Gamification-Augmented hoMe-based Exercise for Peripheral Artery Disease: Rationale and Design of the GAMEPAD Study

Abstract

Background:

Supervised treadmill exercise improves walking performance, functional capacity, and quality of life in patients with peripheral artery disease (PAD). However, few patients with PAD are enrolled in supervised exercise programs, and there are a number of logistical and financial barriers to their participation. A home-based walking intervention is likely to be more accessible to patients with PAD, but no fully home-based walking program has demonstrated efficacy. Concepts from behavioral economics have been used to design scalable interventions that increase daily physical activity in patients with atherosclerotic vascular disease, but whether a similar program would be effective in patients with PAD is uncertain.

Study Design and Objectives:

GAMEPAD (NCT04536012) is a pragmatic, virtual, randomized controlled trial designed to evaluate the effectiveness of a gamification strategy informed by concepts from behavioral economics to increase daily physical activity in patients with PAD who are seen in cardiology and vascular surgery clinics affiliated with the University of Pennsylvania Health System. Patients are contacted by email or text message, and complete enrollment and informed consent on the Penn Way to Health online platform. A GAMEPAD substudy will evaluate the effectiveness of opt-in versus opt-out framing when approaching patients for study participation. Patients are then provided with a wearable fitness tracker, establish a baseline daily step count, set a goal to increase daily step count by 33–50%, and are randomized 1:1 to the gamification or control arms. Interventions continue for 16 weeks, including a 4-week period during which goal step count is gradually increased in the gamification arm, with follow-up for an additional 8 weeks to evaluate the durability of behavior change. The trial has met its enrollment goal of 102 participants, with a primary endpoint of change from baseline in daily steps over the 16-week intervention period. Key secondary endpoints include change from baseline in daily steps over the 8-week post-intervention follow-up period and changes in patient-reported measures of PAD symptoms and quality of life over the intervention and follow-up periods.

Conclusions:

GAMEPAD is a virtual, pragmatic randomized clinical trial of a novel, fully home-based walking intervention informed by concepts from behavioral economics to increase physical activity and PAD-specific quality of life in patients with PAD. Its results will have important implications for the application of behavioral economic concepts to scalable home-based strategies to promote physical activity in patients with PAD and other disease processes where physical activity is limited by exertional symptoms.

Clinical trial registration:

clinicaltrials.gov; NCT04536012

INTRODUCTION

Peripheral artery disease (PAD), defined as atherosclerotic narrowing or occlusion of at least one peripheral artery, affects > 200 million people worldwide and is associated with substantial morbidity and healthcare costs.(1–3) Symptomatic PAD most often involves the lower extremity, and often presents as intermittent claudication, which is characterized by leg soreness or heaviness with ambulation, and often causes substantial functional limitation.(4) Though a minority of patients with peripheral artery disease have classic intermittent claudication, a greater proportion have lower extremity symptoms, and even those who do not report overt symptoms may have substantial unrecognized functional limitation.(4,5)

In patients with intermittent claudication and exertional limb symptoms, randomized controlled trials have demonstrated that supervised treadmill exercise therapy improves walking performance, functional capacity, and quality of life.(6–8) Based on these trials, supervised exercise therapy is a class IA recommendation in the American Heart Association/American College of Cardiology and Society for Vascular Surgery clinical practice guidelines.(9,10) However, despite the strong evidence in its favor and recent decision by the Centers for Medicare and Medicaid Services to cover supervised exercise for patients with PAD,(11) few patients with PAD attend supervised exercise programs due to logistical and financial barriers.(12,13) A home-based walking intervention is likely to be more accessible to patients with PAD, and some studies evaluating hybrid home- and center-based walking have shown promise(8,14–18); however, no fully home-based walking program has proven effective.

In designing a fully home-based walking program for patients with PAD, it is important to consider barriers that limit exercise in patients with PAD. Claudication or other lower extremity symptoms with exertion represent one barrier to increasing physical activity in these patients, but struggling with behavior change is a universal human tendency. Behavioral economics is a scientific field of inquiry that uses principles from economics and psychology to understand and influence how individuals make decisions. In classical experiments, behavioral economists showed that people commonly make certain decision errors, leading to the concept of “bounded rationality.”(19) For example, individuals are more motivated by immediate rather than delayed gratification, by losses rather than gains, and by the desire to avoid feelings of regret.(20,21) In prior studies, gamification based on behavioral economic concepts increased physical activity more than control in patients with or at risk for ASCVD.(22,23) However, whether such a gamification intervention improves physical activity in patients with PAD, who have both psychological and physiological barriers to increasing exercise, is uncertain.

