The Aging^PLUS trial: Design of a randomized controlled trial to increase physical activity in middle-aged and older adults

Abstract

Background:

Negative views of aging (NVOA), low self-efficacy beliefs, and poor goal planning skills represent risk factors that undermine adults’ motivation to engage in physical activity (PA). Targeting these three risk factors may motivate adults to become physically active.

Objective:

To assess the efficacy of Aging^PLUS, a 4-week educational program that explicitly targets NVOA, low self-efficacy beliefs, and poor goal planning skills compared to a 4-week health education program. The study also examines the role of NVOA, self-efficacy beliefs, and goal planning as the mechanisms underlying change in PA.

Design:

This randomized controlled trial (RCT) utilizes the experimental medicine approach to assess change in PA as a function of modifying three risk factors. The RCT recruitment target includes 288 mostly sedentary adults ranging in age from 45–75 years.

Methods:

Eligible middle-aged and older adults are recruited through community sources. Participants are randomized to either the Aging^PLUS or the control group. Participants in both groups are enrolled in the trial for eight months total, with four assessment points: Baseline (pre-test), Week 4 (immediate post-test), Week 8 (delayed post-test), and Month 6 (long-term follow-up). The intervention takes place over 4 consecutive weeks with 2-hour sessions each week. PA engagement is the primary outcome variable. Positive changes in NVOA, self-efficacy beliefs, and goal planning are the intervention targets and hypothesized mediators of increases in PA.

Summary:

By utilizing a multi-component approach and targeting a cluster of psychological mechanisms, the Aging^PLUS program implements the experimental medicine approach to health behavior change.

1. Introduction

Most adults do not engage in regular physical activity (PA; ²), and individuals over the age of 50 are the most sedentary segment of the population (33,40). Self-report data suggest only about 20% of all adults, and about 12% of adults over 65, meet the official physical activity recommendations (19). These numbers are the reality despite overwhelming evidence showing that engagement in PA has many health benefits, including improved cardiovascular, respiratory, and musculoskeletal health (24); improved resistance to Type 2 diabetes (94) and cancers (28,36,99); and improved affective and cognitive functioning (24,46,47,78). Given these benefits, the official PA recommendations are that adults engage in moderate-intensity PA (i.e., slight increase in heart rate) for about 30 minutes each on 5 days/week.

Nielsen and Reiss (74) suggested that a lack of motivation and/or ineffective self-regulation may be key causes of an inactive lifestyle. Thus, interventions should address motivational and self-regulation barriers, such as negative views of aging (NVOA), low self-efficacy beliefs, and deficient goal planning skills. These risk factors undermine adults’ behavior and keep them from adopting and maintaining a PA routine (62,74).

NVOA refer to individuals’ negative attitudes and self-perceptions about growing old(er), and about older adults as a group (35). This also includes beliefs that aging is all negative, uncontrollable, and irreversible (35). Longitudinal studies have shown that NVOA predict poorer functional health (13,58,69,112), greater decline in physical and cognitive functioning (45,82,85), slower recovery from disability (59), and earlier mortality (45,58). Although the exact mechanisms by which NVOA contribute to these outcomes are not well understood, there is evidence that processes of negative self-stereotyping are involved (44,57,113) and that adults with NVOA do not engage in health-promoting behaviors (113).

Self-efficacy beliefs are a key factor in health-promoting behavior (4,5). Numerous studies have shown that individuals with a greater sense of self-efficacy are more likely to exercise regularly, eat healthier, have better health, and are less likely to become disabled (26,49,51). Moreover, self-efficacy beliefs developed during a behavioral intervention predicted long-term adherence to an exercise regimen (71). These findings support a reciprocal link between self-efficacy beliefs and health-promoting behavior (48).

Another obstacle to the adoption and maintenance of health-promoting behaviors are individuals’ deficient goal planning skills (4,5,74,89). Adults with deficient goal planning engage in fewer health-promoting behaviors (91) and often give up quickly when facing obstacles in maintaining a behavior change over time (74,84,89).

This trial targets NVOA, low self-efficacy beliefs, and deficient goal planning as a cluster of motivational risk factors (74) that keep middle-aged and older adults from engaging in regular PA. By targeting these factors, the Aging^PLUS trial is innovative and unique for multiple reasons: (1) It focuses on NVOA as a key motivational factor, which is still a relatively unexplored topic; (2) it contributes to the evidence base on the efficacy of motivation-based interventions; (3) it implements the principles of the experimental medicine approach (81; see explanation below); and (4) it has the potential to become a valuable instrument in the public health toolkit (74,84).

2. Methods

The Aging^PLUS trial is funded by the National Institute on Aging (R01 AG051723), National Institutes of Health. Pilot work that generated the preliminary data for the application to the National Institute on Aging was funded by a faculty seed grant from the Colorado School of Public Health at Colorado State University and a pilot grant from the Colorado Clinical and Translaitonal Sciences Institute (UL1 TR002535; National Center for Advancing Translational Sciences, National Institutes of Health). The Institutional Review Board (IRB) of Colorado State University approved all components of the study, and the trial is registered with ClinicalTrials.gov (NCT03299348). All study procedures adhere to the Consolidated Standards of Reporting Trials (CONSORT; ⁶⁸) guidelines to assure the scientific rigor of the trial and the reproducibility of the protocol and the anticipated findings.

2.1. Study aims

The Aging^PLUS trial has three aims: (1) to conduct a randomized controlled trial (RCT) examining the efficacy of the Aging^PLUS program, a 4-week intervention that targets motivational mechanisms to promote adults’ engagement in PA; (2) to test a conceptual model of the mechanisms underlying the intervention effects; and (3) to conduct a 6-month follow-up to examine the long-term effects of the intervention. The trial adopted the guidelines of the National Institutes of Health (NIH) Stage Model of Behavioral Intervention Development (75) and implements the principles of the experimental medicine approach (62,81). Specifically, the experimental medicine approach (81), involves four distinct steps: (1) Identification of an intervention target; (2) development of measures to verify the target; (3) activation of the target through experimentation or intervention; and (4) examination of the degree to which target engagement produced the intended behavior change. The Stage Model of Behavioral Intervention Development and the experimental medicine approach were adopted because they have become the methods of choice for NIH applications that focus on behavior change.

