Skip to main content
NCBI home page
The Gerontologist logoLink to The Gerontologist
. 2023 Mar 14;63(9):1456–1466. doi: 10.1093/geront/gnad028

A Social Exergame Intervention to Promote Physical Activity, Social Support, and Well-Being in Family Caregivers

Xin Yao Lin 1,2,✉, Lin Zhang 3, Saiyeon Yoon 4, Ruoying Zhang 5, Margie E Lachman 6
Editor: Suzanne Meeks
PMCID: PMC10581379  PMID: 36916022

Abstract

Background and Objectives

Family caregivers often experience a high level of stress, social isolation, a sedentary lifestyle, and poor mental and physical health. An exergame intervention was developed to promote physical activity and well-being in family caregivers and to test social support as a mechanism for behavior change.

Research Design and Methods

The current study was a randomized pilot trial (N = 76) to compare the effectiveness of Go&Grow (social vs nonsocial exergame) to promote well-being through increased social support and physical activity for family caregivers over a 6-week intervention.

Results

The treatment group increased significantly more than the control group in well-being (management of distress) and social support (satisfaction with contact quality). Social support served as a mechanism (mediator and moderator): The treatment group increased more than the control group in satisfaction with social contact quality, which led to more positive affect and less loneliness. Moreover, those in the treatment group who increased more in overall social support and knowing others’ experiences increased their steps more than those with less support, whereas the change in steps for the control group was not related to a support level. Those in the treatment group who used more social features of the app had a greater increase in steps compared with those who used it less.

Discussion and Implications

Social support in technology interventions is a promising direction to promote caregivers’ well-being and physical activity. Social support served as a mechanism of behavior change that can inform more engaging, sustainable, portable, and scalable interventions in the future for sedentary and socially isolated family caregivers.

Clinical Trial Registration Number: NCT05032872

Keywords: Behavior change, Fitness technology, mHealth, Smartphone-based intervention

Family caregivers provide care to their loved ones with disabilities or illnesses (National Alliance for Caregiving, 2020). The aging baby boomer population is increasingly requiring more care, with over 41 million Americans providing care for adults over the age of 50 in 2020 (National Alliance for Caregiving, 2020). Family caregivers play essential roles in their family settings, and they have responsibilities for the welfare of the loved ones they care for. Family caregivers may experience stress daily, including difficulties related to caregiving, and other aspects of their lives. They can also experience social isolation, a sedentary lifestyle, and poor mental and physical health (Bressan et al., 2020; Zarit et al., 2014). In addition, both physical inactivity and lack of social support can lead to a heightened risk of poor health and low well-being in later life (Cunningham et al., 2020; Lachman & Agrigoroaei, 2010). However, family caregivers often have limited time and energy to seek support or engage in physical activity. In addition, an overload of responsibilities can increase their stress level, and minimize caregivers’ positive affect while maximizing their negative affect (Bressan et al., 2020; Zarit et al., 2014). Less social support is also related to more perceived loneliness, which can elevate the likelihood of morbidity and mortality for lonely family caregivers (Kelly et al., 2016; Zarit et al., 2014).

There is a vital need for sustainable, scalable, and enjoyable programs to support social, physical, and emotional well-being among family caregivers and to help manage their own health and well-being so they can provide more effective care for their care recipients (Koumakis et al., 2019; Waligora et al., 2019).

Social Exergames to Address Caregiver Needs

Social exergames are digital games involving real-world physical activity and social activities with others. An exergame study conducted in an assisted living setting reported encouragement and increases in social connectedness for older adults while playing the exergame (Brox et al., 2014). Playing games with others is associated with more satisfaction and enjoyment, and social interactions in online exergames can motivate older adults to exercise more and prevent loneliness (Brox et al., 2014; Li et al., 2018). Many prior exergame studies have been developed and tested with digital coaching systems, whereby exercises are demonstrated in a stationary video gaming context (e.g., using the Nintendo Wii) with younger adults or nursing home residents (Chao et al., 2015; Li et al., 2018). However, given the limited time that family caregivers have for their own activities, interventions must be sensitive to their time and location constraints (Kelly et al., 2016). Moreover, past exergame studies focused on in-person game socialization rather than virtual social experiences for family caregivers (Brox et al., 2014; Chao et al., 2015). However, given the limited time that family caregivers have for their own activities, interventions must be sensitive to their time and location constraints (Kelly et al., 2016; Stowell et al., 2019). Additionally, past studies have mainly focused on physical activity, whereas other outcomes such as social support and well-being are also important to consider (Cunningham et al., 2020; Lachman & Agrigoroaei, 2010). These limitations and literature gaps leave a strong need for social exergames that can provide family caregivers with enjoyable and convenient mobile gaming activities that promote their physical activity, social support, and well-being.

Go&Grow is a garden-themed social exergame app for caregivers to plant and grow flowers based on their daily steps and exercises. An initial Go&Grow pilot study tested 18 caregivers of family members with Alzheimer’s disease and related disorders (Lin et al., 2020) using a nonrandomized single-arm trial (Stage 1a of the National Institute of Health [NIH] Stage Model for Behavioral Intervention). Results showed an increase in caregivers’ physical activity and social contact with other caregivers, and a decrease in their caregiving stress (Lin et al., 2020; Stowell et al., 2019). Online social engagement could allow caregivers to interact with others for an exchange of information, guidance, and encouragement on experiences and opinions on physical activity and well-being (Hopwood et al., 2018; Li et al., 2018). Such social engagement could foster caregivers’ social support, physical activity, and well-being, and may reduce the negative experiences in the caregiving burden (Hopwood et al., 2018; Newman et al., 2019).

