Combined intervention approaches for initiating and maintaining physical activity in depressed individuals: design and rationale of the Project MOVE randomized clinical trial

Abstract

Introduction:

Regular engagement in physical activity decreases risks for many chronic conditions, and may also improve depression symptoms. However, rates of physical activity and adherence to exercise interventions remain low among depressed individuals relative to non-depressed individuals.

Methods:

This is a study protocol for Project MOVE. This study is a theoretically-driven, 3-arm randomized controlled trial for increasing physical activity with depressed adults. Each successive arm includes an added component that may serve to increase and maintain physical activity. The arms are: 1) Brief advice (BA) to exercise alone (minimal treatment control condition); 2) BA + supervised and home-based exercise (SHE) + health education (HE; serves as contact control for CBEX); and 3) BA + SHE +cognitive-behavioral sessions focused on increasing and maintaining exercise (CBEX). The target sample size is 240. Assessments are conducted at baseline, Month 1.5, end of intervention (month 3), and at 6 and 9 months. The primary outcome is minutes of moderate-to-vigorous physical activity, assessed via an accelerometer. Secondary outcomes include cardiorespiratory fitness, body composition, and depression, and maintenance of moderate-vigorous physical activity through 6 and 9 month follow-ups. Mediators and moderators derived from behavior change theories, including the Health Behavior Model, Self-Determination Theory, and Social Ecological Theory, will be examined.

Conclusion:

Project MOVE is designed to test primarily whether both a structured exercise program (SHE) and a cognitive-behavioral group (CBEX) increase physical activity in depressed adults during both a 3-month intervention period, and during the 6-months that follow.

Introduction

Regular engagement in physical activity (PA) protects against numerous health conditions [1, 2]. US Department of Health and Human Services guidelines recommend that adults engage in at least 150 minutes per week of moderate-vigorous PA (MVPA) [2]. Although depressed individuals are at increased risk for numerous physical health problems [1–7], converging evidence suggests that many do not exercise regularly [4, 5]. For example, Behavior Risk Factor Surveillance System data suggest that only 45% of people with “serious psychological distress,” compared to 67% of those without distress, reported at least 150 mins per week of MVPA [6]. Epidemiologic and intervention studies show that baseline depression consistently predicts poorer adherence to recommended PA levels [7–13], even after controlling for variables such as cigarette smoking, percent body fat [13], and health conditions [10]. Core features of depression such as anhedonia (lack of enjoyment) and fatigue, as well as poor problem-solving abilities, may make it especially difficult to develop motivation and cope with cognitive/affective and environmental barriers to PA [8, 10].

Previous interventions with depressed individuals have focused almost exclusively on whether PA promotion programs can ameliorate depression symptoms [e.g. 14-24], with evidence suggesting that increased PA leads to decreased symptoms [20, 21]. However, there is little existing research on HOW best to increase PA among depressed populations. Only a few studies with depressed individuals provide data comparing the impact of different types of exercise interventions on the overall amount of PA [25–30]. For example, one study examined supervised group vs. home-based exercise [25], finding that frequency of exercise was greater with home-based exercise, but physical fitness improved more with supervised exercise. Another study examined different “doses” of prescribed exercise [26], finding that when a higher dose of exercise was prescribed, participants engaged in more frequent/intense exercise.

In other populations, cognitive-behavioral interventions have been successful in increasing PA [25]. Such an intervention could target both depression-specific and general motivators and barriers to PA by increasing salience of exercise benefits, using problem-solving skills to manage barriers, learning and practicing goal setting, developing skills for mobilizing social support, and choosing activities that meet one’s own psychological needs. The combination of a structured exercise program and a cognitive-behavioral group might yield the best outcomes in terms of amount of MVPA per week. No study to date has addressed whether an adjunctive cognitive-behavioral intervention increases MVPA in depressed individuals.

A second question is whether a structured exercise program by itself, or a structured exercise program with a cognitive-behavioral component, can increase MVPA in the time period following the acute intervention. Since the benefits of exercise may not persist once exercise is discontinued [32], understanding whether exercise interventions lead to continued longer-term maintenance of PA in depressed individuals is critical. The little available previous research in depressed individuals is inconclusive [33, 34].

The current study is designed to address these questions to understand key components of an effective exercise intervention for depressed individuals. This paper describes the design and rationale for a randomized controlled trial called Project MOVE.

Methods

Detailed aims

In Project MOVE, we will test whether a cognitive-behavioral intervention (CBEX) that targets depression-related exercise barriers and teaches problem-solving skills, combined with a supervised and home-based aerobic exercise program (SHE), is superior to SHE + a health education (HE) contact control in both increasing and maintaining MVPA in depressed individuals. Moreover, this study will assess whether both programs are superior to a minimal, brief advice for exercise control arm (BA). We will assess physical activity both during a 3-month intervention period, and in the 6 months that follow.

Study design

This study is a 3-arm randomized controlled trial with adults with depression. Each successive arm includes an added component that may serve to increase and maintain physical activity. Arms are: 1) Brief advice (BA) to exercise alone (minimal treatment control condition); 2) BA + supervised and home-based exercise (SHE) + health education (HE; serves as contact control for CBEX); and 3) BA + SHE +3) cognitive-behavioral sessions focused on increasing and maintaining exercise (CBEX). The intervention occurs over a 3-month period. Assessments are conducted at baseline, midpoint (Month 1.5), end of intervention (3 months), and 6 and 9 months follow-up.

The focus of this study is to help participants initiate and maintain a moderate-to-vigorous level of leisure-time physical activity. This may include brisk walking, aerobic exercise with exercise machines, or other types of exercise. We focused on leisure-time physical activity as it is beneficial to both psychological and physical health (whereas occupational physical activity may present risks to physical health) [35] and it is more easily modifiable than occupational physical activity.

The acute intervention period is 3 months. We chose 3 months because there is evidence that a 3-month intervention timeline can adequately produce measurable changes in physical activity, including in depressed individuals [36, 37]. The goal of this study is to assess both acquisition and maintenance of regular physical activity. It is possible that a 3-month intervention is not sufficient for maintenance of physical activity; that is one of the questions we will investigate. We plan to follow participants for 6 months after the acute intervention period to test whether the interventions have a sustained impact on MVPA.