2.2. Setting, recruitment, and inclusion and exclusion criteria

The Aging^PLUS trial takes place at Colorado State University, a land-grant university in the Front Range area of Colorado. Baseline psychological and physical assessments are conducted in one-on-one sessions in the facilities of a human performance and clinical research laboratory. The intervention portion of the trial, which follows the Baseline assessment, is delivered in a classroom setting in a health and medical center building.

Participants are recruited from a two-county area via newspaper announcements, flyers and newsletters of local civic organizations, faith-based organizations, and other community organizations. Middle-aged participants are recruited through the workplace wellness program of the university. Based on prior recruitment experiences, men and members of minority groups are oversampled by 20–30% to ensure proportional representation.

Inclusion criteria are: (1) age 45–75 years; (2) being mostly sedentary (i.e., PA not exceeding 30 minutes of daily moderate-intensity PA for 3 days per week) but having the intention to become physically more active (41,89); (3) English-speaking; (4) willing to be randomized to one of two educational groups¹; (5) willing to take part in the physical fitness tests and wear an accelerometer; (6) willing to commit to follow-up testing (e. g., not moving out of the area during the study period); and (7) physician clearance to take part in a submaximal exercise test and to begin an exercise program.

These criteria were adopted for various reasons. The age range was chosen because adults become increasingly aware of their age during midlife (age 45–50) and NVOA become more self-relevant (23,53). In terms of the upper age limit, adults in their 70s commonly do not yet face the biological vulnerabilities of very old age (3) and still have reserve capacities that become more limited during the fourth age. This age range is also consistent with the ranges reported in PA research with older adults (15,41,46). The PA inclusion criteria are important because the intervention focuses on increasing adults’ engagement in PA. Physician clearance forms are required before participants are enrolled in the study to ensure that their inclusion is safe and without any risk for their health and well-being.

Exclusion criteria are: (1) Indication of cognitive decline defined as more than four errors on the Short Portable Mental Status Questionnaire (77); (2) severe vision and hearing loss as obtained by self-report; (3) serious problems with mobility as obtained by self-report; and (4) a history of neurological, mental, or substance abuse disorders as obtained by self-report. If adults are excluded from the study due to a health concern, the project coordinator encourages the participant to meet with a primary care physician (PCP) for a follow-up.

2.2.1. Research design

The study employs a randomized single-masked pre-test-post-test control group design (see Figure 1). This design was chosen to achieve robust and unbiased results. Reproducibility and transparency are also important criteria, assuring that the study’s methods and analyses can be replicated and extended by other investigators.

Figure 1.

Aging^PLUS Study Design

Upon establishing eligibility via a phone interview, adults complete two Baseline sessions scheduled 2–3 days apart. The first session involves the administration of all psychosocial measures and takes 1 ½ – 2 hrs. A second 2-hour session involves the assessment of physical fitness (i.e., cardiorespiratory fitness) and fitting a waist-mounted ActiGraph accelerometer (ActiGraphs, LLC; Pensacola, FL). Participants are instructed to start wearing the accelerometer for the next 7 days to provide baseline data on their level of PA.

After the Baseline sessions, participants are randomly assigned to either the Aging^PLUS group or the control group. In order to ensure that the treatment groups are balanced by gender and age group, the random assignments are stratified by gender and age group (45–54, 55–64, and 65–75). Within each stratum, varying blocks, sized from 4 to 8, are constructed and randomization is conducted in advance by the study statistician using a random number generator within the SAS software package. Assignment allocations are then placed in sealed, sequentially labeled, opaque envelopes and not opened by research staff until the moment of randomization for each participant occurs. These procedures were adopted to be compliant with the sequence generation, allocation concealment, and recommendations in the 2010 CONSORT Statement (68). Randomization occurs as close as possible to the start of the intervention to minimize participant dropout. Although group facilitators are, by definition, aware of the content of their own educational program, they are blind to the content of the other condition. The affiliated institution’s IRB requires that study participants be informed in the consent form that they will be randomly assigned to one of two possible educational programs. Participants do not know the content or focus of the alternative program. Data collectors are also blind to the content of the treatment or control group. To avoid any contact between facilitators and participants of the treatment and the control group, the groups are conducted on two different days.

The intervention lasts 4 weeks and is followed by an immediate post-test at the end of the last class meeting (i.e., Week 4 assessment). A delayed post-test is performed at Week 8 after participants have practiced their self-chosen PA goal for 4 weeks, and a long-term follow-up is performed 6 months after the immediate post-test (i.e., Month 6 assessment). Between the Week 8 and the Month 6 follow-up, half of the participants in the Aging^PLUS group are randomly assigned to a monitoring condition. Specifically, in one predetermined week in each month, participants in the monitoring condition complete a daily activity log and wear the accelerometer for 7 days. Across the 5-month period, the week of monitoring rotates within participant and across months so that the different weeks within a month are equally represented across the 6-month period. In addition, participants in the monitoring group receive a brief phone call from a staff member on the first and fourth day of the week in which they are monitored. This phone call is a check-in to make sure that participants fill out the daily activity log correctly and wear the accelerometer as instructed. The daily activity logs and accelerometers are returned by the participant at the end of each week. The daily activity log and the script for the phone call were successfully tested during the feasibility study (10). The purpose of the monitoring group is to test if intermittent monitoring facilitates the maintenance of the intervention effects in the Aging^PLUS group. Starting with the Baseline assessment and ending with the Month 6 follow-up, participation in the study lasts a total of 10 months. Participants receive reimbursements for their time and effort at predetermined time points throughout the study up to a total of $280. Participants who do not complete the full study still receive partial reimbursement for their contribution. Table 1 gives an overview of the occasions of assessment and the administered measures.