The Current Study

The NIH Science of Behavior Change (SOBC) emphasized examining the underlying mechanisms and factors associated with behavioral changes (Nielsen et al., 2018). However, past exergame studies, including the Go&Grow pilot study (Lin et al., 2020; Stowell et al., 2019), had a small sample size (Brox et al., 2014; Unbehaun et al., 2018), and no comparison groups for testing the feasibility and efficacy of the exergames and the underlying mechanisms (Chao et al., 2015; Li et al., 2018). The current study explored the feasibility and efficacy of Go&Grow in a broader sample of sedentary family caregivers with a randomized two-arm pilot trial (Stage 1b of the NIH Stage Model for Behavioral Intervention) to test the purported mechanism, social support, with a treatment (Go&Grow full version) and an active control group (Go&Grow without social features) to promote their physical activity, social support, and well-being. Also, in line with the goal of the SOBC program (Nielsen et al., 2018) and by having an active control group, the current study tested social support and physical activity as mechanisms of change for well-being.

To our knowledge, the current study is the first randomized pilot trial testing the underlying mechanisms linking an exergame intervention with key outcomes (physical activity, social support, and well-being) for family caregivers. The current exergame study can offer new insights into implementing behavior change mechanisms to create engaging, sustainable, and scalable future interventions and solutions for sedentary and socially isolated family caregivers who can improve their health with modifiable factors such as physical activity and social support.

Hypotheses

We hypothesized that compared with the control group (nonsocial exergame), the treatment group (social exergame) would increase more in the following:

  • 1) Well-being (increase in positive affect, and decrease in loneliness, stress, and negative affect from pretest to post-test and baseline to the end of intervention [weekly] and decrease in each subscale for caregiver stress from pretest to post-test).

  • 2) Social support (overall social support from pretest to post-test and each subscale of the social connectedness scale from baseline to the end of intervention [weekly]).

  • 3) Physical activity (steps and minutes active measured by Fitbit from pretest to post-test and baseline to the end of intervention [weekly], and each subscale for the self-reported physical activity from pretest to post-test).

  • 4) Secondary outcomes from the pretest to post-test (life satisfaction, exercise efficacy, and sense of control).

It was also predicted that increased social support and physical activity at the weekly level would serve as mechanisms linking the intervention to better well-being from pretest to post-test in the treatment group compared with the control group.

Method

The current study was approved by the university Institutional Review Board (20104R). Informed consent was obtained from each participant before the study. The clinical trial was registered at the U.S. National Institutes of Health (ClinicalTrials.gov) # NCT05032872.

Participants

Power analyses showed that 56 people were required for a power of 0.80 at p < .05, with a medium effect size for the analysis of changes in outcome measures from the pretest to the post-test. Recruitment took place from August 2021 to January 2022 by outreaching to the caregiver and older adult organizations via social media and participant recruitment sites (e.g., caregiver forums, registry). Inclusion and exclusion criteria are shown in Supplementary Figure 2. Participants kept the Fitbit and received $25 for completing all aspects of the study.

Figure 1 shows the CONSORT diagram. A total of 80 participants gave their oral informed consent and were randomized into either the control condition (n = 42) or the treatment condition (n = 38). Of the 80 randomized participants, 76 participants completed the pretest, and 72 participants started the intervention in one of the three cohorts (Cohort 1: n = 22, Cohort 2: n = 22, Cohort 3: n = 28) to make the social groups manageable, with an equal number of people in each condition for each cohort. Sensitivity analyses showed that all baseline characteristics (except age and health) of the three cohorts did not differ (Supplementary Table 1).

Figure 1.

Figure 1.

Open in a new tab

CONSORT diagram of the study. A total of 14 people dropped out of the study after randomization. Some did not respond to our follow-up inquiries (n = 6); the reasons for other dropouts (n = 8) included family emergency, personal sickness, and lost or broken Fitbit.

Study Design

Both conditions

Go&Grow has a simple interface for physical activity-related gamification (based on the user’s completion of daily step goals, flowers can be grown and planted in their gardens). Participants could also follow workout tutorials implemented in Go&Grow from the NIH Go4life (National Institute on Aging at NIH, 2020), and log and track their workouts. Both groups could add flowers to plant to their gardens if they logged at least 14 workouts in a week. The app sent daily notifications to remind all participants to use the app and asked participants to rate their daily mood using a mood slider.

The treatment group (social exergame condition) used the full version of Go&Grow with the social contact features, which allowed users to view, like, reply to, and post a story and view other users’ gardens. They also could unlock new flowers to grow/plant each week if they posted stories. More details of Go&Grow can be found in Lin et al. (2020).

The control group (exergame nonsocial condition) used Go&Grow without social contact and story features; therefore, they always received the same type of flower (Supplementary Figure 1).

Procedure

The current study was a two-armed randomized control trial with a pretest (baseline Weeks 1 and 2), weekly assessments (Weeks 1–8), and a post-test (after Week 8). All recruitment materials specify: “Caregivers: Join a fitness study and increase your physical activity!” Participants were blinded to their conditions. Both groups received the same frequency of contact from researchers and followed the same procedures throughout the study. All contact and data collection were done remotely. Supplementary Figure 2 shows the flow diagram of the study procedure.

Session 1 (pretest)

Research assistants trained all participants to use Fitbit by phone and with an instructional manual, an approach that has been effective in previous studies (Robinson et al., 2019).

Week 1–2 (2-week baseline)

All participants wore the Fitbit for 2 weeks to obtain baseline steps that were averaged to create a baseline score. Both conditions also completed a weekly survey for Weeks 1 and 2.

Week 3–8 (6-week intervention period)

On the first day of Week 3, research assistants trained participants on how to use Go&Grow with the instructional guide and phone consultation for both conditions. After the training, all participants started using the app while continuing to wear the Fitbit and completing weekly surveys for 6 weeks. They were reminded daily with app-based notifications to wear their Fitbit and use the Go&Grow app. The notification for the treatment condition also reminded them to share stories to unlock new flowers.

After Week 8 (post-test)

All participants completed a post-test (the same as the pretest).

Measures

The current study consisted of measures from the pretest, weekly assessments (Weeks 1–8), and post-test (after Week 8). The same measures were used for both conditions. All questionnaires were administered through Qualtrics links sent to each participant via Eztexting or email. Participants’ electronic data (steps and app usage) were collected using a secure server from Brandeis University, and the data were kept confidential. A summary of all measures is shown in Supplementary Table 2.

Primary Outcome Measures

Well-being

Loneliness was measured from the pretest, post-test, and weekly surveys using the 20-item UCLA loneliness scale (Russell, 1996).

Stress was measured from the pretest and post-test based on the Perceived Stress Scale (Cohen et al., 1983). Weekly stress was assessed with the Brief Symptom Inventory (Derogatis & Melisaratos, 1983).

Affect was measured from the pretest, post-test, and weekly surveys using the Positive and Negative Affect Schedule (Watson et al., 1988).

The Caregivers’ Stress Scales (Pearlin et al., 1990) were assessed at pretest and post-test with the following subscales: overload, relational deprivation, job and caregiving conflict, role captivity, sense of self, caregiving competence, personal gain, management of situation, management of meaning, management of distress, and expressive support.

Physical activity

Physical activity was measured by steps and self-reported physical activity. Steps were automatically synced from the Fitbit for both conditions daily. Days with 500 or fewer recorded steps were coded as missing. Weekly step averages were calculated for weeks with 4 or more days not coded missing. The first 2 weeks of baseline steps were averaged to form the baseline steps.

Percent change of steps from baseline to post-test was calculated as the difference between the average weekly steps during the intervention and baseline steps divided by baseline steps.

Minutes spent engaging in light and moderate to vigorous activities were collected from Fitbits.

We also examined percentages of participants in each condition who reached the threshold for minimal clinically important difference (MCID) in steps, in which participants who increased in steps by 10% or more during the intervention were coded as 1, and those who did not reach the threshold for MCID were coded as 0 (Ramsey et al., 2021).

Self-reported physical activity was assessed at pretest and post-test using the International Physical Activity Questionnaire (Craig et al., 2003). This included metabolic equivalent of task (MET)-minutes per week calculated from each category: work, transportation, yard work, leisure, walking, moderate, vigorous, and total physical activity. Each MET-minutes per week score from each subscale was calculated by multiplying METS for each activity by the number of days and daily time. An aggregated MET-minutes per week score was also calculated by averaging across all subscales.

Social support

Overall social support was measured at the pretest and post-test using the 18-item version of the Lubben Social Network Scale (Lubben & Gironda, 2014). Weekly social support was measured with the Social Connectedness Scale by van Bel et al. (2009) with the following subscales: relationship salience, shared understandings, knowing other’s experiences, satisfaction with the contact quantity and quality, and dissatisfaction with contact quantity (van Bel et al., 2009).

Secondary outcome measures

Secondary outcomes at pretest and post-test were assessed because of their known associations with physical activity and well-being (Neupert et al., 2009; Robinson et al., 2019). Exercise self-efficacy was measured with a nine-item scale (Neupert et al., 2009). Life satisfaction was measured with a five-item scale (Diener et al., 1985). Sense of control was measured using the scale used in Midlife in the United States (Lachman & Weaver, 1998a, 1998b).

Covariates

Age (continuous), sex (1 = male, 2 = female), education (years), and functional health (36-item Short-Form Health Survey general health subscale) were obtained from the demographic survey and served as covariates because of their previously recognized relationships with the outcome variable (Radler & Ryff, 2010). App use served as a measure of participant engagement or dose of the intervention, measured by the average number of times each day participants opened the app during the intervention period. App use between the two conditions during the intervention did not significantly differ between the treatment group (M = 1.48, standard deviation [SD] = 1.98; range = 0–11) and the control group (M = 1.32, SD = 0.92; range = 0–5).

Data Analyses

First, we examined all baseline characteristics between both conditions (Supplementary Table 2) and three cohorts (Supplementary Table 1). To test the research questions, multilevel models (MLM) and the R package lme4 were used. Analyses were conducted for the full sample with the intent to treat everyone who completed the pretest (N = 76), as missing data in outcome variables could be handled through multilevel analysis. Changes in well-being, physical activity, and social support were examined with condition-by-time analyses examining whether the treatment and control conditions differed in the amount and direction of change controlling for the covariates in primary and secondary outcomes. Simple slope tests using Johnson–Neyman were conducted for all significant interactions. Planned comparisons were conducted for more specific comparisons based on a priori predictions. Furthermore, MLM was tested to examine whether social feature usage for the treatment group predicted changes in physical activity. Using the PROCESS Macro model in SPSS (Hayes, 2013), mediation models including covariates were tested for mechanisms of change (i.e., whether increased social support and physical activity, in turn, led to better well-being).

Results

Participants

The demographic information for each condition and the whole sample is shown in Supplementary Table 3. Randomization was checked by comparing two conditions and three cohorts with a series of independent t tests and χ2 tests. There were no significant differences between conditions (Supplementary Tables 1 and 3) or cohorts at baseline (Supplementary Table 1) for all primary and secondary measures.

Effects of the Intervention

Condition × time effects on primary outcomes

Well-being

As predicted, a significant condition-by-time interaction was found for the management of distress (β = 0.19, p = .01), in which the treatment group significantly increased in management of distress (β = 0.17, p = .01) compared with the control group (Figure 2A). The planned comparison showed the treatment group reported better well-being from the pretest to post-test in two dimensions: A decrease in loneliness (95% confidence interval [CI] −0.35 to −0.002, p < .05) and more caregiver’s management of meaning (95% CI 0.06–0.51, p = .01), whereas the control group did not change. There was no significant condition-by-time interaction predicting loneliness, stress, or positive and negative affect (Table 1 and Supplementary Table 4).

Figure 2.

Figure 2.

Open in a new tab

Relationship of condition and time predicting outcomes. (A) shows there was significant time and condition interaction on management of distress (β = 0.19, p = .01), in which the treatment group significantly increased in management of distress (β = 0.17, p = .01). (B) reveals there was a significant interaction of time and condition on satisfaction with contact quality (β = 0.06, p = .04), in which the treatment group significantly increased in satisfaction with contact quality compared with the baseline (β = 0.06, p = .01).

Table 1.

Pretest and Post-Test Measures by Condition

Variable Range Treatment group Control group Condition × time effect
B (p value)
Pretest (N = 38) Post-test (N = 31) Range Pretest (N = 38) Post-test (N = 35)
M (SD) M (SD) M (SD) M (SD)
Self-reported physical activity
Work MET 0–3,225 572.05 (617.37) 677.72 (931.94) 0–3,390 424.00 (680.48) 478.26 (900.93) 95.21 (.692)
Transport MET 0–1,386 269.08 (325.21) 282.99 (305.94) 0–1,188 290.42 (352.51) 224.06 (360.49) 62.73 (.532)
Yard work MET 0–3,690 531.67 (645.58) 645.58 (558.85) 0–2,415 403.28 (522.74) 491.73 (633.69) −23.35 (.902)
Leisure MET 0–2,655 376.70 (542.93) 657.33 (683.91) 0–2,693 306.44 (584.51) 400.09 (613.16) 174.30 (.313)
Walking MET 0–2,425 462.17 (524.37) 635.41 (565.07) 0–2,178 433.60 (538.93) 458.23 (499.86) 130.63 (.336)
Moderate MET 0–3,810 702.36 (760.14) 910.68 (678.67) 0–3,375 586.12 (762.96) 692.81 (941.80) 49.34 (.836)
Vigorous MET 0–2,880 359.16 (522.09) 520.77 (762.58) 0–2,880 214.74 (391.45) 224.46 (566.89) 197.67 (.174)
Total MET 0–6,140.5 1,523.69 (1,324.89) 2,066.86 (1482.13) 0–6,363 1,234.46 (1,382.82) 1,375.50 (1,640.83) 377.88 (.341)
Average MET 0–2,302.69 587.20 (493.79) 800.31 (568.19) 0–2,386.13 492.36 (550.84) 551.96 (653.31) 141.64 (.374)
Social support 5–46 32.32 (7.98) 32.23 (9.81) 8–47 30.64 (9.26) 28.40 (9.15) 1.30 (.481)
Loneliness 1–3.7 2.21 (0.63) 2.08 (0.75) 1–3.8 2.21 (0.73) 2.24 (0.76) −0.12 (.292)
Positive affect 0.15–4 2.32 (0.85) 2.14 (1.05) 0.31–3.92 2.13 (0.93) 2.02 (1.03) −0.11 (.509)
Negative affect 0–3.32 1.09 (0.80) 1.00 (0.92) 0–2.85 0.90 (0.67) 0.90 (0.74) −0.04 (.830)
Stress 7–39 23.08 (7.92) 22.32 (8.44) 6–40 22.82 (8.66) 22.20 (7.78) 0.85 (.510)
Exercise efficacy 9–34 19.74 (6.31) 22.87* (6.70) 9–36 19.10 (6.99) 22.86* (6.68) −0.20 (.893)
Life satisfaction 5–33 19.84 (7.50) 20.61 (7.43) 5–35 19.95 (7.95) 18.66 (8.73) 1.58 (.299)
Sense of control 14.5–41.5 31.96 (6.05) 31.10 (7.32) 8–42 29.77 (6.76) 29.07 (7.82) −0.10 (.522)
Caregiving stress scale
Overload 1–4 1.84 (0.92) 1.77 (0.80) 1–3.25 1.61 (0.68) 1.58 (0.73) 0.04 (0.750)
Relational deprivation 1–4 1.97 (0.97) 1.83 (0.86) 1–4 1.83 (0.85) 1.83 (0.77) −0.09 (0.594)
Job caregiving conflict 1–3.8 2.18 (0.72) 2.25 (0.73) 1–4 2.11 (0.60) 2.27 (0.75) −0.07 (0.508)
Role captivity 1–4 1.88 (0.90) 1.82 (1.02) 1–4 1.93 (0.83) 2.13 (1.04) −0.12 (0.251)
Sense of self 1–4 3.66 (0.51) 3.55 (0.68) 2–4 3.34 (0.58) 3.39 (0.57) −0.14 (0.339)
Caregiving competence 1–4 3.34 (0.51) 3.32 (0.64) 1.5–4 3.13 (0.75) 3.07 (0.60) 0.00 (0.986)
Management of situation 1–4 2.84 (0.60) 2.73 (0.72) 1–4 2.65 (0.63) 2.66 (0.47) −0.07 (0.612)
Expressive support 1.5–4 3.16 (0.50) 3.12 (0.52) 1.5–3.88 3.10 (0.41) 3.04 (0.61) −0.01 (0.956)
Management of distress 1.38–3.25 2.18 (0.37) 2.37 (0.31) 1.25–3.25 2.25 (0.35) 2.23 (0.39) 0.19* (0.010)
Management of meaning 1.89–4 3.20 (0.52) 3.25 (0.44) 1.67–3.89 2.94 (0.51) 2.97 (0.46) 0.01 (0.924)
Personal gain 1–4 3.32 (0.65) 3.37 (0.70) 1–4 2.93 (0.81) 2.98 (0.76) 0.05 (0.748)
Open in a new tab

Notes: MET = metabolic equivalent of task; SD = standard deviation, M = Mean, B = Coefficient.

Physical activity

As predicted, the main effect of time was significant from baseline to intervention period for percent change in steps (β = 0.06, p = .05) and light activity minutes (β = 37, p = .05; Supplementary Table 4). Caregivers who used the app more times a day also significantly increased in steps (β = 0.03, p = .003) and light activities minutes (β = 0.19, p = .009). Contrary to predictions, no significant main effects of condition or interaction effects of time and condition on other physical activity measures were found (Table 1). There was also no significant difference between the two conditions in the percentage of participants who reached the threshold for MCID (treatment: 38%; control: 30%).

For the Go4life entries, 50% of the treatment and 47.5% of the control condition logged workouts throughout the intervention. Total workouts logged were: treatment: M = 21, SD = 28, range = 0–109, and control: M = 18, SD = 33, range = 0–187).

Social support

As predicted, significant interaction effects were found predicting weekly social support (satisfaction with contact quality; β = 0.06, p = .04), in which the treatment group significantly increased in satisfaction with contact quality compared with baseline (β = 0.06, p = .01; Table 1 and Figure 2B). Time effects or interaction results for the other weekly subscales and the overall social support were not significant (Table 1).

Condition × time effects for secondary outcomes

Exercise efficacy increased for both control (β = 3.46, p < .01) and treatment groups (β = 3.27, p < .01). There were no significant interaction effects on any secondary outcome measures.

Indirect Effects

As predicted, a significant indirect effect was found for change from baseline to post-test for satisfaction with contact quality, predicting loneliness (indirect effect: −0.11, 95% CI −0.29 to −0.14) and positive affect (indirect effect: 0.19, 95% CI 0.0002–0.40). Compared to the control group, the treatment group increased in satisfaction with contact quality, which in turn led to decreased loneliness and increased positive affect (Table 1). No other indirect effects were found.

Social Support and Change in Physical Activity

Exploratory analyses examining social support as a moderator of condition and physical activity

Exploratory analyses were conducted to test condition by time by social support (overall social support) and each weekly social support subscale predicting the percent change of steps during the intervention.

A significant interaction of time, condition, and change of overall social support from pretest to post-test was found to predict weekly percent change of steps during the intervention (β = 0.001, p = .02). A simple slope test showed that for the treatment condition, but not for the control group, those with an increase in overall social support increased significantly in steps from baseline to the end of the intervention (β = 0.17, p = .01; Supplementary Table 6 and Figure 3A).

Figure 3.

Figure 3.

Open in a new tab

Relationship of condition, time, and social support variables predicting percent change of steps. (A) shows a significant interaction of time, condition and change of overall social support from pretest to post-test was examined in relation to weekly percent change of steps during the intervention (β = 0.001, p = .02). Simple slope test using the Johnson–Neyman approach showed that for the treatment condition, those with higher overall social support increased significantly in steps from baseline to the intervention (β = 0.17, p = .01). (B) reveals a significant interaction of time, condition, and knowing the others’ experiences predicting percent change of steps from baseline to the intervention (β = 0.02, p = .04), in which those with a high level of knowing others’ experiences increased significantly in steps from baseline to the intervention (β = 0.16, p = .02). SD = standard deviation.

For the weekly social support subscale, there was a significant interaction of time, condition, and knowing others’ experiences predicting percent change of steps from baseline to the intervention (β = 0.02, p = .04), in which those who increased in knowing others’ experiences increased significantly more in steps than those who decreased in knowing other’s experiences (β = 0.16, p = .02; Supplementary Table 6 and Figure 3B). Significant interaction results for other subscales were not found.

Within-group analyses: time × social engagement/social comparison for the treatment group

Only the treatment group had social contact features. MLM examined whether social features’ usage predicted changes in physical activity. Social engagement was measured by summing the total number of times caregivers posted, liked, replied, and viewed their own and others’ stories throughout the whole intervention (M = 39, SD = 71.09, range = 1–351). A social comparison was measured by summing the total number of times caregivers viewed others’ gardens (M = 72.86, SD = 129.47, range = 1–663).

Social engagement

As predicted, social engagement significantly predicted change in physical activity, measured by Fitbit for percent change in steps (β = 0.001, p = .001), and minutes spent engaging in light (β = 0.24, p = .002) and moderate to vigorous activity (β = 0.03, p = .04; Supplementary Table 7, Supplementary Figure 3, and Figure 4). Social engagement also predicted change in self-reported physical activity measures: METs for work (β = 17.04, p = .008), moderate (β = 5.56, p = .01), yard (β = 6.29, p < .001) total METs (β = 8.10, p = .02), and overall average METs (β = 3.54, p = .01; Supplementary Table 7 and Supplementary Figure4). Simple slopes test showed that those with more social engagement in the treatment group significantly increased in steps (β = 0.24, p < .001), METs for work (β = 519, p = .01), yard work (β = 524, p = .01), moderate (β = 552, p = .02), total METs (β = 1,054, p = .01), and overall average METs (β = 438, p < .001), whereas those with less social engagement in the treatment group engaged in significantly less light (β = −19.10, p = .03) and moderate to vigorous activity (β = −3.36, p = .03) and METs for yard work (β = −383, p = .05). The interactions were not significant for the other physical activity outcomes.

Figure 4.

Figure 4.

Open in a new tab

Significant indirect effects with all paths showing standardized coefficients. (A) depicts a significant indirect effect of satisfaction with contact quality mediating effects of condition on loneliness (indirect effect: −0.11, 95% CI −0.29 to −0.14). (B) shows a significant indirect effect of satisfaction with contact quality mediating effects of condition on positive affect (indirect effect: 0.19, 95% CI 0.0002–0.40). CI = confidence interval. *p < .10; **p < .05; ***p < .01.

Social comparison

As predicted, greater use of the social comparison features was related to a greater increase in physical activity from pretest to post-test for light activity (β = 11, p = .01), and METs for yard (β = 3.35, p = .001) and moderate work (β = 2.99, p = .01; Supplementary Table 8 and Supplementary Figure 5). Simple slopes test showed that those with more social comparisons in the treatment group significantly increased in METs for yard (β = 516, p = .01) and moderate work (β = 547, p = .02). The interactions were not significant for the other physical activity outcome measures.

Discussion

Compared to noncaregivers, family caregivers often experience a higher level of stress, social isolation, a sedentary lifestyle, and poor mental and physical health (Bressan et al., 2020; Cohen et al., 2015). These are all well-established risk factors for poor health and all-cause mortality (Cunningham et al., 2020; Kelly et al., 2016). There is a vital need for enjoyable, sustainable, and scalable resources to support social, physical, and emotional well-being among family caregivers.

Effect of the Intervention on Well-Being

The treatment group decreased in caregiver’s stress (i.e., increased in management of distress) more than the control group, which was as predicted and the replicated pilot study’s results (Lin et al., 2020). Planned comparisons showed that the treatment group reported an increase in well-being (decrease in loneliness over time and greater management of stress), whereas the control group did not change. Condition-by-time interaction effects were not found for other well-being measures, which could be explained by the unpredictability of the caregiver’s daily life, in which caregivers often do not have control over the conditions of their loved ones, which may affect their own well-being (Bressan et al., 2020; Cohen et al., 2015). Moreover, many caregivers’ stress subscales involved instrumental or emotional support from their own family members, which was not directly related to the intervention. Nevertheless, results suggested that the social exergame app was beneficial for caregivers to better manage their stress and participating in an intervention with other caregivers may have helped with reducing loneliness. Given the limited time that family caregivers have for leisure activities, a physical activity intervention must be sensitive to their time and location constraints (Cohen et al., 2015; Hopwood et al., 2018; Stowell et al., 2019). Go&Grow was developed as a smartphone app that addresses such constraints, and current results showed success in reducing caregiver stress for the treatment group. This suggests that gamified technology-based physical activity interventions, accessible at any time and in any setting, can be engaging, sustainable, and scalable for the time-constrained family caregivers.

Social Support and Change in Well-Being

Social support (satisfaction with contact quality) mediated the relationship between condition and well-being. Other social support subscales did not have significant indirect effects. These results suggested that satisfaction with the quality rather than the quantity of one’s online social network was important for well-being. Moreover, satisfaction with one’s quality of social network played an essential role in well-being. In other words, having caregivers engage in more in-depth conversations and increasing satisfaction with the conversation are keys to promoting their well-being, whereas exchanging information may not be enough.

Physical Activity and Well-Being

Both conditions significantly increased in physical activity (steps and minutes spent engaging in light activity), and exercise efficacy (secondary outcome), and physical activity did not mediate the relationship between the condition and any well-being measures. This may be because both groups had a Fitbit and the Go&Grow app. Both groups were rewarded with flowers if they reached their goals, motivating them to exercise. Moreover, both groups’ initial motivation to participate in a fitness study and the study incentives (Fitbit and $25) could also serve as factors to promote step increases. There was also a lot of variability within groups in the physical activity measures (e.g., steps measured by Fitbit and calculated METs from self-report).

Social Support and Physical Activity

Exploratory analyses examining the effects of social support on physical activity (percent change in steps) as a function of time and condition showed those in the treatment group with increased overall social support and knowing others’ experiences increased steps more than those with lower support, whereas the change in steps for the control group was not related to a support level. Results also showed that more social engagement and social comparison predicted an increase in physical activity (steps, minutes active, and METs) for the treatment group. These findings suggest the importance of including social features in apps to motivate physical activity. The results of this study offered new insights into implementing behavior change mechanisms using exergames for sedentary and socially isolated family caregivers to improve their health with modifiable behavior factors such as physical activity and social support. Family caregivers successfully managing their own health and well-being needs can translate directly into more effective care for their care recipients (Koumakis et al., 2019; Waligora et al., 2019).

Limitations and Future Study

Despite the current study’s strengths and innovative approaches, one limitation was that the app was only operational on Android smartphones, preventing outreach to a wider population (e.g., iPhone users). Additionally, we did not have information about participants’ weekly goals. Participants could decide to stick with the same or lower their step goal rather than increase their steps each week to make it easier to grow flowers. Although if they increased their goal, they could choose a more desirable flower. Results of the self-report measures indicated that 60% of all caregivers in the study did not meet the recommended 150 min of moderate to vigorous physical activity each week by the end of the intervention, although this was not specified to participants as a goal. Although self-reported physical activity data may be subject to recall bias (Prince et al., 2020), and our data showed significant correlations between the self-reported and objective activity data. App use data also did not indicate the amount of time participants spent on the app each day. Additionally, the control group also got rewarded when achieving goals, and both conditions received the Fitbit and $25 as compensation, which may explain why both conditions showed similar increases in steps. Future studies could add a passive control group that receives a different app without the exergame features. Future work can also consider adding qualitative questions to better understand caregivers’ experiences. The current study did not have a follow-up period to examine the maintenance of caregivers’ social support, physical activity, and well-being after the intervention ended, which is important to consider in future work. Nevertheless, social exergames can offer promising directions for future research and practice because they are a convenient and enjoyable approach for caregivers. Future work can consider working with clinicians and caregiver organizations to introduce the social exergame and its benefits to caregivers, and encourage caregivers to engage in more self-care activities to promote their social support, physical activity, and well-being.

Conclusion

The current study addressed some of the limitations in past exergame studies (e.g., single group design and small sample size) and included a control condition without social features to better examine the effects of a social exergame. The treatment group improved in well-being (i.e., caregiver’s management of distress) and social support (i.e., satisfaction with contact quality) compared to the control group. Both conditions increased in steps and exercise self-efficacy. Social support in technology interventions is a promising direction to promote caregivers’ well-being and physical activity. The results of the current study could be used to offer new insights into potential mechanisms of behavior change for more engaging, sustainable, portable, and scalable interventions in the future.

Supplementary Material

gnad028_suppl_Supplementary_Material
Click here for additional data file. (3.1MB, docx)

Acknowledgments

We would like to thank our participants and funding resources. We would also like to thank Dr. Andrea Parker and the Wellness Technology Lab at Northeastern University for their collaboration in developing the initial Go&Grow app supported by a supplement grant to P30 AG048785, and Dr. Herman Saksono for the continuous help with the app (codes and server-related matters) and various app logistics. The current study is a part of the first author’s dissertation and a part of the study was previously presented at the Gerontological Society of America annual meeting. The first author acknowledges support from a National Institute on Aging (NIA)-funded T32 Training Program in Behavioral Geriatrics [NIA T32 AG049666].

Contributor Information

Xin Yao Lin, Department of Psychology, Brandeis University, Waltham, Massachusetts, USA; Division of Geriatrics and Palliative Medicine, Weill Cornell Medicine, New York, New York, USA.

Lin Zhang, Department of Psychology, Brandeis University, Waltham, Massachusetts, USA.

