The control group receives a list of four hyperlinks on a unique webpage of available online information that is Canadian, French, and that included information on walking. Three major Canadian nonprofit or public organizations are included: Montreal Heart Institute, Heart and Stroke Foundation, and Canadian Society for Exercise Physiology. In addition, because key recommendations on walking post‒ACS-related hospitalization were derived from a patient education booklet published by the Montreal Heart Institute that is available online, the walking program in this booklet is also included in the list of hyperlinks. All websites provide information in text format without the use of videos.

The experimental group receives access to TAVIE en m@rche. The central feature of TAVIE en m@rche is the prerecorded videos of a nurse (see Multimedia Appendix 2) who presents the fully automated tailored intervention content delivered according to patients’ baseline assessments of autonomous motivation, perceived competence, and walking behavior. Other modes of delivery include online text that appears beside the videos to allow simultaneous reading of the video’s content, and downloadable PDF files referred to by the nurse. Access to TAVIE en m@rche starts at randomization between the fourth and fifth week after hospitalization, which depends on when the baseline online assessment is submitted. The suggested completion time of the intervention is 4 weeks but access to the intervention ends at 11 weeks postrandomization. An additional brief “booster” is added at 8 weeks postrandomization. We estimate about 60-75 minutes is sufficient to complete the intervention. TAVIE en m@rche consists of 73 videos, each lasting on average nearly 1 minute with most (n=68) lasting less than 2 minutes. The TAVIE platform has simple webpage layouts and is easy to navigate [41].

The intervention goal is to encourage a progressive increase in walking behavior, up to the recommended 150 minutes per week at moderate intensity, which is determined by an adapted version of the Borg Rating of Perceived Exertion [42]. This walking level is recommended to all patients at discharge for an ACS-related hospitalization by their treating physician unless a contraindication is present such as comorbid physical condition or an environmental constraint. Such patients are ineligible for study participation.

The intervention is based on a theoretical framework that integrates SBNC with SDT. SBNC focuses on nursing values such as fostering a collaborative partnership with the person, supporting the person’s self-determination in their decisions and actions, and working with the person’s strengths in the aim of achieving health and healing. The SDT on human motivation specifies theoretical constructs for physical activity to be targeted by the intervention strategies and to drive the tailoring process. The intervention strategies are specifically targeted toward increasing self-reported perceived autonomy support, autonomous motivation, and/or confidence (combined perceived competence and barrier self-efficacy).

The appeal of using videos as the main mode of delivery, rather than text-only format, is the greater ability to convey the strengths-based nursing way of being that is manifested in part by nonverbal behaviors such as tone of voice (eg, energetic vs neutral) and body language (eg, smiling vs a sincere nonjudgmental expression), and by verbal behaviors (ie, the nurse’s script). This script, consistent with both SBNC and SDT, drawn from our past literature review [43], followed five global strategies: being collaborative, being strengths-focused, providing choice, offering rationale, and expressing empathy. These global strategies can be thought of as the fabric in which the entire intervention content is interwoven. As such, we expect that the use of videos instead of text-only format will be more interesting and motivating to participants because the SBNC way of being will be better conveyed.

The intervention consists of four specific strategies targeting increasing perceived autonomy support, autonomous motivation, and/or confidence aimed at increasing walking behavior: Strategy (1) Providing information and feedback to build motivation and confidence; Strategy (2) Exploring reasons to build motivation; Strategy (3) Exploring personal strengths to build confidence; and Strategy (4) Developing an action plan to build and consolidate motivation and confidence. These strategies are operationalized by 19 behavior change techniques. The terminologies of those 19 techniques were made consistent with those of the CALO-RE taxonomy [44] (Table 1). These behavior change techniques were drawn from three main literary sources: (1) SDT-based physical activity face-to-face or Web-based interventions and Motivational Interviewing due to its consistency with SDT [45], (2) facilitators of physical activity such as improved cardiac health and quality of life, and barriers of physical activity such as lack of time and the presence of fatigue and depressive symptoms found in cardiac patients, and (3) two patient education booklets on the secondary prevention of ACS [42,46]. The content validity of comparing intervention strategies (global and specific) with the theoretical background was done by the Université de Montréal scientific PhD jury of the first author. One cardiac nurse reviewed the operationalization of the entire intervention, and two clinical kinesiologists in secondary prevention reviewed the operationalized information on walking.

The strategies are conveyed through a set of videos that build toward participants’ commitment to developing their own action plan. The order of the strategies is determined by the primary static tailoring method that is driven by participants’ baseline self-reported autonomous motivation (low vs high), confidence (low vs high), and walking behavior, which resulted in four tailored profiles: A, B, C, and D (Figure 1). Profiles A, B, and C are assigned from scores that are below the recommended 150 minutes per week of walking. Profile A receives Strategies 1 (information), 2 (reasons), 3 (strengths), and 4 (action plan) because this profile is low in motivation and confidence. Profile B receives Strategies 1 (information), 2 (reasons), and 4 (action plan) because this profile is low in motivation. Profile C receives Strategies 1 (information), 3 (strengths), and 4 (action plan) because this profile is low in confidence. Profile D receives Strategies 1 (information), and 4 (action plan) because this profile is high in motivation and confidence. In addition, participants who attained the recommended minutes per week of walking between hospital discharge and baseline receive Profile D.

Secondary methods are the use of tailored messages based on “yes” versus “no” responses to questions after intervention login on identifying symptoms of exercise intolerance in the past 7 days (Introduction). Participants who respond “yes” to having identified symptoms of exercise intolerance are provided an onscreen video message asking them to not initiate the intervention, to consult a free 24-hour province-wide phone service for general health problems if the symptoms are nonurgent, or to call 9-1-1 or go to the emergency department if the symptoms are urgent, and then to log out of the intervention. