Rationale and design of a smartphone‐enabled, home‐based exercise program in patients with symptomatic peripheral arterial disease: The smart step randomized trial

Abstract Background Supervised exercise therapy (SET) is recommended in patients with symptomatic peripheral arterial disease (PAD) as first‐line therapy, although patient adoption remains low. Home‐based exercise therapy (HBET) delivered through smartphones may expand access. The feasibility of such programs, especially in low‐resource settings, remains unknown. Methods Smart Step is a pilot randomized trial of smartphone‐enabled HBET vs walking advice in patients with symptomatic PAD in an inner‐city hospital. Participants receive a smartphone app with daily exercise reminders and educational content. A trained coach performs weekly phone‐based coaching sessions. All participants receive a Fitbit Charge HR 2 to measure physical activity. The primary outcome changes in 6‐minute walking test (6MWT) distance at 12 weeks over baseline. Secondary outcomes are the degree of engagement with the smartphone app and changes in health behaviors and quality of life scores after 12 weeks and 1 year. Results A total of 15 patients are randomized as of December 15, 2019 with a mean (SD) age of 66.1 (5.8) years. The majority are female (60%) and black (87%). At baseline, the mean (SD) ABI and 6MWT were 0.86 (0.29) and 363.5 m, respectively. Enrollment is expected to continue until December 2020 to achieve a target size of 50 participants. Conclusions The potential significance of this trial will be to provide preliminary evidence of a home‐based, “mobile‐first” approach for delivering a structured exercise rehabilitation program. Smartphone‐enabled HBET can be potentially more accessible than center‐based programs, and if proven effective, may have a potential widespread public health benefit.


| INTRODUCTION
An estimated 8.5 million Americans (or 7.2% of the US adult population) are estimated to have peripheral arterial disease (PAD), defined as an ankle-brachial index (ABI) of <0.9. 1 Nearly, 10% of PAD patients are symptomatic with claudication-a reproducible leg pain with ambulation that is relieved with rest. 2 Following the onset of symptoms, patients with PAD are at an increased risk of adverse outcomes including worsening functional status, amputation, myocardial infarction, stroke, and death. 1,3 African Americans in particular are at an increased risk of PAD and experience greater walking impairment and more severe disease than non-Hispanic white individuals. 4,5 Supervised exercise therapy (SET), a 12-week program with multiple in-person weekly sessions, is recommended as a first-line therapy in symptomatic PAD. 6 Current guidelines for SET include that participants perform intermittent bouts of treadmill-based walking exercise at least three times per week. During each session, patients are advised to walk in intervals with an intensity level to induce moderate to moderately-severe leg symptoms within 5 to 10 minutes, at which point they should rest until the pain dissipates. Walking intervals should be repeated with a goal of accumulating up to 60 minutes of total exercise time during each session. 7 Participation in SET has been shown to improve walking ability, overall functional status, and health-related quality of life (QoL) in symptomatic PAD, and is considered first-line therapy with a Class I recommendation from established practice guidelines in patients with claudication. 8,9 Despite the benefits, SET remains significantly underutilized among eligible patients. 10,11 In a recent meta-analysis, less than 25% of patients with claudication enrolled in SET. 10 Participation in SET has been limited by a lack program availability, poor reimbursement, inconvenience, and patient motivation. 12 Even despite a National Coverage Determination by the Centers of Medicare and Medicaid Services in 2017 for SET reimbursement, the availability and uptake of SET programs have remained exceedingly low. 13 This suggests that SET may remain an unlikely option for many patients, especially in low-resource, public-hospital settings where third-party payer coverage for healthcare services is limited.
Home-based exercise therapy (HBET) has been proposed as an alternative for patients with limited access to facility-based SET. 6,14 Structured, home-based programs have the potential to be a more convenient and acceptable option for patients than supervised exercise. 14 Although several randomized trials have documented improved walking ability and QoL from HBET over usual care, more recent trials have failed to confirm these findings. [15][16][17][18][19] 1.1 | Rationale for the smart step study Smartphone ownership is on the rise 20 and new opportunities exist to efficiently provide services with wider accessibility without increases in personnel and capital resources. Smartphones have been used to successfully deliver home-based cardiac rehabilitation (CR), a program similar in broad scope and design to HBET for PAD, with high adherence and low attrition. 21,22 Although several ongoing studies are also evaluating smartphones for various exercise applications in PAD, 23,24 Smart Step is the first study, to our knowledge, to assess the potential for smartphones to deliver HBET in a low resource, public hospital setting with an economically disadvantaged population.
We recently showed the feasibility of smartphone-enabled CR within the Department of Veterans Affairs with high adherence (90%) and satisfaction (80%) and a 20% gain in functional status. 25 Based on knowledge gained from this endeavor, we designed the Smart Step Study to help PAD patients overcome similar barriers to SET participation.

| Hypotheses and objectives
The overarching goal of the Smart Step Study is to evaluate the feasibility of a smartphone-enabled, home-based exercise program for PAD in a low-resource setting where access to a traditional, facilitybased SET program is essentially absent. The primary outcome of interest is functional improvement after 12-weeks of enrollment. We will also assess for participant engagement with the smartphone app and phone-based coaching. If successful, this study will generate novel information on how PAD patients in economically disadvantaged settings respond to a smartphone-based exercise program and provide evidence to proceed with larger trials to determine efficacy.   Table 1.

| Study procedures and follow-up
Upon enrollment, participants are randomly assigned to either the smartphone-enabled HBET or usual care arm and followed prospectively for 12 weeks. Regardless of the treatment arm, all subjects receive a Fitbit Charge 2 (Fitbit, San Francisco, California) to record various measures of physical activity including steps, exercise activity, and heart rate (HR). All participants are then prescribed an exercise plan according to their study arm as outlined below.

| Standard exercise therapy
Patients randomized to the usual care arm receive counseling by their physician to perform self-monitored walking exercise according to established practice guidelines for PAD. This represents current standard practice at Grady Memorial Hospital and includes a recommendation to perform between 30-and 45-minutes walking exercise three times per week. This is based on prior meta-analyses which have found that walking at least 30 minutes is more beneficial than shorter durations, and that benefit appears to peak at 45 minutes. 9,30,31 Participants are instructed to walk at an intensity that induces onset of claudication within 3 to 5 minutes and moderate to moderately severe claudication within 8 to 10 minutes, at which point they should stop and rest until their claudication subsides. After their symptoms dissipate, they should resume this same cycle of exercise until their target Inclusion and exclusion criteria Inclusion criteria • Age 18-89 years • Able to provide informed consent • Clinically stable intermittent claudication • Diagnosis of peripheral arterial disease (PAD) with one of the following: Ankle-brachial index (ABI) ≤ 0.9 or ABI > 0.9 but ≤1.00 at rest with 20% drop in ABI with exercise or heel-rise test or ABI > 0.9 with other evidence of lower extremity PAD such as a positive anatomic study (ie, ultrasound, computed tomography or invasive angiography) or prior lower extremity revascularization (surgical or endovascular) for PAD or ABI > 1.3 with additional evidence of PAD (ie, pulse volume recording [PVR], toe-brachial index or arterial duplex studies). • Possess a working Android or iOS smartphone to install the Movn app Exclusion criteria Justification 1. Individuals whose walking is limited by a condition other than PAD (eg, severe angina or dyspnea, arthritis, muscle weakness or pain).
Intervention is designed to improve PAD related walking impairment. Exercise may be potentially unsafe for these patients.
3. Active behavioral conditions such as uncontrolled schizophrenia or illicit drug addiction that, in the opinion of the study team, will interfere with active participation.
May interfere with ability to fully engage in the study.
exercise duration is reached. In addition to exercise, participants in the standard exercise group receive counseling on healthy lifestyle behaviors including dietary changes, medication adherence, and smoking cessation if needed.

| Smartphone-enabled exercise therapy
Patients in the smartphone-enabled arm receive an app (Movn, Moving Analytics, Los Angeles, California) to guide their exercise program in addition to the fitness tracker which also synchronizes with the app. Our study team has successfully piloted the use of Movn in the delivery of home-based CR to veterans with ischemic heart disease. 25 Movn is a commercially available home-based exercise rehabilitation program based on the MULTIFIT, 32 a case-management system for secondary prevention and patient surveillance after acute MI devel-

| Study measures
The primary outcome changes in 6-minute walk distance between

| 6-minute walk test
The 6-minute walk test (6MWT) distance is an objective and wellvalidated measure of walking ability that predicts mobility loss and mortality in PAD with excellent test-retest reliability, and is our

(A) Daily Reminders (C) Provider-Case Management Dashboard (B) Two-Way Messaging
F I G U R E 2 Movn smartphone app and dashboard primary endpoint. 12,34,35 Distances walked between repeat 6MWTs by PAD patients have a high-reliability coefficient (>0.90) and a low coefficient of variation (~10%). 34,35 The 6MWT offers several advantages over treadmill testing in PAD as it correlates more closely with physical activity levels in the community and is not associated with a learning effect when repeated testing is performed. 36,37 Participants perform a 6MWT by walking back and forth along a 100-ft hallway using a standard protocol. 19 The total distance walked, claudication onset time, defined as the time at which ambulation could not continue due to maximal pain, and the pain-free walking distance, are recorded to quantify the severity of claudication. In a recent randomized study, the minimal clinically important difference (MCID) for 6MWT distance for small, moderate and large changes after 3 months of home-exercise were 11, 28, and 45 m, respectively. 38