Table 1.

Timeline of measures administered throughout the RCT

Construct	Measure	Occasions of Assessment	Reference
Primary Outcomes
Physical Activity	CHAMPS	Baseline, Week 8, Month 6	(100)
	Daily Activity Log ActiGraph Accelerometer	Baseline, Week 8, Monitoring Months, Month 6	(10) (ActiGraph, LLC)
		Baseline, Week 8, Monitoring Months, Month 6
Views of Aging	Awareness of Age-Related Change (AARC)	Baseline, Week 4, Week 8, Month 6	(23,39)
			(43)
	Age Stereotypes	Baseline, Week 4, Week 8, Month 6	(86)
	Expectations Regarding Age (ERA)		(107)
		Baseline, Week 4, Week 8, Month 6
	Essentialist Beliefs about Aging Scale (EBA-S)
		Baseline, Week 4, Week 8, Month 6
Secondary Outcomes
Physical Fitness	Submaximal Exercise Test (VO2 Max)	Baseline, Week 8, Month 6	(9,73)
		Baseline, Week 8, Month 6
	Blood Pressure	Baseline, Week 8, Month 6
	Heart Rate	Baseline, Week 8, Month 6	(32)
	Short Physical Performance	Baseline, Week 8, Month 6
	Battery (SPPB)	Baseline, Week 8, Month 6
	Handgrip Strength	Baseline, Week 8, Month 6
	Body Mass Index (BMI)
	Hip and Waist
	Circumference
Implicit Views of Aging	Lexical Decision-Making	Baseline, Week 8, Month 6	(18)
	Task	Baseline, Week 8, Month 6	(31,37,98)
	Implicit Attitudes Test (Age- Specific)
Self-Efficacy Beliefs	General Self-Efficacy Scale	Baseline, Week 4, Week 8, Month 6	(90)
	Self-Regulation Scale		(21)
		Baseline, Week 4, Week 8, Month 6
Goal Planning	Behavioral Intentions	Week 4, Week 8, Month 6	(80,89,96)
	Action and Coping Planning	Week 4, Week 8, Month 6	(80,89,96)
	Action Control	Week 4, Week 8, Month 6	(80,89,96)

Note. All measures were selected because they (a) are well validated and have excellent psychometric properties; (b) have been used in previous studies; (c) are sensitive to intervention-induced change; and (d) do not create undue burden for the participants.

2.2.2. Treatment condition: Aging^PLUS group

The intervention targets the negative effects of NVOA, low self-efficacy beliefs, and deficient goal-planning skills as the motivational mechanisms that may keep adults from engaging in regular PA. This reasoning is supported by the findings from a small number of studies (8,56,87,111) and our own preliminary data (10). Evidence from these studies suggests that adults’ NVOA can be decreased and positive views of aging (PVOA) can be increased via implicit and explicit interventions. There is also evidence showing that adults’ self-efficacy beliefs and goal planning behavior can be improved through interventions (10,87). Additionally, improvements in these psychological mechanisms were significantly associated in all prior studies with either improved physical function (i.e., improved walking speed, gait and balance) or increased PA (i.e., number of daily steps walked).

The Aging^PLUS program consists of 4 weekly 2-hour in-class sessions (total of 8 hours) delivered by a trained facilitator and co-facilitator to a group of 5–15 participants. Table 2 displays the general topics covered in each weekly session. In addition to the presentations, the participants also engage in reflection exercises contained in a workbook, and these individual exercises are followed by brief group discussions guided by the two facilitators.

Table 2.

Session content for the Aging^PLUS intervention and the control group.

Session	AgingPLUS Program	Health Education Program
1	Introduction; age stereotypes and their sources; common misconceptions about aging; negative self-stereotyping; immunization against negative self- stereotyping	Introduction; preventing heart disease; blood pressure levels; hypertension and the associated risks; cholesterol levels and why it matters; modifiable lifestyle factors
2	Review of Week 1; the concept of plasticity; taking control of aging; benefits of physical activity; official recommendations for physical activity to achieve health benefits	Review of Week 1; regulating and testing blood glucose; Type 1 and Type 2 diabetes; risk factors for and prevention of Type 2 diabetes
3	Review of Weeks 1 and 2; basic rules of effective goal planning; defining a physical activity goal; making an action plan; planning for obstacles; keeping track of one’s physical activity; making physical activity a daily habit	Review of Week 2; causes and symptoms of clinical depression; treating clinical depression; benefits of social engagement; cognitive impairment and decline; types of dementia; promoting mental health
4	Review of Weeks 1, 2, and 3; Reflection on practicing the physical activity goal for a week; strategies for long-term success; graduation	Review of Weeks 1, 2, and 3; modifiable and non-modifiable cancer risk factors; importance of cancer screenings for men and women; graduation

The two facilitators are certified following a standardized training and certification protocol modeled after the protocol of the Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) trial (38). The facilitator training is conducted by two master trainers and the training activities are supervised by an RCT implementation and evaluation expert with extensive experience in training and certification of facilitators and testers.

2.2.3. Control condition: Health education group

To control for the effects of social contact and content, half of the participants are randomly assigned to a generic health education curriculum, known as the “10 Keys”™ to Healthy Aging program (72). This curriculum is similar to control condition programs used in other RCTs (27,76), and controls for educational content, group contact time, group size, attention, and facilitator contact time.

To provide the exact same amount of contact time and within-group clustering, the content of the full “10 Keys”™ program was adjusted for the purposes of this RCT. Specifically, we identified the most relevant health topics (i.e., cardiovascular disease, cancer, type 2 diabetes, clinical depression, etc.) and adopted the program’s content so that the selected topics can be delivered in 4 weekly 2-hour sessions in groups of 5–15 participants (see Table 2). NVOA, self-efficacy beliefs, and goal planning are not discussed in any way in this group. The specific health benefits of engaging in regular PA are also not discussed in the control group.