Saiyeon Yoon, Department of Psychology, Brandeis University, Waltham, Massachusetts, USA.

Ruoying Zhang, Department of Psychology, Brandeis University, Waltham, Massachusetts, USA.

Margie E Lachman, Department of Psychology, Brandeis University, Waltham, Massachusetts, USA.

Funding

The current study was funded by the Boston Roybal Center, P30 AG048785, from the National Institute on Aging, National Institutes of Health (NIA/NIH), American Psychological Association Dissertation Award, Sigma Xi Research Award, and the Brandeis University Dissertation Provost Award. The funding sources had no other involvement other than financial support.

Conflict of Interest

None declared.

Data Availability

The clinical trial was registered at the U.S. National Institutes of Health (ClinicalTrials.gov) #NCT05032872. Deidentified data will be deposited at the Boston Roybal Center and ­investigators interested in the data can contact the corresponding author X Lin at xylin@brandeis.edu or xyl4001@med.cornell.edu or the Boston Roybal Center at roybal@brandeis.edu.

References

  1. Bressan, V., Visintini, C., & Palese, A. (2020). What do family caregivers of people with dementia need? A mixed- method systematic review. Health & Social Care in the Community, 28(6), 1942–1960. doi: 10.1111/hsc.13048 [DOI] [PubMed] [Google Scholar]
  2. Brox, E., Evertsen, G., Åsheim-Olsen, H., Burkow, T., & Vognild, L. (2014). Experiences from long-term exergaming with elderly. Proceedings of the 18th International Academic MindTrek Conference on Media Business, Management, Content & Services - AcademicMindTrek’, 14, 216–220. doi: 10.1145/2676467.2676483 [DOI] [Google Scholar]
  3. Chao, Y. -Y., Scherer, Y. K., & Montgomery, C. A. (2015). Effects of using Nintendo WiiTM exergames in older adults: A review of the literature. Journal of Aging and Health, 27(3), 379–402. doi: 10.1177/0898264314551171 [DOI] [PubMed] [Google Scholar]
  4. Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385–396. doi: 10.2307/2136404 [DOI] [PubMed] [Google Scholar]
  5. Cohen, S. A., Cook, S., Kelley, L., Sando, T., & Bell, A. E. (2015). Psychosocial factors of caregiver burden in child caregivers: Results from the new national study of caregiving. Health and Quality of Life Outcomes, 13, 120. doi: 10.1186/s12955-015-0317-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Craig, C. L., Marshall, A. L., Sjöström, M., Bauman, A. E., Booth, M. L., Ainsworth, B. E., Pratt, M., Ekelund, U., Yngve, A., ­Sallis, J. F., & Oja, P. (2003). International physical activity questionnaire: 12-Country reliability and validity. Medicine and Science in Sports and Exercise, 35(8), 1381–1395. doi: 10.1249/01.MSS.0000078924.61453.FB [DOI] [PubMed] [Google Scholar]
  7. Cunningham, C., O’Sullivan, R., Caserotti, P., & Tully, M. A. (2020). Consequences of physical inactivity in older adults: A systematic review of reviews and meta-analyses. Scandinavian Journal of Medicine & Science in Sports, 30(5), 816–827. doi: 10.1111/sms.13616 [DOI] [PubMed] [Google Scholar]
  8. Derogatis, L. R., & Melisaratos, N. (1983). The brief symptom inventory: An introductory report. Psychological Medicine, 13(3), 595–605. doi: 10.1017/S0033291700048017 [DOI] [PubMed] [Google Scholar]
  9. Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49(1), 71–75. doi: 10.1207/s15327752jpa4901_13 [DOI] [PubMed] [Google Scholar]
  10. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press. [Google Scholar]
  11. Hopwood, J., Walker, N., McDonagh, L., Rait, G., Walters, K., Iliffe, S., Ross, J., & Davies, N. (2018). Internet-based interventions aimed at supporting family caregivers of people with dementia: Systematic review. Journal of Medical Internet Research, 20(6), e216. doi: 10.2196/jmir.9548 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Kelly, S., Martin, S., Kuhn, I., Cowan, A., Brayne, C., & Lafortune, L. (2016). Barriers and facilitators to the uptake and maintenance of healthy behaviours by people at mid-life: A rapid systematic review. PLoS One, 11(1), e0145074. doi: 10.1371/journal.pone.0145074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Koumakis, L., Chatzaki, C., Kazantzaki, E., Maniadi, E., & Tsiknakis, M. (2019). Dementia care frameworks and assistive technologies for their implementation: A review. IEEE Reviews in Biomedical Engineering, 12, 4–18. doi: 10.1109/RBME.2019.2892614 [DOI] [PubMed] [Google Scholar]
  14. Lachman, M. E., & Agrigoroaei, S. (2010). Promoting functional health in midlife and old age: Long-term protective effects of control beliefs, social support, and physical exercise. PLoS One, 5(10), e13297. doi: 10.1371/journal.pone.0013297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Lachman, M. E., & Weaver, S. L. (1998a). Sociodemographic variations in the sense of control by domain: Findings from the MacArthur studies of midlife. Psychology and Aging, 13(4), 553–562. doi: 10.1037//0882-7974.13.4.553 [DOI] [PubMed] [Google Scholar]
  16. Lachman, M. E., & Weaver, S. L. (1998b). The Midlife Development Inventory (MIDI) personality scales: Scale construction and scoring. Waltham, MA: Brandeis University, Vol. 7, (pp. 1–9). [Google Scholar]
  17. Li, J., Erdt, M., Chen, L., Cao, Y., Lee, S. Q., & Theng, Y. L. (2018). The social effects of exergames on older adults: Systematic review and metric analysis. Journal of Medical Internet Research, 20(6), e10486. doi: 10.2196/10486 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lin, X. Y., Saksono, H., Stowell, E., Lachman, M. E., Sceppa, C., & Parker, A. G. (2020). Go&Grow: An evaluation of a pervasive social exergame for caregivers of loved ones with dementia. Proceedings of the ACM on Human-Computer Interaction, 4(CSCW2), 1–28. doi: 10.1145/3415222 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lubben, J., & Gironda, M. (2014). Measuring social networks and assessing their benefits. Ashgate Publishing Company. [Google Scholar]