We therefore designed the Gamification-Augmented hoMe-based Exercise for Peripheral Artery Disease (GAMEPAD) randomized controlled trial to test the effectiveness of gamification plus automated coaching, compared with control, to increase physical activity over a 16-week intervention with 8-week follow-up in patients with PAD (Figure 1).

Figure 1:

Summary of the design of the GAME PAD trial

Patients with peripheral artery disease (PAD) are eligible for enrollment into GAME PAD. Patients complete informed consent and baseline questionnaires on the Penn Way to Health platform and are mailed a Fitbit wearable fitness tracker. They wear the fitness tracker for 2 weeks to establish a baseline step count, and patients with baseline step count > 7500 are excluded. The remaining patients set a goal to increase step count by 33–50% above baseline and are then randomized to gamification + automated coaching or attention control.

METHODS

GAMEPAD is funded by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR001878, and supported in part by the Institute for Translational Medicine and Therapeutics of the Perelman School of Medicine at the University of Pennsylvania. Prior to enrollment of the first patient, the trial was registered at clinicaltrials.gov (NCT03911141) and the study protocol was approved by the University of Pennsylvania Institutional Review Board. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents.

Study Population

GAMEPAD is recruiting participants with PAD from the more than 60 primary care, cardiology, and vascular surgery practices affiliated with the University of Pennsylvania Health System, located in southeastern Pennsylvania and New Jersey. To increase generalizability, inclusion criteria are broad with few exclusion criteria: Participants are eligible if they are ≥ 18 years old, own a smartphone or tablet operating the iOS or Android operating system, and have PAD, defined as ankle-brachial index < 0.90, lower extremity CT scan or ultrasound consistent with PAD, angiography with ≥ 70% stenosis in any lower extremity artery, or a history of medical or surgical revascularization. They are excluded if they are unable or unwilling to provide informed consent, have chronic limb-threatening ischemia (defined as rest pain, ulceration, or tissue loss involving the lower extremity), have planned lower extremity revascularization, have prior above or below knee amputation, require a wheelchair or use of a walking aid other than a cane, are currently participating in a supervised exercise program for patients with PAD, have anticipated life expectancy less than 6 months, if they take > 7500 steps/day at baseline, or if there is any other reason why it is not feasible to complete the entire 6-month study.

Potentially eligible patients are identified using data from the electronic health record (EHR) and the health system’s clinical data warehouse, or by direct referral from their cardiologists and vascular surgeons.

Study Procedures

After potentially eligible patients are identified via the EHR, they are directly contacted by email or text message. In a substudy of the GAME PAD trial, patients will be randomized to one of two different messages, framing participation in an opt-in or opt-out manner. Both messages introduce the study and provides a link to the study’s webpage on Penn Way to Health,(24) a research and care delivery platform that automates the delivery of behavior change interventions. Patients randomized to the opt-in approach will be instructed to call a study coordinator or visit the Way to Health platform to enroll in the study; patients randomized to the opt-out study will be notified that they will be called by a study coordinator unless they decline to participate. Patients referred directly from cardiologists and vascular surgeons are not included in the opt-in versus opt-out substudy, and will be called by a study coordinator to discuss the study and enrollment.

All participants enroll via a study webpage on the Way to Health platform. On the study webpage, participants create an account, confirm their eligibility, complete the informed consent process, and answer questions regarding demographics, medical history, health status, and health-related quality of life. Participants complete the enrollment process on their own, outside of clinical encounters, with study staff available to answer any questions by phone.