The two group facilitators for the control group are also certified following the same standardized training and certification protocol modeled after the protocol of the ACTIVE trial (38). Two master trainers manage the facilitator training, and an RCT implementation and evaluation expert with extensive experience in training and certification of facilitators and testers supervises the training activities.

2.2.4. Intervention and assessment fidelity

A major concern of any RCT is intervention and assessment fidelity. Intervention fidelity refers to the accuracy and consistency with which the intervention procedures are implemented (7,25) and is comparable to internal validity in experiments. To ensure strong intervention fidelity (7), we implement a rigorous fidelity monitoring protocol, involving the following measures: (1) Facilitators are trained for the intervention sessions using a structured training and certification protocol modeled after the ACTIVE trial (38). (2) Facilitators are also trained to use checklists to monitor critical steps in each intervention session. These checklists are reviewed by two master trainers and feedback is provided immediately (114). (3) The intervention sessions of each facilitator are video-taped, and the tapes are reviewed immediately by two independent master trainers. If the video tapes show any violations of the treatment protocol, then the master trainers provide immediate feedback and one of them visits the next intervention session as observer. Additional feedback is provided after this observation session to make sure that the intervention content is implemented according to the established protocol. To further promote intervention fidelity, facilitators complete refresher trainings in-between study waves and prior to a new wave that may follow a holiday break or short inactive period.

To assure assessment fidelity, we implemented several measures: (1) Data collectors are blind to group assignment and study hypotheses. (2) All data collectors are trained in the correct administration of the measures and interview procedures by a master trainer following a structured training protocol. The training protocol is modeled after the training protocol used in the ACTIVE trial (38) and includes (a) detailed instruction in the administration of each measure; (b) demonstrations of each measurement instrument; and (c) practice sessions with other data collectors. Data collectors will not be allowed to meet with study participants and collect data until they have been successfully certified for testing by the independent expert on assessment fidelity (38). The same training procedures are used to train the phone interviewers who make the phone calls to participants during the recruitment and monitoring periods. Like the data collectors, phone interviewers are blind to intervention group assignment and study hypotheses. An expert in RCT training and implementation supervises the fidelity trainings, attends the training sessions, and assists with the quality control measures.

2.2.5. Primary outcomes

2.2.5.1. Physical activity.

A strength of the Aging^PLUS program is that PA, which is the primary outcome, is assessed subjectively and objectively. For self-reported PA, participants are queried about their PA using the CHAMPS Physical Activity Questionnaire for Older Adults (100) and daily activity log. Objective data are provided by waist-mounted accelerometers. PA data are collected at Baseline, the Week 8 post-test, and the Month 6 follow-up (see Table 1).

The CHAMPS assesses participants’ weekly frequency and duration of various physical activities over the past 4 weeks (100). Caloric expenditure/week as well as frequency/week in all PA-related and moderate-intensity PA-related activities can be calculated. Caloric expenditure calculation is based on the metabolic equivalent (MET) values recommended by the Compendium of Physical Activity (1). The CHAMPS is particularly appropriate for this project given the age of the study participants and the expected increases in PA (100).

The daily activity log was developed and tested during the feasibility study (10). For each day in a designated week, participants record their minutes of mild, moderate, and/or vigorous PA, as well as the type of activity (i.e., walking, biking, etc.). The daily activity log is the primary self-report measure during week 8 when participants practice their self-chosen activity goal. Additionally, the daily activity log is used during the monitoring period. Total active minutes are calculated for week 8 averaged across the 7 days. A conversion into MET values is performed to derive a measure of caloric expenditure based on self-report.

Participants’ engagement in PA is objectively measured using the ActiGraph accelerometer (ActiGraphs, LLC; Pensacola, FL). Accelerometer data provide objective measures of the types (i.e., sitting vs. walking), as well as the intensity, duration, and frequency of activity throughout the day (14,101). We use the ActiGraph GT3X (ActiGraph, LLC; Pensacola, FL), which has been validated for quantifying PA when worn on the waist (102). On the first day of data collection, trained staff teach participants how to attach an upright ActiGraph accelerometer with an elastic, adjustable waist belt onto the right hip. The accelerometer is initialized to sample acceleration data at 30 hertz. During the 7-day Baseline data collection, participants are instructed to not remove the device, except at night while sleeping and during any water activity (i.e., bathing, showering, swimming), and to maintain typical activity patterns. They are also told that they will not be able to see the data recorded by the unit.

2.2.5.2. Views of aging (VOA).

Participants’ explicit views of aging, both positive and negative, are assessed at all occasions using three measures (see Table 1). Awareness of age-related change (AARC; ²³) is measured by a 10-item questionnaire with a 5-point scale (1 = Not at all; 5 = Very much) that assesses adults’ positive (gains) and negative (losses) self-perceptions of aging (i.e., “With my increasing age, I realize that I appreciate relationships and people much more”; “With my increasing age, I realize that my mental capacity is declining”; ³⁹). The reliability, validity, and factor structure of this measure have been established (11).

Age stereotypes are measured with a multidimensional questionnaire (43) assessing general VOA in 8 life domains: Physical and mental fitness, health and appearance; family and partnership; friends and acquaintances; personality and way of living; work and employment; religion and spirituality; money-related issues; and leisure activities and civic commitment. Each domain is assessed by 3–5 items, so domain-specific scores and a total score can be calculated. Participants’ expectations regarding aging (ERA) are assessed with a 12-item measure by Sarkisian et al. (86). Items are answered on a 4-point scale (1 = Definitely true; 4 = Definitely false) and measure the extent to which individuals expect to experience positive or negative age-related changes (i.e., “Having more aches and pains is an accepted part of aging”). Finally, the Essentialist Beliefs about Aging Scale (EBA-S; ¹⁰⁷) includes 4 items, which assess how malleable participants view the aging process (i.e., “To a large extent, a person’s age biologically determines his or her abilities”). Items on the EBA-S are answered on a 7-point scale (0 = Do not agree; 6 = Absolutely agree). The reliability and validity of the VOA measures have been established in prior research.