The Institutional Review Board (IRB) of Butler Hospital where this study is conducted has approved all study procedures (IRB ID #793390–53). The research team provides ongoing monitoring of the trial, with an external Safety Monitoring Committee reviewing recruitment, retention, and adverse events twice per year. This trial is registered at ClinicalTrials.gov (NCT02691845).

Theoretical framework for Project MOVE

The interventions tested in Project MOVE (i.e., BA, CBEX, SHE) are based on existing theories about which factors are likely to influence adherence to a PA program and acquisition and maintenance of MVPA among depressed individuals. The Social Ecological Theory of PA suggests that individual-level psychological constructs (e.g., self-efficacy), social environment factors (e.g., social support for PA), and physical environment characteristics (e.g., weather, access to parks, safe neighborhoods, etc.) predict both initiation and maintenance of PA [38]. Studies assessing Social Ecological Theory constructs in PA behaviors have been conducted primarily with non-depressed populations. However, the characteristics of the social and physical environment may have a particularly strong impact on people with depression. Depression can be characterized by poor social skills [38] and low levels of social support [39]. To the extent that depression is correlated with socioeconomic status [39], depressed people may live in physical environments that are less resource-rich. Further complicating this, depression is associated with poor problem-solving skills [40] and low motivation. When the environment cannot be directly manipulated, good problem-solving skills, and motivation to use those skills may enable the individual to manage environmental barriers. Thus, interventions that teach problem-solving skills, coach people to engage social supports, or provide direct support for finding safe environments in which to exercise may be important for people with depression.

The Health Belief Model and Self-Determination Theory also inform the current trial. Both have been used extensively as the foundation for theory-based interventions to promote PA [39–47]. The Health Belief Model suggests that perceived susceptibility to a health problem, perceived severity of a health problem, perceived benefits of a behavior change, perceived barriers to action, cues to action, and self-efficacy predict health behavior change [44–47]. Several of these barriers or facilitators may be influenced by depression. For example, relative to non-depressed individuals, depressed individuals may perceive fewer benefits of engaging in specific health behaviors such as PA (e.g., [45]). Similarly, depressed individuals may have difficulty believing that PA will help them to feel better physically or mentally. Even if they do experience improved mood with acute bouts of exercise, they may be less likely to recall this experience later. Indeed, previous research suggests that individuals with depression may not use their reinforcement history to modulate current behavior [46]. Even if a choice was previously rewarded, they are less likely, relative to non-depressed people, to make that choice again in the absence of an immediate reward. Therefore, depressed individuals may need assistance making benefits of exercise salient and may require increased environmental reminders and cues not only to action, but also as to why exercise is personally important to them.

Moreover, according to the Health Belief Model, depressed individuals may perceive more barriers to action due to low energy, feelings of being overwhelmed, low self-worth (including poorer body image), low social support, unsafe environments, or desire to withdraw from interactions with others (see [43, 46] for empirical support). When depressed individuals have difficulty engaging in exercise, they may be prone to self-blame and self-criticism [43]; indeed, fear of failure may be a barrier to initiating an exercise program. Finally, self-efficacy for physical activity is associated with increased exercise adherence (e.g., [45, 46]), and this may be particularly impaired with depression (see [47] for empirical support).

Self-Determination Theory further elaborates on which perceived benefits may serve as motivators to behavior change [48, 49]. Self-Determination Theory suggests that health behavior change is influenced by intrinsic motivation based on one’s values (i.e., exercising to promote good health) and extrinsic motivation based external controls (i.e., exercising to achieve a certain body weight). Previous research has consistently shown that intrinsic motivation is associated with a greater likelihood of maintaining increased PA over time [see 50]. Among individuals with depression, for whom motivation can be low, the experience of intrinsic motivation to exercise may be challenging to attain. Research has shown that anhedonia, a core feature of depression, predicts less physical activity, and that this association is mediated by less enjoyment of physical activity [51–54]. Therefore, it may be important for individuals with depression to have assistance in building intrinsic motivation for exercise.

Project MOVE Interventions

We next describe each of the study interventions with an emphasis on their relation to theories described above. All interventions except Brief Advice are provided in a group format. This is similar to interventions tested in previous work, and is less costly than providing interventions in an individual format. Further, participants may benefit from and enjoy the sense of community that comes from being in a group.

Brief Advice (BA).

All participants receive BA [55] prior to randomization. For this intervention, participants meet with a trained research staff member once for 45 minutes and discuss the following: public health guidelines for physical activity (i.e., gradual increase in activity up to at least 150 minutes of MVPA per week), physical and mental health benefits of PA, components of typical workout (including warm-up, aerobic zone, cool-down, and knowing when to stop), and how to determine exercise intensity using rate of perceived exertion [RPE, 56] by estimating heart rates with radial and carotid pulse. Participants are given lists of free/lost cost resources to help them select activities that can be done inside (e.g., aerobics DVDs), outside (e.g., bike paths), regardless of the weather (e.g., mall walking), and are easily integrated into daily life (e.g., walking dog). Moreover, all participants receive a booklet that covers basics of depression treatment options (i.e., medication, psychotherapy), and stresses the importance of PA.

Lastly, as part of BA, participants are guided to develop an initial plan for PA that identifies: (a) a list of physical activities (e.g., group exercise at a fitness center, brisk walking or jogging) that they plan to engage in during the 3-month intervention, (b) important barriers that may prevent engagement in these activities (e.g., time management, mood, lack of social support, access to walkable environments), and (c) strategies to overcome some barriers. For example, the research staff and participants may discuss the lack of neighborhood walkability and generate ideas about alternative places for exercise (e.g., nearby malls and walking hours, nearby bike paths and walking trails). Thus, BA provides some basic information about exercise, as well as information and problem-solving that may help to address social and environmental barriers to exercise.

Supervised and home-based exercise (SHE).