Once eligible patients complete the baseline questionnaire, they are mailed a wrist-worn activity tracker (Fitbit Charge 5) and instructions for how to connect the device to the Way to Health platform. Participants are told to wear the device for two weeks to get comfortable using it, but are not explicitly told that baseline physical activity data is collected during this time period. During the second week of this two-week run-in phase, baseline activity measures (daily step counts, minutes of moderate physical activity) are estimated, as done previously.(22,25) The first week of data is ignored to diminish the potential upward bias of the estimate from higher activity during initial device use. To prevent potential mismeasurement, days with step values < 1000 are ignored, as previous studies have shown that these are unlikely to represent capture of actual activity during the entire day.(26,27) For participants with < 4 days with step count ≥ 1000 during the second week of the run-in period, the baseline period is extended until at least 4 valid days of data are collected. Baseline daily step count is calculated as the mean daily step count from the second week of the run-in period; days where the participant had a step count < 1000 are excluded from both the numerator (steps) and the denominator (days). Participants who do not complete the run-in phase or had baseline step counts > 7500 steps/day are not randomized into the trial and are asked to return their Fitbit device.

Once baseline measures are established, participants are contacted via text message or email, given their baseline step count, and asked to set a goal step increase of 33%, 40%, or 50%, or a custom goal at least 1500 steps greater than baseline. This approach was selected to give participants the option of setting their own goal, which we have shown is the most effective approach for goal-setting,(23) while nudging them to choose between different goals that were ambitious but achievable. They are then randomized 1:1 to either gamification + automated coaching or attention control, stratified by baseline step count (< 2500 steps, 2500–5000 steps, > 5000 steps) using an electronic number generator through the Way to Health platform. Treatment assignment is necessarily open-label, but patients are not explicitly informed about the existence or details of other treatment arms.

To achieve their goal, all patients receive a message recommending that they begin walking for 15 minutes daily at their own pain-free pace or with minimal pain, gradually increasing the duration to 30–60 minutes daily. They are told to stop and rest if the pain becomes more significant, then resume walking. Though this is not the regimen typically prescribed for supervised exercise therapy—in which patients are usually instructed to walk at a pace that will induce claudication after 5–10 minutes of walking, then rest until claudication resolves, then repeat as necessary to complete a 60-minute session(11)—a simpler strategy was selected to facilitate adherence to a walking plan without face-to-face support from an exercise physiologist.

During the study period, patients are asked to complete questionnaires at 16 and 18 weeks’ follow-up. Participants receive $25 for enrolling in the study and completing baseline surveys, $25 for completing the 16-week intervention and 16-week survey with step data available for ≥ 60% of days, and $25 for completing 24-week survey. Patients are also allowed to keep the Fitbit at the conclusion of the study.

Study Treatments

In the gamification + automated coaching arm, participants will receive biweekly text messages including encouragement to walk for exercise, reminders about their goal step increase, reminders about optimal practices for home-based walking in patients with PAD, and information about the benefit of exercise in improving PAD symptoms and functional capacity. They will have a four-week ramp-up toward their step goal. The net difference between baseline and their goal will be divided by 4 and the participant will be asked to achieve the 25% increase each week for the 4-week ramp-up. For example, a participant with a baseline of 5000 steps and goal of 7400 steps will be asked to achieve goals of 5600 steps in week 1, 6200 in week 2, and so on through the 4-week period. Following the ramp-up period, the participant will be asked to maintain their goal during the remainder of the study.

Participants are also entered into a game that leverages insights from behavioral economics to address predictable barriers to behavior change (Table 1). The components are the following: 1) Pre-commitment: Each participant signs a contract agreeing to try their best to achieve their daily step goal, an approach shown to motivate behavior change.(28) 2) Points: At the start of each week, the participant receives 70 points (10 for each day of the week). Participants are endowed with points rather than given points after achieving a milestone to leverage loss aversion, a concept from prospect theory that indicates that individuals are more motivated by losses than gains.(20) Each day the participant is informed of their step count from the day prior. If the step goal was achieved, the participant retains his or her points; if the step goal was not achieved, they are informed that they had lost 10 points. Points are replenished at the start of each week to leverage the “fresh start effect” – the concept that individuals are more motivated for aspirational behavior around temporal landmarks like the start of a new week.(29) 3) Levels: At the end of the week, participants with 40 points or more advance one level; participants with less than 40 points drop down a level. The levels are blue (lowest), bronze, silver gold, platinum (highest). Each participant begins in the silver level to create a sense of achievable goals and use loss aversion to motivate ongoing efforts not to lose status.(30) Every eight weeks, individuals in the blue and bronze levels are restarted back at silver, and offered a chance to adjust their step goal, as long as they remain within the range of a 33 to 50% increase from baseline. This allows for another “fresh start,” creates a new endowment, and avoids participants becoming discouraged if they set their goals too high at the start of the study. 4) Supportive sponsor: Each participant selects a family member or friend of their choice who receives a weekly email with the participant’s progress, including accumulated points, level in the game, and average step count. The supportive sponsor helps enhance social incentives to motivate the participant toward his or her goal. Prior to starting the study, supportive sponsors participate in a three-way phone call with the participant and study staff to discuss the rules of the game and ways that they can help the participant reach their goals. Every eight weeks, participants in the blue and bronze levels are contacted, along with their sponsor, to discuss ways the sponsor can better help participants achieve their goals moving forward. The game lasts for the duration of the 16-week intervention. After 16 weeks, the participants in the gamification arm receive a daily text message noting their step count from the day prior for an additional 8-week follow-up period.