2.2.5.3. Self-efficacy beliefs.

Self-efficacy beliefs are assessed in terms of general and domain-specific self-efficacy beliefs. The General Self-Efficacy Scale (90) measures individuals’ dispositional sense of self-efficacy (8 items; i.e., “I can always manage to solve difficult problems if I try hard enough”), and the Self-Regulation Scale (21) measures adults’ ability to self-regulate their behavior when they encounter obstacles (10 items; i.e., “I can concentrate on one activity for a long time, if necessary”). Both scales use a 4-point response format (1 = Not at all true; 4 = Completely true).

Domain-specific self-efficacy is assessed in terms of exercise self-efficacy (65), motivational self-efficacy, and volitional self-efficacy (89). Exercise self-efficacy evaluates participants’ confidence in their physical capability to successfully complete 5-minute periods of walking (5–50 minutes) at a moderately fast pace. For each of the 10 items, participants indicate their confidence to execute the behavior on a 100-point percentage scale in 10-point increments, ranging from 0% (not at all confident) to 100% (highly confident). The measure of exercise self-efficacy is the average confidence rating across the total number of items (65,66).

Motivational self-efficacy assesses the perceived ability to initiate a PA program (i.e., “I am certain that I can be physically active on a regular basis, even if it is difficult”), whereas volitional self-efficacy assesses the perceived ability to adhere to a PA program (i.e., “I am capable of continuous physical exercise on a regular basis even if I need several tries until I am successful”; ⁸⁹). Each scale consists of 3 items rated on a 6-point scale (1 = Totally disagree; 6 = Totally agree; ⁹¹).

2.2.5.4. Goal planning.

Participants’ exercise-related goal planning is assessed in terms of behavioral intentions (7 items; i.e., “Going forward, I intend to be physically active at a moderate intensity for 30 minutes per day on 2–3 days per week during most weeks”), action and coping planning (7 items; i.e., “For the upcoming week, I have already planned which physical activity I will perform”), and action control (7 items; i.e., “During the past week, I consistently monitored when, where, and how long I was physically active”). A 6-point scale is used for the behavioral intentions (1 = Not at all true; 6 = Absolutely true), action and coping planning (1 = Totally disagree; 6 = Totally agree) and action control items (1 = Strongly disagree; 6 = Strongly agree). These scales have been extensively used in clinical and non-clinical samples and their psychometric properties are well established (80,89,96).

2.2.6. Secondary outcomes

2.2.6.1. Physical fitness assessment.

Objective health and physical functioning data are obtained from participants’ physical fitness tests, which are completed at Baseline, Week 8, and the Month 6 follow-up (see Table 1). Fitness is assessed to examine the extent to which participants’ change in PA results in a physiological training effect. Along with measuring blood pressure (BP), heart rate (HR), body mass index (BMI), and hip and waist circumference, the trained staff in the Human Performance and Clinical Research Lab (HPCRL) supervises participants in completing a handgrip strength test and the Short Physical Performance Battery (SPPB; ³²). Under the supervision of the same trained staff, participants also complete a submaximal cardiorespiratory fitness test, using a cycle ergometer and following an age-validated protocol. This test is used to estimate VO₂max capacity. The test was chosen because the PA eligibility criterion requires adults to be mostly inactive and, thus, the submaximal cardiorespiratory fitness test was considered most appropriate and safe. After taking participants’ BP and HR, they begin the cycle ergometer test, which lasts about 13–15 minutes and includes 3 stages: warm-up, guided exercise, and cool down. While revolutions per minute remain consistent throughout the test, resistance is changed by the trained staff across the 3 stages based on participants’ rate of perceived exertion, or how intensely their body is currently working. BP and HR are monitored throughout the stationary bike test, as well as after the completion of the test. Estimates of VO₂max capacity are calculated using sex-specific formulae that account for a person’s age, HR, and power output (9,73).

2.2.6.2. Implicit views of aging.

Participants’ implicit views of aging are assessed using a lexical decision-making task (LDMT; ¹⁸) and an age-specific brief implicit attitudes test (BIAT; ^31,37,98) at Baseline, Week 8, and the Month 6 follow-up (see Table 1). The LDMT is administered on a computer screen and reaction times are measured in response to words presented in a specific order. That is, the target categories “young” vs. “old” are followed by either real words (i.e., “wise”) or non-words (i.e., “flurb”). The real words are either positive or negative and either characteristic of younger or older adults (i.e., 4 groups of words). Participants press a designated key on the keyboard to indicate if a word is a real word or a non-word. The reasoning behind this test is that participants focus their attention on making correct word-related judgments rather than making potentially value-laden age-related evaluations (30,37). The developer of the lexical decision task serves as consultant to the project to ensure the correct administration of the test.

The BIAT is also computer-administered and requires participants to categorize stimuli shown on the computer screen. The stimuli represent one of four categories: young faces, old faces, good words, and bad words. Participants see two of the four categories at the top of the screen (i.e., “YOUNG or GOOD”, “OLD or GOOD”) and quickly decide by pressing one of two designated keys if the face (i.e., young face, old face) or word (i.e., “joy,” “horrible”) appearing in the center of the screen matches the categories at the top of the screen. It is crucial that participants complete the test as quickly as possible to assess the true magnitude of the relation between stimuli (i.e., older adults and good words). In theory, participants will more quickly categorize stimuli that are presented they typically think match (i.e., older OR younger adults AND good).

2.2.7. Other variables

2.2.7.1. Personality.

Because participants’ enduring personality traits may influence how they respond to the intervention, personality is assessed with the 60-item NEO Five-Factor Inventory (NEO-FFI; ^20,67). The NEO-FFI questionnaire is widely used to study adult personality and assesses the Big Five master traits of Extraversion, Neuroticism, Conscientiousness, Agreeableness, and Openness to Experience. The NEO-FFI utilizes a 5-point scale (1 = Strongly disagree; 5 = Strongly agree).