In considering the optimum type of structured exercise program, we note that most PA interventions with depressed patients have involved multiple weekly sessions of supervised exercise [see 57 ]. Although these approaches have shown some efficacy in decreasing depression symptoms, attending multiple sessions each week may not be feasible for many patients. Home-based or partially home-based PA interventions have shown promise in depressed populations [26, 27]. Given core depression symptoms of amotivation, anhedonia, and fatigue, and poor problem-solving, a combination of supervised and home-based exercise (similar to Trivedi [26] ) may be the best balance of external encouragement and monitoring vs. convenience for depressed patients.

Participants assigned to SHE [57] attend aerobic exercise sessions supervised by an exercise physiologist once a week at the Butler Hospital Fitness Facility. The exercise session lasts ~1 hour and includes: review of self-monitoring logs, warm-up (5–10 mins), aerobic exercise (beginning at 20 minutes per session with weekly gradual increases based on heart rate monitoring by the exercise physiologist, with the target being the moderate-intensity range of 64–76% of age-predicted maximal heart rate [58] for each participant), cool-down (5–10 minutes), stretching (5 minutes), and exercise plans for the next week (5 minutes). Participants have the option of choosing from several types of exercise equipment, including treadmills, elliptical machines, and stationary bicycles.

At each session, participants and the exercise physiologist set exercise goals to engage in MVPA 2 to 4 additional times a week in the context of their own environment (e.g., at home or through community resources). On a week-to-week basis, the exercise physiologist and participant set a short-term goal that gradually increases MVPA based on amount completed the previous week. The exercise physiologist discusses with the participant type, frequency, and duration options for achieving weekly exercise goals. Participants are encouraged to select physical activities that are feasible and attractive for them. They are given the option of exercising in multiple short bouts (e.g., 10 minutes) or in longer bouts (e.g., 30 minutes). As part of the intervention, participants are asked to self-monitor their exercise by filling out a weekly exercise log with the exercise activities they engaged in during the week, the duration of each activity, and their self-reported rate of perceived exertion for each activity.

After the 3-month intervention, participants receive calls from the exercise physiologist on a monthly basis for the duration of the study (i.e., months 4–9). During these calls, the exercise physiologist asks about current activity level, helps the participant to set an appropriate exercise goal for the next few weeks based on how much they exercised in the past month, problem-solves with the participant vis-à-vis any barriers to exercise, and encourages the participant to continue to keep an exercise log.

Consistent with the Health Belief Model [49,52], SHE aims to increase perceived self-efficacy for PA by teaching participants how to set goals for PA and monitor their daily and total weekly minutes of MVPA. SHE also provides them with social support, i.e., a group-based exercise environment wherein other people with depression are engaging in PA. Corresponding to the SDT, as participants engage in the supervised exercise sessions, and experience a positive effect on their mood state, their perceived benefits of PA in relation to mental health or depression management will likely become more salient. This would likely increase intrinsic motivation for PA, which is essential for long-term maintenance. Moreover, returning to the Health Belief Model, the monthly phone calls from the exercise physiologists after the 3-month intervention provides participants with continued support to help them to overcome ongoing or emerging barriers during the follow-up periods.

CBEX.

Participants in the CBEX [55] condition participate in a 30-minute group counseling session led by a research therapist focused on teaching cognitive-behavioral skills to help with increasing motivation and adherence to exercise. This session occurs weekly throughout the 3-month intervention period, either immediately before or after their supervised exercise session. CBEX directly targets: 1) anhedonia, amotivation, increased fatigue, and other individual-level barriers associated with major depression; 2) individual-level determinants of behavior change described by the Health Belief Model [43] and Self-Determination Theory [52]; and 3) social environment and physical environment factors associated with exercise [38, 39] (through social skills coaching and problem-solving training) [48, 49]. Each session begins with a review of personal barriers and facilitators to exercise, and the therapist assists the patient with using skills learned in CBEX to manage barriers. Sessions emphasize making personal choices consistent with goals and values, and participants have menus of options whenever possible. Participants receive a workbook to use for this intervention. The session topics are described in Table 1.

Table 1.

CBEX intervention description.

Session	Content	Targets
** All sessions	Discussion of participant-specific facilitators and barriers to exercise in previous week; application of skills (as they are learned) to manage barriers and increase facilitators
1. Benefits of exercise	Introduction of potential benefits; exploration of benefits most important to the patient; ways to increase salience of benefits to the patient.	PB, IdM
2. Barriers to exercise	Identification of individual, social, and environmental barriers to exercise; introduction to cognitive and behavioral strategies for managing barriers	B
3. Problem- solving	Using problem-solving to address individual, social, environmental barriers to exercise	B
4. Goal setting	Goal setting as a skill that is acquired through practice; specific steps toward developing manageable and realistic exercise goals	EnSM, SE
5. Getting social support	Identification of supportive people who will help the participant overcome exercise barriers and encourage/ reward exercise	PB, B
6. Getting motivated, staying motivated	How to select activities that meet individual psychological needs, i.e., are enjoyable, increase a sense of accomplishment, or allow socializing	PB, InM
7. Exercise & mental health	Benefits of exercise for depression and anxiety	PB, IdM
8. Time management	Key skills for time management such as prioritizing and planning activities with a special emphasis on decreasing barriers to exercise.	B
9. Getting motivated, staying motivated, part 2	Focus on reasons for exercise; self-statements that will remind one of reasons to exercise; visualizing success	PB, IdM
10. Getting back on track	How to identify situations considered high-risk for deviating from an exercise program; developing coping strategies for these situations.	B, SE
11. Physical activity and health	Review of strategies for engaging in physical activity when one has chronic pain or other chronic conditions	PB, B, IdM

Note. PB = perceived benefits; B = barriers (individual, social, environmental); SE = self-efficacy; InM = intrinsic motivation; IdM = identified motivation; EnSM = Enhanced self-management skills

Health Education (HE).

Health education sessions serve as a contact control for CBEX and are 30 minutes in length and group-delivered by a research therapist. HE topics include: nutrition; caffeine; colds, germs, and flu; preventing cancer; diabetes; heart health; sleep; complementary and alternative medicine; and being a smart patient. We use a health education manual previously developed for other studies and used in educational workshops serving people with depression [55]. Classes are interactive in nature and do not include components designed to increase PA through problem-solving or other skills training. Because participants may come in with varying levels of health literacy, the interventionist tailors presentation of information to participants’ level of knowledge of the topic. That is, the interventionist may present only basic information if the class seems relatively uninformed about a topic, or they are prepared to provide more in-depth information, including results from new research studies, if the class seems relatively well-informed. If there seem to be disparate levels of knowledge, the interventionist is instructed to attempt to balance the needs of the class participants.