Table 1:

Behavioral economic principles used to inform the design of the interventions

Principle	Predictably irrational tendency	Implications for intervention design
Status quo bias	People favor the path of least resistance and avoid initiating change	Without the intervention, patients are unlikely to change their physical activity; therefore, the gamification intervention runs automatically rather than requiring the individual to actively participate, other than by working to achieve their physical activity goals
Immediacy	Immediate rewards are more motivating than rewards far into the future (e.g. health outcomes)	Points are rewarded every day to create an immediate ‘benefit’ that links to future benefits
Loss aversion	People are more motivated when the same situation is framed as a loss rather than a gain	Points are endowed at the beginning of each week and can be lost for not achieving step counts
Social ranking	Social influences from networks impact people’s behavior	Participants select a social support partner who will identify ways to help them in their journey and receive a weekly update on their progress.
Goal gradients	People try harder when goals are within reach	Participants start in the middle level and if they perform poorly, they get a fresh start every 8 weeks, and are moved back to the middle level

In the attention control arm, patients receive a text message each day telling them whether they achieved their step goal on the prior day for the entire 24-week study period.

Study Endpoints

The primary endpoint is change in mean daily steps from baseline through the 16-week intervention, excluding the 4-week ramp-up phase. Other outcomes include change in mean daily steps from baseline through the 24-week study period, change in light, moderate, or vigorous physical activity from baseline through the 16-week intervention and 24-week follow-up periods, change in Walking Improvement Questionnaire scores from baseline through 16-week and 24-week follow-up, and change in PROMIS mobility, pain interference, and satisfaction with social roles and activities scores from baseline through 16-week and 24-week follow-up. These outcomes were selected to reflect functional mobility in the community, an outcome of importance to patients with PAD.(31)

Mean daily steps are captured by the Fitbit devices and automatically uploaded to the Penn Way to Health platform. Step data are transferred to the Way to Health platform every 4 hours. To identify minutes of moderate and vigorous physical activity (MVPA), we will use data established in previous studies validating a threshold of 100 steps per minute as the minimum level of activity to be considered MVPA.(32,33)

STATISTICAL CONSIDERATIONS

The primary analysis will fit generalized linear mixed effect regression models to evaluate changes in physical activity and quality of life outcomes adjusting for each participant’s baseline measure, time, and calendar-month fixed effects (fitted as a nominal variable), and participant random effects to account for repeated measures. For analyses of daily steps, data captured during the entire study period will be used; this approach increases power by using all participant data, and provides a more complete picture of daily step count over the entire study period. All analyses will be performed according to the intention-to-treat principle.

Patients with PAD walk approximately 3900 steps/day (SD 2689 steps).(17) With a home-based exercise program, the mean change in steps/day over 4.5 months for patients with PAD was a decrease by 200 steps (SD ~2000 steps).(17) With 100 patients (50 in each arm), the study will have approximately 80% power to detect an 1100 step difference in change in daily step count between the two arms with alpha set at 0.05.

As in prior studies with similar design,(23,34) our approach to handling days with missing step count data will be to impute daily step counts for these days in our primary analysis, using an imputation model that includes study arm, calendar month (fitted as a nominal variable), week of study, baseline daily steps, age, sex, race/ethnicity, educational level, marital status, household income level, self-reported health, and participant random effect. We will also perform a sensitivity analysis that uses collected data without imputation (such that days with missing step counts are recorded as zero steps). In prior studies using similar statistical methods, results of such sensitivity analyses have been similar to the primary analyses.(23,34)

For the opt-in versus opt-out recruitment substudy, the primary outcome will be the proportion of patients who enroll in the study as a fraction of all patients contacted, compared between potential participants randomized to opt-in versus opt-out framing. We will also test for an interaction between method of recruitment (opt-in versus opt-out) and the effect of the intervention on the primary outcome of change from baseline in mean daily steps.