  20. National Alliance for Caregiving. (2020, May 11). Caregiving in the US 2020. The National Alliance for Caregiving. Retrieved August 1, 2022, from https://www.caregiving.org/caregiving-in-the-us-2020/ [Google Scholar]
  21. National Institute on Aging at NIH. (2020). Try these exercises. Go4Life. https://www.nia.nih.gov/research/blog/2014/07/go4life-nia-health-education-campaign
  22. Neupert, S. D., Lachman, M. E., & Whitbourne, S. B. (2009). Exercise self-efficacy and control beliefs predict exercise behavior after an exercise intervention for older adults. Journal of Aging and Physical Activity, 17(1), 1–16. doi: 10.1123/japa.17.1.1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Newman, K., Wang, A. H., Wang, A. Z. Y., & Hanna, D. (2019). The role of internet-based digital tools in reducing social isolation and addressing support needs among informal caregivers: A scoping review. BMC Public Health, 19(1), 1–12. doi: 10.1186/s12889-019-7837-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Nielsen, L., Riddle, M., King, J. W., Aklin, W. M., Chen, W., Clark, D., Collier, E., Czajkowski, S., Esposito, L., Ferrer, R., Green, P., Hunter, C., Kehl, K., King, R., Onken, L., Simmons, J. M., Stoeckel, L., Stoney, C., Tully, L., Weber, W., & NIH Science of Behavior Change Implementation Team (2018). The NIH Science of Behavior Change Program: Transforming the science through a focus on mechanisms of change. Behaviour Research and Therapy, 101, 3–11. doi: 10.1016/j.brat.2017.07.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Pearlin, L. I., Mullan, J. T., Semple, S. J., & Skaff, M. M. (1990). Caregiving and the stress process: An overview of concepts and their measures. Gerontologist, 30(5), 583–594. doi: 10.1093/geront/30.5.583 [DOI] [PubMed] [Google Scholar]
  26. Pfeiffer, E. (1975). A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. Journal of the American Geriatrics Society, 23(10), 433–441. doi: 10.1111/j.1532-5415.1975.tb00927.x [DOI] [PubMed] [Google Scholar]
  27. Prince, S. A., Cardilli, L., Reed, J. L., Saunders, T. J., Kite, C., Douillette, K., & Buckley, J. P. (2020). A comparison of self-reported and device measured sedentary behaviour in adults: A systematic review and meta-analysis. International Journal of Behavioral Nutrition and Physical Activity, 17(1), 1–17. doi: 10.1186/s12966-020-00938-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Radler, B. T., & Ryff, C. D. (2010). Who participates? Accounting for longitudinal retention in the MIDUS national study of health and well-being. Journal of Aging and Health, 22(3), 307–331. doi: 10.1177/0898264309358617 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Ramsey, K. A., Meskers, C. G. M., & Maier, A. B. (2021). Every step counts: Synthesising reviews associating objectively measured physical activity and sedentary behaviour with clinical outcomes in community-dwelling older adults. Lancet Healthy Longevity, 2(11), e764–e772. doi: 10.1016/S2666-7568(21)00203-8 [DOI] [PubMed] [Google Scholar]
  30. Robinson, S. A., Bisson, A. N., Hughes, M. L., Ebert, J., & Lachman, M. E. (2019). Time for change: Using implementation intentions to promote physical activity in a randomised pilot trial. Psychology & Health, 34(2), 232–254. doi: 10.1080/08870446.2018.1539487 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Russell, D. W. (1996). UCLA Loneliness Scale (Version 3): Reliability, validity, and factor structure. Journal of Personality Assessment, 66(1), 20–40. doi: 10.1207/s15327752jpa6601_2 [DOI] [PubMed] [Google Scholar]
  32. Stowell, E., Zhang, Y., Castaneda-Sceppa, C., Lachman, M., & Parker, A. G. (2019). Caring for Alzheimer’s disease caregivers. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), 1–27. doi: 10.1145/3359232 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Unbehaun, D., Vaziri, D. D., Aal, K., Wieching, R., Tolmie, P., & Wulf, V. (2018). Exploring the potential of exergames to affect the social and daily life of people with dementia and their caregivers. Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, April 2018, 1–15. doi: 10.1145/3173574.3173636 [DOI]
  34. Van Bel, D. T., Smolders, K. C., IJsselsteijn, W. A., & De Kort, Y. A. W. (2009). Social connectedness: concept and measurement. In Intelligent Environments 2009 (pp. 67–74). IOS Press [Google Scholar]
  35. Waligora, K. J., Bahouth, M. N., & Han, H. R. (2019). The self-care needs and behaviors of dementia informal caregivers: A systematic review. Gerontologist, 59(5), e565–e583. doi: 10.1093/geront/gny076 [DOI] [PubMed] [Google Scholar]
  36. Watson, D., Anna, L., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales, Vol. 8. [DOI] [PubMed] [Google Scholar]
  37. Zarit, S. H., Kim, K., Femia, E. E., Almeida, D. M., & Klein, L. C. (2014). The effects of adult day services on family caregivers’ daily stress, affect, and health: Outcomes from the daily stress and health (DaSH) study. Gerontologist, 54(4), 570–579. doi: 10.1093/geront/gnt045 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gnad028_suppl_Supplementary_Material
Click here for additional data file. (3.1MB, docx)

Data Availability Statement

The clinical trial was registered at the U.S. National Institutes of Health (ClinicalTrials.gov) #NCT05032872. Deidentified data will be deposited at the Boston Roybal Center and ­investigators interested in the data can contact the corresponding author X Lin at xylin@brandeis.edu or xyl4001@med.cornell.edu or the Boston Roybal Center at roybal@brandeis.edu.

Articles from The Gerontologist are provided here courtesy of Oxford University Press

RESOURCES