2.2.7.2. Health status and functional health.

Participants’ health status is assessed with an established medical diagnoses checklist and functional health with the SF-36 Health Survey (Version 2.0; ¹⁰⁶). The medical diagnoses checklist queries participants on a variety of health and lifestyle questions, such as questions related to menopause for women, diagnoses of chronic health conditions (i.e., cardiovascular disease, cancer, Type 2 diabetes, etc.), frequency of doctor visits per year, and amount of weekly alcohol consumption. The SF-36 provides an 8-scale profile of a person’s functional health and well-being as well as summary scores for physical and mental health. The SF-36 is one of the most widely used instruments for assessing functional health in survey research with community-residing adults (38,44,71).

2.2.7.3. Participant expectations.

Participants’ perceptions of education group effectiveness and group experience are assessed with 7 questions and 1 open-format item. For perceptions of education group effectiveness, participants are queried about their expectations for the group’s potential (i.e., “Education groups like the one I will participate in have the potential to change how I view my own aging.”). Perceptions of education group experience items assess participants’ motivation to be part of the group (i.e., “In my education group, I am motivated to learn more about healthy and active aging.”). All items are rated on a 7-point scale (1 = Very strongly disagree; 7 = Very strongly agree). Lastly, the open-format item gives participants the opportunity to share what they hope to learn or gain from the program.

2.3. Safety monitoring

In addition to the Aging^PLUS program being in full compliance with the requirements of the Institutional Review Board, multiple measures have been adopted to ensure participant health and well-being throughout the study. First, as part of the enrollment procedures, the research team requests approval from the prospective participant’s PCP, asking for confirmation that the interested individual is healthy enough to engage in PA. If a prospective participant does not have a PCP, or the PCP does not respond to the team’s request, then participants have the option to sign a waiver of liability. In cases where the PCP states that it is not safe for an individual to start physical activity, the person will not be enrolled in the study.

Second, during physical fitness assessments, trained professionals monitor all activities. BP is routinely measured during these assessments and if a systolic resting BP reading is at or above 150 and/or a diastolic BP reading is at or above 90, the test is not started, and the values are immediately reported to the on-staff physician and the project coordinator. If the prospective participant is agreeable, he or she will be rescheduled at a later point in time for a second trial and after a consultation with his or her PCP. If the second trial again raises concerns due to high BP readings, then the participant is excused from the study and is strongly encouraged to visit his or her PCP to get proper hypertension treatment. The participant is also informed that he or she may be eligible for future waves of the program if proper control of the hypertension has been achieved. Lastly, the trained professionals overseeing the physical function assessments, the on-staff physician, the PI, and the research team are regularly in communication regarding the health and safety of participants, and adverse events reports are filed with IRB and the data safety monitoring officer whenever necessary.

2.4. Sample size justification

The sample size for this RCT was determined by considering issues of study design, statistical power, and participant recruitment. We estimated the required sample size to detect a medium effect size of f₂ = .15 for mixed linear models/multilevel models (63,108) with two groups at Level 2 (i.e., treatment conditions) and four times of measurement at Level 1 (Baseline, Week 4, Week 8, Month 6 follow-up) with a power of .80 at a significance level of α = .05. This calculation showed that a total sample size of 240 participants would be required (i.e., 120 adults per group). Accounting for a 20% drop-out rate over the course of the study resulted in 150 participants per group and a total sample size of 300 adults. We applied a higher dropout rate in our calculations compared to the drop-out rate in the feasibility study (i.e., 13.1%; ¹⁰) because this RCT is more demanding and takes longer to complete, making attrition more likely. This sample size also has a power of .95 to detect large effect sizes if they are obtained. Moreover, this sample size provides sufficient power for the planned multilevel analyses. The power analysis also takes into account that in RCTs with multiple times of assessment, the outcome variables for individuals in the same group are correlated.

2.5. Statistical analysis

2.5.1. Aim 1. To conduct a RCT examining the efficacy of the Aging^PLUS program, a 4-week intervention that targets motivational mechanisms to promote adults’ engagement in PA.

Using the Proc Mixed procedure in the SAS® 9.4 statistical package (SAS Institute Inc., Cary, NC), multilevel models as a form of mixed linear models (61,108) and intent-to-treat (ITT) analyses (12,110) will be performed. In terms of the specific analyses for Aim 1, we will examine the change in the dependent variables across the occasions of measurement (i.e., occasion) at Level 1 and the effect of treatment condition (i.e., treatment) at Level 2. Analyses will be performed individually for each measure of NVOA, self-efficacy beliefs, goal planning, and PA to examine if the intervention effects are consistent across measures. Although we expect consistency across these individual measures, even if the data reveal sporadic inconsistencies, this information will be valuable with regard to the key mechanisms and the selection of the most appropriate measures in future work. The cross-level Treatment × Occasion interaction will be significant (p < .05), if in comparison to control group participants, Aging^PLUS participants: (1) show significant improvement in NVOA, self-efficacy beliefs, and goal planning at Week 4 and Week 8 in comparison to Baseline, (2) have more positive implicit attitudes about aging at Week 8, and (3) show a significantly higher level of PA (via daily activity log and accelerometer) at Week 8 and the Month 6 follow-up. After these primary hypotheses have been tested, further questions can be addressed to examine the effects of potential moderators of the treatment effects, such as the effects of age, sex, level of education, or starting level.

ITT analyses will be performed, implementing the approach proposed by White et al. (110). First, every effort will be made to follow up with all randomized participants, even if they withdrew from the assigned treatment. To the extent possible, reasons for withdrawal will be documented and reported. Next, we will perform a set of main analyses with all available data. By using SAS Proc Mixed (or alternatively Mplus; ⁷⁰), these analyses will include cases with partial data. Third, a detailed analysis of the nature of the missing data will be performed. This step is crucial to understand the extent and pattern (i.e., occurrence and order) of the missing data, as well as the nature/hierarchy of the missingness (i.e., missing completely at random vs. missing at random vs. missing not at random; ⁸⁸). The gained knowledge will determine the statistical options (which are also software-dependent) for the best estimation of missing data. Because recent discussions of the ITT method have concluded that the last observation carried forward (LOCF) approach is outdated, we will examine whether missing data imputation (i.e., multiple imputation) provides a sensitive solution to handling the missing data beyond the flexibility provided by our multilevel models, which include cases with partial data. Lastly, we will perform sensitivity analyses to compare findings across analytic models and gain an understanding of the effect of departures from the assumptions made in the main analyses. In the interest of reproducibility and transparency, we will provide detailed descriptions of these analyses in all publications resulting from this project (68). An experienced intervention methodologist/biostatistician will supervise these analyses.

2.5.2. Aim 2. To test a conceptual model of the mechanisms underlying the intervention effects.

Mplus (Version 8; ⁷⁰) will be used to test a multiple mediator model (34), examining whether improvements in NVOA, self-efficacy beliefs, and goal planning measured immediately after the intervention (i.e., week 4) independently and jointly mediate the association between the treatment and change in PA at week 8. This temporal ordering of the mediators is required to establish support for the posited causal pathways (83). This model will be tested in a two-stage approach. In Stage 1, individual models will be tested for the measures of the three key mechanisms using the adjusted change score method recommended by Roth and MacKinnon (29,83). That is, pre-test-post-test change scores of all mediators and the outcome variables will be adjusted for Baseline values (not shown in Figure 2).

Figure 2.

Multiple Mediator Model

Using the same method, in Stage 2, a comprehensive model with all significant mediators identified in Stage 1 will be tested to examine unique vs. shared mediation and to calculate the total mediated effect across all significant mediators. An alternative to this two-stage approach is the testing of a latent mediator model. The latent mediator approach requires that the individual measures for each intervention mechanism are moderately correlated. A bootstrapping approach of drawing 5000 random samples from the data will be used to evaluate if the total indirect effects are significantly different from zero, and 95% confidence intervals will be estimated.

Mplus (Verson 8; ⁷⁰) will be used to test the direct (c’) and indirect (a1×b1, a2×b2, a3×b3) pathways of this mediation model. To make decisions on full, partial, or no mediation, the percentage of the total effect that is mediated or unmediated will be calculated. Although significance of the indirect effect will also be assessed, significance is a function of effect size and sample size, which is why it is important to consider alternative methods of testing for mediation.

We hypothesize that change in NVOA, self-efficacy beliefs, and goal planning will be significant mediators (p < .05) of the effect of the Aging^PLUS program on participants’ PA at week 8. Our review of the goal planning literature (93) also suggests that measures of NVOA and self-efficacy beliefs will be significantly stronger mediators than measures of goal planning, and we will test this assumption explicitly. Results from testing this model will elucidate the differential role NVOA, self-efficacy beliefs, and goal planning play as mechanisms of action to increase adults’ engagement in PA (81). Of note, this model can also be extended to a three-wave model (83), in which the effects of the mediators on participants’ PA at the Month 6 follow-up can be examined.

2.5.3. Aim 3. To conduct a 6-month follow-up to examine the long-term effects of the intervention.

The purpose of this aim is twofold: (1) To examine the time course of the intervention effects across the 4 times of measurement (i.e., Baseline, Week 4 post-test, Week 8 post-test, and Month 6 follow-up) and (2) because half of the treatment group is randomly assigned to a monitoring condition, to examine the effect of monitoring on the maintenance of the treatment effects. We will use longitudinal multilevel modeling (MLM; ^95,97,108) to model the time course of the intervention effects. MLM is the method of choice to analyze such data for the reasons already described in the first sub-section under Statistical Analysis. The dependent variables (DVs) to be modeled are the measures of NVOA, self-efficacy beliefs, goal planning, as well as measures of PA. Separate models will be performed for each DV, and composite scores will be used whenever appropriate for parsimony reasons.

At Level 1, we will examine the simple effect of time (i.e., within-person change) by using weeks in study to clock time. Both linear and curvilinear effects of time will be examined, as some decay of the intervention effects can be expected. At Level 2, we will incorporate two between-person predictors of change. First, for one set of analyses, treatment condition will be included as a dummy-coded variable, using the control group as the reference group. Second, for another set of analyses, condition of monitoring will be included as a dummy variable with the non-monitored group serving as the reference group. Findings from these analyses will reveal whether time-related change on the DVs varied between treatment group and monitoring condition.

Although the overall purpose of these analyses is descriptive in nature, we expect that long-term maintenance of the intervention effects will be significantly greater in the Aging^PLUS group compared to the social contact control group. Similarly, we expect that long-term maintenance of the intervention effects will be significantly greater in monitored individuals compared to non-monitored individuals. Findings from these analyses will be crucial in informing the next phase of our research program, which will focus on optimizing the intervention with personalized monitoring or targeted feedback.

3. Discussion

Being physically active on a regular basis has been described as “one of the most promising non-pharmacological, noninvasive, and cost-effective, methods of health-promotion” (50). For this reason, engagement in PA has been identified as one of the most promising approaches to promote healthy and successful aging (50). The many health benefits associated with regular PA extend to a variety of health conditions (24) and to psychological health and cognitive functioning (24,78). Considering the many advantages of PA engagement, intervention researchers should focus on how to motivate adults to engage in PA to promote healthy aging (17,60,104).