| Survey data
Data from both the Walking Impairment Questionnaire (WIQ) and Short-Form 36 (SF-36) are collected as secondary outcomes. All surveys are self-administered by the participant. The WIQ is a validated PAD-specific measure of patient-reported walking ability including walking speed and distance, and the ability to climb stairs. 39 Each domain is scored on a scale of 0 to 100 with 0 representing the most significant limitation and 100 representing no difficulty. Changes in walking ability as a result of therapeutic interventions, such as exercise, have been correlated with improvements in the WIQ score. 18,19 The SF-36 is a well-validated QoL assessment that is frequently used to assess changes in response to therapeutic interventions in PAD, including trials of home exercise. 19,40 The SF-36 includes eight health dimensions (eg, physical functioning, bodily pain, general health) which will be converted to a scale ranging from 0 (worst possible score) to 100 (best possible score). 40

| Daily ambulatory activity
Daily ambulatory activity is assessed as a secondary endpoint using a wrist-worn activity monitor (Fitbit Charge HR) which measures step counts, actigraphy, and resting HR. Fitbit devices have been shown to be more valid than other consumer activity monitors in measuring walking activity. 41 Participants are instructed to wear the device daily during waking hours and remove it at night to recharge. The device is connected via Bluetooth to their personal smartphone device via the Fitbit app. The device records the number of steps taken and other data points on a minute-to-minute basis. With participants' consent, a member of the study team remotely accesses the participants' activity data using the Fitbit online dashboard.

| Ankle-brachial index
All participants undergo measurement of the ABI at baseline to determine eligibility. The ABI is associated with the degree of functional impairment and the rate of functional decline among older adults in PAD. 28 Since prior randomized trials of exercise therapy in PAD have not demonstrated an improvement in ABI after exercise, the ABI is not repeated after the 6MWT or during any follow-up visit. After 10 minutes at rest, an ABI is obtained from the more severely diseased lower extremity using a standard protocol. 42 Both the ankle and brachial systolic pressures are measured by Doppler. Measurements are taken from the posterior tibial and dorsalis pedis arteries at the ankle in both legs, and the higher of the arterial pressures from the more severely diseased legs is recorded as the resting ankle systolic pressure. Similarly, the higher systolic pressure is recorded as the brachial systolic pressure for the ABI, calculated as the ratio of the ankle systolic pressure to the brachial systolic pressure.

| Key endpoints
Our primary outcome is change in 6MWT distance from baseline to 12 weeks (ie, end of program completion). All remaining outcomes are secondary and include survey data from the WIQ and SF-36 questionnaires, changes in daily ambulatory activity from the Fitbit, modifiable PAD risk factors (such as blood pressure and smoking status), selfreported exercise and medication adherence collected via the smartphone app, the degree of participant engagement with the app individual features (eg, chats, educational videos), and pre-specified

| Power and sample size calculations
We powered the study on the primary outcome of 6MWT distance.
All other outcomes are secondary. In a recent study of HBET, the esti-

| Statistical analysis
For analyses involving comparisons of groups we will perform twosample, independent t-tests for normally distributed data, and logarithmic transformations for positive skewed distributions. For highly skewed data where log transformation does not adequately correct the skew, we will use non-parametric rank-order tests such as the Wilcoxon rank sum. For between-group differences, 95% confidence intervals will be constructed, and multivariable adjustments using linear regression will be used where large group imbalances exist. Analyses will be performed as intention-to-treat, but we will also do sub-group analyses based on the degree of engagement with the program.
Multiple imputations will be used for missing data using established methods, 44 and per-protocol analyses will also be conducted. The a priori level for statistical significance was a two-sided P < .05.
There are no sub-studies planned, but we will perform predefined subgroup analyses to explore efficacy according to several baseline characteristics including age, ABI results, functional status, revascularization history, and degree of engagement with the smartphone app and coaching program. Step study is the first study to evaluate a smartphone-enabled HBET program in a low resource, public-hospital setting with a majority of African American patients. The Smart Step trial also integrates a contemporary wearable activity monitor with the Fitbit Charge 2 which will complement that of several other trials which assessed the value of mobile technologies in the evaluation and management of symptomatic PAD. [17][18][19]43 The Smart Step Study is unique because it primarily targets patients with lower socioeconomic status (SES) in an inner-city hospital. Although SET program availability is limited in general, PAD patients in underserved populations may have an even higher risk of poor outcomes due to lack of adequate insurance and reliance on lower-resource healthcare systems without available SET programs.  19 Participants in the home-exercise group completed 12-weeks of exercise, initially with weekly sessions at the hospital for 4 weeks followed by 8 weeks of completely home-based exercise with telephone coaching and with the wearable monitor. ischemic heart disease and stroke. 47

| CONCLUSION
Access to a structured exercise walking program is of paramount importance to patients with PAD; however, the multiple barriers including limited availability, inconvenience, and cost has made delivery of facility-based SET programs challenging. A smartphone-enabled HBET program, if effective, could have far-reaching implications for the growing number of PAD patients especially in lower resource settings where participation in SET is virtually nonexistent.