Project MOVE research goals and hypotheses

The primary outcome of this study is minutes of MVPA per week across the three intervention conditions at end of intervention (3 months). We hypothesize that BA + SHE + CBEX will be superior to BA + SHE + HE, and that BA + SHE + HE will be superior to BA alone in terms of objectively measured MVPA levels. Hypotheses for the 6- and 9-month long-term follow-up mirror those for intervention endpoint. Secondary outcomes will include depression and cardiorespiratory fitness. As above, we hypothesize that BA + SHE + CBEX will be superior to BA + SHE + HE, and that BA + SHE + HE will be superior to BA alone on these outcomes. Further, this study will examine the mediating role of minutes of MVPA on the association between intervention condition and change in depressive symptoms. In addition, theory-relevant constructs such as benefits of exercise, perceived barriers to exercise, exercise self-efficacy, and intrinsic and extrinsic motivation to exercise will be examined as mediators of the association between condition and minutes of MVPA. Lastly, study analyses will assess whether social and environmental barriers to exercise predict exercise maintenance.

Sample size and power calculation

Kraemer et al. [59] recommend that RCTs be powered to detect minimally clinically significant differences. To determine what a minimally clinically significant effect size for minutes of MVPA per week might be, we looked at the literature to estimate a standard deviation in depressed individuals. Using statistics from Hoffman et al. [60] in their report of 1-year follow-up of depressed individuals after an exercise intervention, a sample standard deviation of 80 minutes of MVPA per week was estimated. Next, differences in minutes per week of MVPA that might be considered clinically significant were reviewed--these varied widely. First, there is a dose-response relationship between exercise and physical health: even 60 minutes per week of moderate exercise (compared to no exercise) is associated with a lower risk of mortality and onset of coronary heart disease [2]. Second, in their follow-up of depressed individuals after an exercise intervention, Hoffman [60] found that 180 minutes of exercise (vs. 0) corresponded to a difference in endpoint Hamilton Rating Score for Depression score difference of about 3 points. A 3-point difference is considered “a minimally clinically significant difference” by the National Institute for Clinical Excellence in the UK [61] although others have argued that a smaller difference between groups (such as 2 points) may be clinically significant as well [62]. Third, Babyak et al. [63] followed depressed individuals after an exercise of 64% vs. 48%; this difference seemed clinically significant in that it translated to depressive relapse rates of 30% vs. 52%. Extrapolating from these studies, one could estimate that a minimally clinically significant effect size for the difference in MVPA minutes per week might range from Cohen’s d of .20 to .75 to 1.5 or even 2.2 [64]. Given this wide range, it was concluded wise to power the study at a lower end of the range (e.g., d = .50, considered a medium effect size). Assuming a standard deviation of 80 minutes of MVPA per week, this corresponds to a difference between groups of 40 minutes of MVPA per week.

Power analysis is based on the primary hypothesis of this study that the most intensive intervention, BA+SHE+CBEX, will be superior to BA + SHE + HE, and that BA + SHE+ HE will be superior to BA alone in terms of MVPA per week. The total sample of 240 (n=80 participants per group at baseline, with 64 per group after 20% attrition) is sufficient to have 80% power to detect a medium-sized difference between arms using a two-tailed significance test level with α =.05.

Study procedures

Participant recruitment and screening

Recruitment strategies include web-based advertisements (e.g., Facebook, Craig’s list), a web page, flyers and brochures in the community, primary care, and in psychiatry practices. Interested individuals either call the study telephone to be screened by a research assistant or submit on online request through the study website to be contacted. We track all inquiries, and each inquiry is followed until resolution (i.e., unable to reach, not eligible, enrolled in the study). Inclusion criteria are: 1) adults ages 18–65; 2) low physical activity, (i.e., have not participated in 90 minutes or more of MVPA for more than 6 weeks in the last 12 weeks); 3) able to walk one mile required for cardiorespiratory fitness test; 4) medically-cleared for exercise with no significant medical condition or physical disability that would interfere with physical activity or study participation; 5) weight that does not exceed 375 lbs, due to weight limits on exercise machines; 6) Quick Depression Symptomology, QIDS [65] score greater than or equal to 8 and less than 20, and including endorsement of either sad mood or anhedonia; and 7) able to make one of the two available exercise class times. Participants are excluded if they have: 1) a current anorexia or bulimia (past 3 months); a lifetime history of bipolar disorder, schizophrenia or a chronic psychotic condition; or major depressive disorder with psychotic features in the past six months; 2) moderate or severe substance use disorder in past 6 months; 3) currently pregnant or planned pregnancy in the next year; 4) current suicidality requiring immediate treatment; and 5) unable to consent and complete study assessments. Participants must be able to understand English well enough to consent and complete assessments. Finally, if participants are aged 60 or older, or if there is any concern about their cognitive capacity, participants will undergo a brief cognitive functioning screener, the Montreal Cognitive Assessment (MoCA [66]). Participants must achieve a score within 1 SD of the mean for their age and education level to participate in the study (66).

Informed consent, enrollment and randomization

Individuals who express interest in the study and appear eligible based on a brief telephone screening interview are scheduled for an in-person appointment (i.e., baseline day 1). On baseline day 1, the study research assistant provides a description of study procedures, which includes detailed information about each of three study arms and an explanation of the randomization process, and then completes informed consent with the participant. The participant is given an opportunity to read through the consent form in detail and to ask any questions they may have about the study. After informed consent, additional assessments are collected through interviews and questionnaires, and eligibility for study participation is confirmed. Furthermore, a release of information is obtained to contact the individual’s primary care physician (PCP) to request medical clearance for study participation. Upon receiving medical clearance from the PCP, participants are scheduled for a second in-person appointment (i.e., baseline day 2).

During baseline day 2, participants engage in a 1-mile Rockport treadmill walk test. After research assistants review all study procedures with them, and ascertain that they still wish to participate, participants are then randomized to one of the 3 intervention conditions using a 1:1:1 ratio. Participants are aware that they may be assigned to one of three conditions, and, because this is a behavioral intervention, are not blinded to the condition to which they are assigned.