Despite widespread awareness of the benefits associated with PA, the percentage of adults meeting the official PA recommendations is quite low (16) and engagement in PA tends to decline with age (33,40). At the same time, previous research indicates that NVOA, low personal self-efficacy beliefs, and deficient goal planning are deleterious to individuals’ health and are possibly among the root causes for many adults’ sedentariness (42,74). For example, negative attitudes and self-perceptions of aging have been shown to be linked with poorer cognitive, physical, and psychological health in experimental and quasi-experimental studies (35,52,54,55). Similarly, having low self-efficacy beliefs is associated with a low likelihood of engaging in regular PA, eating a healthy diet, or maintaining better health overall (26,49,51). Finally, individuals with deficient goal planning skills tend to engage in fewer health-promoting behaviors and often give up quickly when they encounter obstacles in executing and maintaining a change in health behavior (74,84,89). To date, only four studies have targeted NVOA in combination with other social-cognitive factors with the intent to increase older adults’ engagement in PA. To address the major public health concerns related to adults’ sedentariness, additional programs need to be developed with a focus on these motivational factors.

The Aging^PLUS program was designed to fill this gap in the research literature by targeting NVOA, low self-efficacy beliefs, and deficient goal planning to increase PA in middle-aged and older adults. The Aging^PLUS RCT is significant and innovative for multiple reasons: (1) The theory-driven delineation of intervention targets; (2) the implementation of the experimental medicine approach; (3) the focus on a cluster of risk factors; (4) the focus on implicit aging attitudes; and (5) the assessment of short- and long-term effects. In the following paragraphs, we will elaborate on each of these points individually.

First, a theoretical model was proposed to examine if and to what extent improvements in NVOA, self-efficacy beliefs, and goal planning produce the observed changes in PA. This is innovative because most intervention studies do not explicitly test theoretical models to confirm or disconfirm the role of the putative mechanisms underlying observed behavior change (74,75,109). In addition, testing such a model helps to elucidate the putative effects of multiple intervention pathways and their differential impact on the outcome behavior. Thus, this approach recognizes that focusing on a single pathway may be too narrow and may miss critical elements of the target behavior under investigation (6,75,79,89). Furthermore, such an approach allows for future fine-tuning of the program with a rigorous focus on the active explanatory mechanisms of change.

Second, using the experimental medicine approach to behavior change (62,81) and adopting the NIH Stage Model of Behavioral Intervention Development (75), the significance and innovation of this RCT is in its focus on the targeted modification of motivational mechanisms that are well-documented barriers to healthy aging. Due to the specific focus on social-cognitive beliefs and motivations, the Aging^PLUS program goes considerably beyond the simple advice and encouragement model often used in public health messages (103) and contributes to a precision approach in prevention science (81).

Third, this RCT is the first to target NVOA, low self-efficacy beliefs, and deficient goal planning as a cluster of psychological mechanisms and risk factors for improving PA in middle-aged and older adults. This work translates basic social-psychological research on NVOA, self-efficacy beliefs, and goal planning into a structured intervention program and examines the purported mechanisms in a rigorous way. Additionally, by addressing and counteracting the negative effects of these three psychological mechanisms and risk factors together, there is a greater likelihood of generating sustained behavior change, permitting the determination to what extent each mechanism contributes to the observed change in PA.

Fourth, because NVOA operate mostly outside of a person’s conscious awareness (37,52), it is also important to show whether and to what extent the intervention changes not only participants’ explicit attitudes about aging, but also their implicit attitudes. That is, Aging^PLUS also addresses the basic question whether it is necessary to change individuals’ implicit aging attitudes in order to promote long-term health behavior change, or whether it is sufficient to change individuals’ explicit attitudes (56). The reasoning underlying this objective is that implicit attitudes tests are free of self-presentational influences (30) and are, therefore, more basic and valid assessments of individuals’ views on topics that may be subject to stereotyping and prejudice, including NVOA and age stereotypes.

Fifth, findings from this study will provide insights into the durability and time course of the achieved intervention effects, and their association with PA and physical health (i.e., cardiorespiratory fitness). Evidence about the durability and time course of intervention effects to increase engagement in PA is currently limited (10,56) and very little is known about the long-term effects of motivational interventions. Following participants over a 6-month period represents a start to address the long-term maintenance of the observed behavior change.

Along with these innovations, the Aging^PLUS program has additional advantages. The program is overall relatively inexpensive and time-efficient in its delivery. Although 8 hours may seem like a large time commitment for participants, this amount of time seems rather modest in light of the fact that individuals develop their NVOA over an entire lifetime (53,92). In terms of venues of delivery, we have worked so far with organizations serving adults in the community, but we have also taken steps to deliver the program as part of workplace-based wellness programs. The advantages of this context are that we are reaching middle-aged adults and individuals who prepare for retirement (22), and that the costs of the program are covered by the employer. In addition, we are exploring the possible dissemination of the program via internet and audiovisual means (i.e., video discs) for in-home use (64,105).

Despite the strengths and advantages of the Aging^PLUS trial, there are also several limitations that need to be noted. Although we limit how often participants must attend in-person sessions throughout the study, the duration of the study from beginning to end spans about eight months. As such, the study requires a strong commitment from participants. If participants are not committed to the 8-month study and drop out, the final sample may be selective, preventing us from generalizing the study findings to less selective populations. Additionally, this trial only included adults aged 45–75 years, but adults younger than 45 or older than 75 could also benefit from the Aging^PLUS program. Lastly, it is important to acknowledge that there are other approaches to increasing adults’ engagement in PA. For example, some programs teach adults directly how to engage in physical exercise or implement an approach that strengthens participants’ social support network as a critical component of behavior change. Although all of these approaches are important to explore and to compare, it is noteworthy that the Aging^PLUS trial focuses on a cluster of risk factors that have not been examined before.

In summary, the Aging^PLUS program has the potential to make a valuable contribution to our understanding of the role of motivational factors in promoting engagement in PA in middle-aged and older adults (74). Understanding the role and function that NVOA, self-efficacy beliefs, and goal planning skills play individually and conjointly may represent a significant step forward in the design of tailored interventions to increase adults’ engagement in PA (50). Finally, we also believe that Aging^PLUS is a cost-effective program that has high potential for dissemination, becoming possibly an important instrument in the public health toolkit.