TRIAL STATUS

The trial started enrollment in October 2020 and completed enrollment in July 2023. Ultimately, 102 patients were enrolled, as patients in the baseline phase when the trial reached its recruitment goal were allowed to continue in the study. The last patient enrolled will finish follow up in January 2024. Baseline characteristics of trial participants are shown in Table 2. Participants’ mean ± standard deviation age is 69.4 ± 8.6; 48.0% of participants are female, 15.7% are Black, 36.2% have diabetes, 80.4% hypertension, and 25.4% prior MI. Overall, 19.6% of participants reported classic claudication symptoms, and 9.8% reported no lower extremity symptoms with walking. Scores on the Walking Impairment Questionnaire were 44.3 ± 33.9 on the distance subscale, 37.6 ± 24.7 on the speed subscale, 44.7 ± 28.3 on the stairs subscale, and 42.2 ± 25.5 overall. Though GAMEPAD did not require patients to have symptoms with walking to be enrolled, these scores are similar to those observed in cohorts of patients with symptomatic PAD.(35) The mean participant took 4385 ± 2041 steps during the baseline run-in period and set a goal of increasing steps by 1356 ± 1555 during the study period.

Table 2:

Baseline characteristics of enrolled patients

Baseline characteristics	Total (n = 102)
Age	69.4 ± 8.6
Male	53 (52.0)
Race
White	80 (78.4)
Black	16 (15.7)
Asian	2 (2.0)
Hispanic	1 (10)
Other	3 (2.9)
Education
Some high school	2 (2.0)
High school graduate	14 (13.7)
Some college	32 (31.4)
College graduate	54 (52.9)
Annual household income
< $50,000	34 (33.3)
$50,000 to $100,000	34 (33.3)
> $100,000	34 (33.3)
BMI	30.0 ± 6.3
Current smoking	11 (10.8)
Hypertension	82 (80.4)
Hyperlipidemia	72 (70.6)
Diabetes	37 (36.2)
Prior MI	26 (25.5)
Stroke	12 (11.8)
Heart failure	12 (11.8)
COPD	19 (18.6)
Kidney disease	19 (18.6)
Baseline daily step count	4385 ± 2041
Goal step increase	1356 ± 1555
Walking Impairment Questionnaire
Distance	44.3 ± 33.9
Speed	37.6 ± 24.7
Stairs	44.7 ± 28.3
Overall	42.2 ± 25.5
San Diego Claudication Questionnaire
Classic claudication	20 (19.6)
Atypical leg pain	52 (50.9)
No symptoms	10 (9.8)

To date, 94% of participants randomized have remained in the study through 18-week follow-up, and 95% of participants submitted step count data for ≥ 60% of days through the 16-week intervention period.

CONCLUSIONS

The GAME PAD trial, a randomized trial of an entirely home-based walking program for patients with PAD informed by concepts from behavioral economics, will provide several important insights. From the standpoint of clinical practice, few eligible patients with PAD participate in supervised exercise programs, and there are substantial barriers to participation,(12) particularly for Black, rural, and low socioeconomic status patients, groups with particularly poor PAD outcomes.(36) Some home-based exercise programs demonstrate promise, but all previously tested home-based exercise programs involve multiple visits to a centralized supervised exercise center and live consultation with an exercise specialist. Requirements for multiple visits may limit participation for many patients, and the need for specialized exercise specialists may strain resources for some health systems. By contrast, GAME PAD is the first trial to test an entirely home-based and virtual physical activity program in patients with PAD, and does not require access to exercise specialists. Should this intervention prove effective, it could be deployed even in communities without access to vascular subspecialists or supervised exercise therapy. Moreover, though other studies have demonstrated the effectiveness of physical activity programs informed by behavioral economic theory to increase physical activity in patients with or at risk for atherosclerotic cardiovascular disease, GAME PAD seeks to expand this paradigm to patients whose physical activity is limited by exertional symptoms.