In order to increase the likelihood of balance in baseline depression severity between treatment arms, we stratify randomization [67] based on depression severity scores, QIDS ≤ 14 vs. QIDS ≥ 15, and use blocking within strata. Prior to the start of the study, the study statistician generates the two randomization sequences (one for each strata). These lists are then uploaded into REDCap [68] by a staff member who does not have any participant assessment responsibilities, and the REDCap database is put into production. One put into production, staff cannot access the lists. Thus, when a participant is enrolled and ready to be randomized, the study staff member indicates to which strata the participant belongs and then presses a button to learn to which arm the participant is allocated.

Data collection and blinding

Data are collected on the same schedule across the three arms. Participants complete assessments at baseline prior to randomization, at Month 1.5, at Month 3 (end of intervention phase), and at 6 month and 9 follow-ups. Moreover, there are brief mood assessments before and after each exercise session during the 3-month intervention. Participants wear an Actigraph (GT3X model [69]) accelerometer for 1 week at baseline, Month 1.5, Month 3 (end of intervention phase) and 6 and 9-month follow-ups. Follow-up assessments of depression symptoms, which are conducted via a structured interview, are completed by research staff who are blind to participants’ intervention arm.

Data management

All data are stored in REDCap. REDCap is a secure, web-based application that is designed to support online or offline data capture for research studies, quality improvement, and operations [68]. REDCap provides easy data manipulation (with audit trails for reporting, monitoring, and querying patient records), real-time data entry validation, and an automated export mechanism to common statistical packages [68]. REDCap’s data are stored on encrypted servers within the data center of our institution. Participants complete self-report forms directly into REDCap. Other data, such as data from interviews or performance-based assessments, are entered by study research staff. Data from our interview-based secondary outcome (depression symptoms) are double-entered in order to verify that they have been correctly entered. Accelerometer data are not stored in REDCap, as they are processed using the ActiLife software [70, 71] for each participant and then exported. This process is explained below. All paper documents are stored in locked file cabinets in the research space, with consent forms and other forms with personal identifiers being stored separately from data without identifiers.

Measures

Table 2 provides information about the study assessments and timeline. The reading level for all self-report measures is grade 8 or below. These instruments have been used widely in previous exercise interventions.

Table 2: Schedule of Assessments	Construct Assessed	Mode	Timepoints
INSTRUMENTS TO ASSESS INCLUSION CRITERIA and CHARACTERIZE SAMPLE
Structured Clinical Interview for DSM-V Axis I Disorders, Patient Version (SCID)	Mood, eating, substance use, and psychotic disorders	Interview	BL1
Demographics		Self-report	BL1
General Health history	General health	Self-report
MEASURES OF TREATMENT UTILIZATION
Treatment Response to Antidepressant Questionnaire (TRAQ)	Antidepressant medication use	Interview	BL1, M1.5, M3, M6, M9
Treatment History Interview (THI)	Psychotherapy, other tx	Interview	BL1, M3, M6, M9
OUTCOME MEASURES
Actigraph (GT3X model)/ Minutes of at MVPA per week	Amount of physical activity **PRIMARY OUTCOME**	Actigraph	BL2, M1.5, M3, M6, M9
International Physical Activity Questionnaire (IPAQ)	Amount of physical activity	Interview	All
Heart Rate, and Blood Pressure	Resting Heart Rate, Blood Pressure	EP	BL2, M3, M6, M9
Rockport Treadmill exercise test,	Cardiorespiratory fitness	Treadmill	BL2, M3, M9
Quick Inventory of Depression Symptoms – Clinician Rating (QIDS)	Depression severity, suicidality	Interview	All
Patient Health Questionnaire (PHQ-9)	Depression severity	Self-report	BL1, M1.5, M3, M6. M9
Short-Form Health Survey (SF-36)	Physical functioning, overall health, pain	Self-report	BL1, M1.5, M3, M6, M9
Systematic Assessment of Treatment-Emergent Events – General Inquiry (SAFTEE-GI)	All adverse events (#, severity, impairment)	Interview	M1.5, M3, M6, M9
Positive and negative affect schedule (PANAS)	Affect	Self-report	BL1, M1.5, M3, M6, M9
State-trait anxiety inventory (STAI)	Anxiety	Self-report	BL1, M1.5, M3, M6, M9
Brief Irritability Test (BITe)	Irritability	Self-report	BL1, M1.5, M3, M6, M9
HYPOTHESIZED MEDIATORS
Motives for Physical Activity Measure	Perceived exercise benefits	Self-report	BL1, M1.5, M3, M6, M9
San Diego Health and Exercise Questionnaire	Barriers to exercise	Self-report	BL1, M1.5, M3, M6, M9
Behavioral Regulation in Exercise Questionnaire (BREQ-3)	Intrinsic & identified exercise motivation	Self-report	BL1, M1.5, M3, M6, M9
POTENTIAL PREDICTORS OF MVPA
Social support from family; Social support from friends/ work colleagues	Social support for PA	Self-report	BL1, M3
Personal safety; aesthetics; walking environment	Physical environment	Self-report	BL1, M3

Note. BL = baseline; other numbers in the timepoint column indicate week (wk) and month (M). EP = Exercise Physiologist

Physical activity measures

The primary outcome measure is accelerometry-determined minutes of MVPA per week. Although public health recommendations for MVPA per week are 150 minutes, increased MVPA that that falls short of this target may still have an impact on depression symptoms [72] and other physical health outcomes [1]. Further, there may be some people for whom this 150-minute target is not feasible. Therefore, MVPA minutes per week rather than the degree of adherence to a 150-minute-per-week target is the primary outcome for the Project MOVE trial.

MVPA minutes is objectively measured using an ActiGraph accelerometer, which is a widely used and valid instrument in exercise research [70, 71]. The device is small and weighs 1.5 oz, and participants wear it on their nondominant wrist. The accelerometer collects data 30 times per second and sums it up across a period of time (epoch). There are numerous options for data output, including number of bouts of moderate (or vigorous) physical activity per week, and minutes of each bout. Consistent with standardized accelerometer data reduction approaches [70], a cut point of ≥ 1, 952 PA activity counts per minute will be used to quantify MVPA minutes. Further, only bouts of 10 minutes or longer of MVPA will be counted when calculating the weekly MVPA. Participants receive instructions on proper wear of this device at each assessment timepoint, along with a log for tracking the time that they wear the accelerometer. Participants have the option of receiving text message reminders to wear the accelerometer. Ten hours of data is be considered a valid day [70, 71]. Data from these devices are used to create a variable of minutes of MVPA per week.

Moreover, as a secondary outcome, participants report their level of physical activity based on the International Physical Activity Questionnaire or IPAQ [73]. This 7-day item instrument is designed to measure total minutes and level of intensity of PA over past seven days [73, 74]. This measure also measures average minutes of sedentariness in the past seven days [73,74].

Cardiorespiratory fitness measure

The Rockport 1-Mile Walk Test is a measure of cardiorespiratory fitness for adults and has been validated for use on a treadmill [75, 76]. Rather than requiring an outdoor or indoor track to complete this standardized protocol, this more convenient adaptation of the 1-mile walk test involves participants walking on a treadmill for one mile at the fastest pace that allows them to comfortably complete the mile. Heart rate and blood pressure are monitored at scheduled intervals [76]. Duration of the walk test and their heart rate upon completion are utilized to calculate the VO2peak – a measure of oxygen consumption and indication of fitness [75, 76]. Rating of Perceive Exertion (RPE) is used throughout the walk test to record the individual’s level of intensity of PA up to his or her age predicted maximum heart rate [76, 56]. Following test termination, each participant undergoes a 3-minute active cool-down and 5–7 minute seated cool-down, during which time heart rate is assessed at each minute with blood pressure measured at 2-minute intervals [76, 56]. Cardiorespiratory fitness is a secondary outcome.

Theory-based measures

Participants complete self-report questionnaires to assess potential mediators and predictors of increased physical activity. The measures are found in Table 2. The Motives for Physical Activity (MPAM) scale consists of 23 items assessing perceived exercise benefits [77]. The San Diego Health and Exercise Questionnaire [78, 79] assesses barriers to exercise. As per the Health Belief Model, both increased exercise benefits and decreased barriers may mediate the impact of SHE and CBEX on amount of physical activity. The Behavioral Regulation in Exercise Questionnaire (BREQ-3) [80], a 19-item scale, is used to assess self-determined motivation in physical exercise. The BREQ-3 measures external, interjected, identified and intrinsic regulation [80]. Consistent with Self-Determination Theory, increased intrinsic and identified motivation may mediate the impact of CBEX on outcome.

We chose potential predictors of physical activity based on the Social Ecological Model of PA. Participants complete a Social Support and Exercise scale, which consists of 23-items that measure individuals’ perceived social support for PA from their family members and friends (two item examples are “[they]helped plan activities around my exercise,” or “[they complained about the time I spend exercising”) [79]. The other is Physical Environment for Exercise, a 33-item scale that assesses features of an individual’s physical environment or neighborhoods that are associated with PA behaviors (e.g., “my neighborhood offers many opportunities to be physically active”) [81].

Clinical measures

The Quick Inventory of Depression Symptoms-Clinician Rating (QIDS) is a commonly-used clinical interview that yields a depression severity score between 0 −27 [65]. It has been used in other NIH-funded depression clinical trials and has good psychometric properties [65, 82]. Research staff are initially trained to competence on the QIDS by one of the study PIs. Subsequently, once every 6 weeks, this PI, the first author, and all interviewers listen to and rate a QIDS interviews in order to obtain a measure of inter-rater reliability. Once all ratings are finalized, raters discuss ratings in order to ensure consistency for future ratings, The Structured Clinical Interview for DSM-V-Patient version (SCID) is used to assess inclusion/exclusion criteria, including mood disorders, psychotic symptoms, substance use disorders, and eating disorders [83]. The Patient Health Questionnaire (PHQ-9) is used as a self-report measure of depression symptom severity [84]. The QIDS and PHQ-9 are secondary outcome measures; the SCID and QIDS are used to assess inclusion criteria at baseline. The QIDS is repeated at month 1.5, end intervention, at follow-ups (6 & 9 months).

Adverse events

Participants are instructed at enrollment to report any injury or adverse events to the research staff. In addition, we administer the Systematic Assessment of Treatment-Emergent Events – General Inquiry (SAFTEE-GI, [85, 86] at follow-up assessments in order to monitor adverse events in all 3 arms. This interview includes a general inquiry about the presence of physical or health problems, regardless of cause, as well as their onset, duration, pattern, current status, severity, and functional impairment. Summary scores, including number of events elicited, average severity, and functional impairment will be calculated.

Suicidality.

Risk for suicidality is assessed formally every month during the intervention phase and every 3 months during the follow-up phase of the study (as part of the PHQ-9 and QIDS). In this study, suicidality is defined as a QIDS item 12 (suicide item) score ≥ 2 [65], or a PHQ-9 item 9 (suicide item) score ≥ 1 [84], or any report of suicide ideation or behavior to a member of study staff. At the time that a risk for suicidality is identified, a study clinician immediately evaluates that person to determine whether a higher level of care is needed.

Clinical deterioration.

At each assessment after baseline, trained research assistants identify participants with significant clinical deterioration, defined by a QIDS score ≥ 16 and an increase of ≥ 6 points from the previous timepoint [65, 84]. If research staff identify a participant with clinical deterioration, a study clinician immediately evaluates that participant to determine if a higher level of care should be recommended.

Other measures

Other secondary outcome measures include the Short-Form Health Survey (SF-36), which assesses physical functioning and overall health [87], the Positive and Negative Affect Schedule (PANAS, [88]) the State-Trait Anxiety Inventory [89], and the Brief Irritability Test (BITe; [90]). To characterize the sample, we use the Treatment History Interview (THI [91, 92]) and the Treatment Response to Antidepressant Questionnaires (TRAQ, [93,94]) assessing, respectively, history and current and past pharmacotherapy treatment. Demographic information including age, gender, race, ethnicity, marital status, education, employment status, and income is collected at baseline via self-report. All the study measures are found in Table 2.

Participant compensation

Participants are compensated $30 for completing assessments at baseline visit-1, baseline visit- 2, and $50 for the Month 1.5 assessment. Subsequently, participants receive $70 compensation for each completed assessment (months 3, 6, and 9). Participants also receive $20 each time that they return an accelerometer (for up to a total of $100). Participants are not compensated for attending intervention sessions.

Retention

In addition to monetary incentives, other techniques to increase retention for assessments are: repeated discussion of study procedures prior to randomization to make sure that the participant is prepared to engage in all study procedures; a personalized calendar of assessments; reminder calls prior to assessments; collection of contact information for an individual who will always know how to contact the study participant for use if we cannot reach the participant; provision of transportation for assessments if needed; and use of phone/ mail to conduct assessments if necessary. Even when participants drop out from their assigned treatment arm, every effort is made to conduct the assessments. We track retention and review retention at weekly study meetings.

Planned analyses

All planned analyses will be conducted on the intent to treat sample; thus, all randomized participants will be included in the analysis. Descriptive statistics will be used to summarize the baseline demographic and clinical characteristics of the sample.

Primary Outcome (Minutes of MVPA in the intervention phase).

Using the mi impute and estimate facilities in Stata 16.0 [95], we will generate 50 fully populated data sets using multiple imputation by chained equations (MICE [96]). MICE is a methodologically principled method to handle missing data under the assumption that data are missing at random (MAR). Variables in the imputation model will include treatment condition, age, gender, available observations of accelerometry-determined minutes of MVPA per week, self-reported minutes of MVPA per week, and to support the assumption of MAR, significant predictors of outcome and attrition. The robustness of our inferences will be evaluated by performing sensitivity analyses in which missing values are systematically replaced by alternative estimates of MVPA.

To test the primary aim that the most intensive intervention, BA+SHE+CBEX, will be superior to BA + SHE + HE, and that BA + SHE + HE will be superior to BA alone during the intervention period, assessments of group differences in MVPA as measured by the accelerometer will be conducted using general linear mixed models (GLMM [97–100]. Models will estimate the average effect of treatment group on minutes of MVPA during the intervention period (assessed at 1.5 and 3 months). Covariates will include age, gender, an indicator variable representing timepoint of assessment, and minutes of MVPA observed per week at baseline. We chose age and gender as covariates since both are typically associated with amount of physical activity (101, 102). Tests of individual hypotheses will be conducted only if the overall Wald χ2-test for between group differences is significant at the .05 level.

Effect Sizes.

Given that minutes per week of MVPA is the primary outcome and is a meaningful metric, we will be able to report between group differences in minutes per week of MVPA along with 95% confidence intervals. To facilitate comparison to other studies, we will also report standardized mean differences or Cohen’s d (103) at follow-up timepoints.

Secondary Outcomes.

Analogous GLLM models will be used to investigate the impact of treatment group on other outcome measures over time, including MVPA in the follow-up phase, and self-reported minutes of MVPA per week, health outcomes, physical fitness (i.e., Peak V0₂₎, depression symptoms, irritability, anxiety, positive and negative affect, and suicidality or clinical deterioration in all phases.

Informed by Self-Determination Theory and the Health Belief Model, we hypothesize that perceived benefits of exercise, perceived barriers to exercise, exercise self-efficacy, intrinsic motivation to exercise, and identified motivation to exercise will mediate the association between treatment arm and amount of MVPA. In addition, we expect that amount of MVPA will mediate the association between treatment arm and depression symptom severity. Tests for mediation using methods suggested by MacKinnon and colleagues will be conducted [100]. We recognize that our ability to make statistical inferences or to test mediational models that might account for the potential inter-dependencies among variables will depend upon outcomes from tests of other aims. Accordingly, results that suggest potential mechanisms that can be more fully tested in structural equation models in future studies will guide the abovementioned analyses.

Finally, analyses of Social Ecological Theory constructs at baseline (e.g., environmental factors) as predictors of MVPA over time will be conducted. Specifically, these predictors will be included as covariates in the analytic models for the primary outcomes to assess their impact on outcome.

Discussion

The Project MOVE randomized clinical trial uses combined intervention approaches to attempt to address various barriers that may hinder physical activity engagement and the maintenance of MPVA among depressed populations. Specifically, this study: a) assesses whether a cognitive-behavioral psychological intervention, combined with an exercise program to increase and maintain PA is superior to an exercise program alone among adults with depression; b) tests the efficacy of a partially home-based, partially supervised exercise program to increase MVPA (vs. a minimal treatment control); and c) seeks to determine if these interventions can help to maintain higher PA levels beyond the time period of the active exercise intervention phase.

There have been many PA interventions conducted to date with depressed groups; however, the overall adherence rate remains poor. We examined papers included in a Cochrane review of exercise for depression [104] that: 1) enrolled individuals with a depression diagnosis; 2) included at least 15 individuals in the exercise arm; and 3) reported on adherence. This yielded 7 studies [25, 28, 105–108]. We added more recent and relevant papers not included in the Cochrane review [26, 109]. The average weighted adherence rate across studies during the intervention phase was 73% [25, 26, 28, 105–109]. Adherence ranged from 38% to 100%, with some of the higher quality studies showing lower adherence rates (e.g., [26, 28, 109]). For example, Trivedi et al [26] compared a low dose of exercise to a “public health dose” (i.e., 150 minutes of MVPA per week) for depressed individuals over 3 months. In the public health dose arm, in the first 6 weeks, average adherence was 85% of the public health dose, but in the next 6 weeks, adherence was only 54%. Further, for the 2nd 6-week period, the adherence rate does not include data from those who dropped out of the assessments. Thus, multiple approaches may be necessary to promote consistent adherence to a PA program among depressed individuals.

As previously mentioned, Project MOVE uses combined intervention approaches that address central barriers to PA, including core depression symptoms (e.g., amotivation, poor problem-solving skills), that may impede participant engagement or adherence as well as motivation for PA among depressed individuals. The CBEX intervention has been developed to help individuals cultivate problem-solving skills based on reported barriers and facilitate successful engagement in PA. CBEX also focuses on helping participants to develop intrinsic motivation to exercise. This type of approach has shown to be effective in increasing PA engagement among physically inactive adults [110–112]; although, to date, most of these studies have been conducted with non-depressed populations.

