariate main effect of condition was nonsignificant, Wilks’ Λ = .99, F(2, 113) = 0.70, p = .50, ${\eta }_{\text{p}}^2=.01$.

Follow-up tests for the multivariate main effect of occasion showed that both left hand-grip strength, univariate F(1, 114) = 16.34, p < .001, ${\eta }_{\text{p}}^2=.13$, and right hand-grip strength, univariate F(1, 114) = 9.45, p = .003, ${\eta }_{\text{p}}^2=.08$, contributed to this effect. Paired sample t-tests indicated that treatment group participants’ left hand-grip strength, t(55) = −4.11, p < .001, d = 0.20, and right hand-grip strength, t(55) = −3.51, p < .001, d = 0.15, were significantly higher at the Week 8 posttest compared with baseline. In contrast, analyses for the active control group showed no significant changes from baseline to Week 8 for either left hand-grip strength, t(59) = −1.76, p = .08, or right hand-grip strength, t(59) = −1.18, p = .24 (Figure 4).

Figure 4 —

Group means for left and right hand-grip strength at baseline and Week 8 (N = 116). *Significant difference at .05 alpha level.

Discussion

Building on the promising findings of a feasibility study (Brothers & Diehl, 2017), the purpose of this study was to examine the efficacy of the AgingPLUS program using a randomized pre-to posttest active control group design. It was hypothesized that participants in the AgingPLUS program would show significantly greater improvements in VoA (i.e., primary outcome), increases in PA (i.e., primary outcome), decreases in BP (i.e., secondary outcome), and increases in hand-grip strength (i.e., secondary outcome) from pretest to posttest compared with participants in the active control group. This overall hypothesis was not supported by the results from the RM-MANOVAs—at least at the multivariate level. However, main effect findings and follow-up analyses showed positive changes in VoA, increases in PA engagement, and improvements in BP and hand-grip strength in both groups from pretest to posttest. Given the improvements among all participants, both the generic successful aging and the new AgingPLUS intervention program resulted in the intended behavior changes and without a differential advantage of the AgingPLUS program—particularly in terms of VoA and engagement in PA. These findings are discussed in more detail later. Along with active control group participants’ improvements providing support for the effect of the generic successful aging program, the obtained results for the treatment group provided qualified support for the efficacy of the AgingPLUS program.

Improved VoA

Given the multifaceted nature of VoA and to fully assess the efficacy of the AgingPLUS program, four measures of VoA were used in the current study. Although no significant multivariate interactions emerged, a significant main effect for occasion was found for age stereotypes and ERA, indicating a significant improvement from pretest to posttest. Specifically, participants in both groups showed similar positive changes in age stereotypes and age expectations from baseline to Week 8 (Figure 1).

Coupled with evidence from the larger literature, this occasion effect demonstrates that negative VoA can be changed to become more positive. This finding is meaningful not only because more positive VoA have been shown to act as a protective factor against negative health outcomes (Levy et al., 2014; Wurm et al., 2008) but also because negative VoA are associated with negative health behaviors and outcomes. The literature suggests that having more negative VoA predicts poorer physical health and cognitive health (Kotter-Grühn & Hess, 2012; Levy et al., 2006; Wurm et al., 2010) and reduces the likelihood of engaging in important health-promoting behaviors (Wurm et al., 2010). Certainly, this change in VoA not only has important implications for health but also adds qualified support that VoA can change in a positive way.

Although it was hypothesized that participants in the treatment group would improve their negative VoA more significantly compared with the participants in the active control group, it is still encouraging that those in the AgingPLUS group experienced a positive shift in VoA. One specific aspect that may, in part, explain the lack of a significant interaction between group condition and occasion resides in the shortened educational session duration compared with the previous feasibility study (Brothers & Diehl, 2017). Specifically, to make the weekly meetings more manageable for participants, the AgingPLUS protocol had been condensed from four 2-hr sessions to four 1-hr sessions. That is, the dosage of the treatment in the AgingPLUS had been cut in half. This shortening resulted in less time for hands-on exercises and group discussions, which, in turn, may have impacted the efficacy of the program in a negative way. Thus, the full effect of the AgingPLUS program may not have been implemented in the present study—an aspect that should be taken into consideration when interpreting the study’s findings. A currently ongoing AgingPLUS trial is addressing this issue by eliminating the 1-hr exercise portion and reverting the weekly educational sessions back to 2 hr. The results from this ongoing trial will provide valuable insight into the impact of the full program.

At the same time, differences in delivery style of material and facilitation may have impacted the group. The facilitators who delivered the AgingPLUS intervention had been trained and certified by a master trainer who had been one of the developers of the AgingPLUS program. The successful aging content in the active control group had been delivered by a master trainer who was very motivated to convey a positive message about the opportunities to age successfully. Thus, the master trainer’s positive presentation style may have inadvertently resulted in a stronger motivational effect than is common for a usual control group program. Specifically, control groups in intervention research commonly control only for social contact (Rebok, 2016) and do not deliver substantive content, as was done in the present study. Thus, in combination, the reduced dosage of the AgingPLUS program and the strong delivery of the successful aging curriculum in the active control group may have inadvertently diminished the full effect of the AgingPLUS program, resulting in a less than optimal outcome.

Increased Engagement in PA

To assess the main aim of the AgingPLUS program (i.e., PA change), the current study used three objective indicators of participants’ engagement in PA: total weekly step count, kilocalorie expenditure, and walking distance. Participants in both intervention conditions increased their PA from pretest to posttest, and the magnitude of these increases was similar in both conditions. Based on this finding, only partial support was provided for the hypothesis about the efficacy of the AgingPLUS intervention with regard to increasing engagement in PA.

The lack of significant group differences from the main analysis may be explained, in part, due to using the same structured exercise trainings for both groups. Because all participants engaged in the weekly exercises before the educational sessions and were given the same examples of ways to engage in PA on their own, the similar improvements in PA at posttest may be attributed to this shared experience. Another factor that may have contributed to the group similarities was the relatively short follow-up period in the study. Findings from previous research suggest that interventions can be subject to “sleeper effects,” meaning that the impact of the intervention or training may not be visible immediately after the intervention but may appear after a longer period of time (see Mühlig-Versen et al., 2012). However, due to constraints in the study’s budget and design, a long-term follow-up (i.e., 6 months later) was not possible, and thus, the presence of delayed effects could not be assessed. Nevertheless, it is reassuring that participants in both groups showed a significant increase in PA.

Positive Changes in Indicators of Physical Health and Physical Performance

Hypertension is considered a risk factor that can be modified to decrease the risk of developing serious cardiovascular disease (United States Department of Health and Human Services, 2021a; Yusuf et al., 2020). This study assessed both SBP and DBP as indicators of physical health. Although the multivariate interaction was not significant, follow-up analyses showed significant decreases in SBP and DBP from pretest to posttest for the AgingPLUS group. Similar decreases, however, were not observed in the participants in the active control group. To note, in comparison with those in the AgingPLUS group, participants in the active control group had significantly lower SBP and DBP at the start of the study despite random assignment. This initial group difference could, in part, have contributed to the resulting significant occasion effect found only for AgingPLUS group participants. Altogether, the decrease in SBP and DBP from pretest to posttest suggests that the effects were going in the right direction but were not strong enough to show statistical significance. This finding provides preliminary support for the effects of the AgingPLUS program on indicators of health—a finding that requires further confirmation.

These results are encouraging given the focus of the AgingPLUS intervention and the current evidence suggesting that holding negative VoA can have deleterious effects on BP (Levy et al., 2000, 2008; Weiss, 2018) and can increase the risk of developing cardiovascular disease (Auman et al., 2005; Chalabaev et al., 2013; Levy et al., 2009). These findings are meaningful for at least two reasons: (a) The documented decrease in SBP and DBP was observed 1 month after participants completed the AgingPLUS intervention; thus, it is possible that a longer period of increased engagement in PA may even have a stronger effect on decreasing hypertensive individuals’ BP, resulting in long-term benefits, and (b) the decrease in BP has larger implications for improving general health as well. Specifically, managing and maintaining a healthy BP decreases the risk of developing serious health conditions, such as stroke (United States Department of Health and Human Services, 2021b).

Widely considered a robust measure of physical performance (McGrath et al., 2018), left and right hand-grip strength were also assessed. Similar to the significant changes in SBP and DBP from pretest to posttest, significant changes in hand-grip strength in both hands were observed in the AgingPLUS group but not in the active control group. The increase in grip strength may have emerged due to feelings of self-enhancement after partaking in an intervention aimed at dismantling negative VoA and increasing self-efficacy (Stephan et al., 2013). In the treatment group, the impact of the intervention may have been compounded with the observed increase in PA to produce this effect. Thus, the study hypothesis found some support via these follow-up analyses, but the difference between the treatment group and the active control group from Weeks 0 to 8 (i.e., the interaction effect) was not statistically significant at the multivariate level. Taken together, the results from the BP and hand-grip analyses provided partial support for the efficacy of the AgingPLUS intervention and associated physical health and physical performance benefits.

Limitations

This study has several limitations. First, given the small sample size, the study was underpowered to detect small effects if they existed. Specifically, the statistical power for the interaction effect was consistently insufficient for detecting small effect sizes in each repeated-measures analysis of variance. The lack of statistical power likely helps to explain the diverging findings between the main analyses and follow-up analyses.

In addition, characteristics of the study sample also may explain the lack of significant findings. First, the sample was fairly homogeneous. The majority of participants were women. White, and well educated and reported being in overall good health. Thus, the sample was lacking in diversity. Second, it is possible that the sample included highly motivated participants. Notably, an inclusion criterion for the study was having the intention of starting regular PA engagement. Both factors may have contributed to the possible underestimation of the efficacy of the program.

Conclusion

The AgingPLUS program was created to promote engagement in PA among middle-aged and older adults by changing their negative VoA to become more positive and, hence, improving their motivational outlook. The reported findings of similar improvements in both groups provided qualified support for the efficacy of both AgingPLUS and the generic successful aging program. Nevertheless, the results from the AgingPLUS intervention point in the right direction and may warrant further examination via a properly powered randomized controlled trial. Such a trial should also strive to target a more diverse sample of middle-aged and older adults.