Moreover, there is limited knowledge on how to promote long-term MVPA, which has shown to be associated with numerous health and mental health benefits, among depressed patients. Because the benefits of exercise may not persist once exercise is discontinued [25], understanding whether exercise interventions lead to continued longer-term maintenance of PA in this population is of critical importance. Our review of the literature shows that only two studies in clinically depressed individuals have looked at the critical period following a formal exercise program to see if exercise was, in fact, maintained. Babyak et al [33] reported on the 6 months following the exercise program and found that 64% of participants in the exercise arm, compared to 48% of participants in the medication comparison arm, reported that they exercised during follow-up. There was no information on amount of exercise. The exercise arm showed a significantly lower rate of depression relapse during follow-up than the medication arm (30% relapsing vs. 52% relapsing). In contrast, Hoffman et al [34] noted that 50% of participants reported that they engaged in no exercise at follow-up, and they also found no difference in exercise rates between groups initially assigned to exercise, sertraline, or placebo. However, more exercise during the follow-up period was associated with a greater likelihood of at least partial remission from depression at 1-year. In sum, both studies suggest that, if exercise is maintained after an initial intervention period, it may be associated with better depression outcomes over time. However, the two studies are mixed as to whether, in depressed individuals, an exercise program can have a long-term impact on amount of exercise after the program. An MVPA program that includes some home-based exercise may have a longer-term impact because participants are exercising in their home environment right from the start of the program.

There are numerous limitations to previous research on PA. First, regarding descriptions of adherence in previous research, with a few exceptions (e.g., [25]), it is often not clear whether participants considered to be dropouts were included in the calculations of adherence. Second, adherence has typically been reported in terms of number of supervised sessions attended during an intervention phase. This significantly limits the extent to which we understand how much total PA (including home-based activity) a person is actually doing. As such, examining minutes MVPA per week (rather than session attendance) as the primary outcome provides needed information on the extent to which depressed individuals exercise in their day-to-day lives. Third, when measuring MVPA outside of an exercise session, objective measures (e.g., accelerometers) in addition to self-report information are necessary. Fourth, no studies in depression (and few studies not in depression) have examined whether theory-based constructs, such as those in the Health Belief Model and Self-Determination Theory, mediate the relationship between intervention and amount of PA. Further, there is little research on how the physical environment in which depressed individuals reside may further impact lack of PA as well as adequate strategies to increase behavior change. In this study, we attempt to address all these limitations.

If efficacious, the flexible approaches used in Project MOVE to promote engagement and maintenance of MPVA could be beneficial among populations from varied racial or ethnic backgrounds. For example, interventions that consider the demands or responsibilities that individuals have (e.g., work, caring for others) that may hinder their engagement in PA are consistent with recommended guidelines toward decreasing disparities in health behaviors among Black or African American populations [2, 113]. In both BA and SHE, participants work with a trained research staff to choose activities as well as times and places for exercise that are best suited for their own preferences and lifestyle. Individuals who are caretakers or have children and report caregiving responsibility as a barrier to exercise are encouraged to choose exercises that incorporate their children or other family members when possible (e.g., walking or biking at their local park, or hiking).

Physical environmental barriers (e.g., lack of sidewalk or parks in the neighborhood) have also been shown to be associated with low rates of PA among some minority communities [113, 114]. In BA, participants are guided to identify alternative places for exercise (e.g., nearby parks, local shopping malls and their walking hours), or other resources (e.g., short exercise videos that can be done at home) to increase PA. In CBEX, participants are asked weekly about environmental barriers (and other barriers) that prevented them from exercising, and exercise facilitators. Then, as a group, they problem-solve to generate strategies to get back on track or remain regularly active.

Furthermore, Project MOVE may appeal to minority populations since it consists of a single weekly supervised exercise session; this is less burdensome than multiple supervised sessions. Interventions that require multiple week sessions may not be feasible for people with multiple caregiving responsibilities or difficulty accessing transportation. Moreover, since minority populations tend to lack physically active role models [113, 114], the supervised group exercise would allow them to be around other people with depression and thus could facilitate a sense of community and support.

Summary

The significant morbidity associated with depression, a low level of PA is an important modifiable lifestyle behavior that could significantly improve the health and overall well-being of individuals with depression and thereby decrease the disease-related economic burden. However, we know little about approaches to increasing PA that are accessible to depressed patients, and that can lead to longer-term maintenance of PA in this population. Depressed individuals may experience high levels of barriers to physical activity that intrinsic to the experience of depression: anhedonia, amotivation, fatigue, poor self-worth, and poor problem-solving skills for coping with barriers. We believe that adding a theory-based cognitive-behavioral counseling intervention to structured exercise programming may increase physical activity when compared to just a structured education program. Further, we plan to assess the impact of exercise interventions after the acute intervention phase has passed. In sum, the proposed study will address existing knowledge gaps regarding specific intervention approaches to increasing physical activity in depressed individuals – a population that could significantly benefit from both the physical and mental health benefits of long-term engagement in exercise. Moreover, if effective, these approaches could be applied to minority populations with depression.

Figure 1.

Intervention Conditions & Follow-ups

Abbreviations:

(PA)

Physical activity

(MVPA)

Moderate-to-Vigorous Physical Activity

(BA)

Brief advice

(SHE)

home-based exercise

(CBEX)

Cognitive-Behavioral Sessions Focused on Increasing and Maintaining Exercise

(HE)

Health Education

(QIDS)

Quick, Depression Symptomology

(PCP)

Primary Care Physician

(IPAQ)

Physical Activity Questionnaire

(MPAM) scale

Motives for Physical Activity

(BREQ-2)

Behavioral Regulation in Exercise Questionnaire

(PHQ-9)

The Patient Health Questionnaire

(SAFTEE-GI)

Systematic Assessment of Treatment-Emergent Events-General Inquiry

(SF-36)

Short Health Survey

(THI)

Treatment History Interview

(TRAQ)

Treatment Response to Antidepressant Questionnaires

(MAR)

Missing at Random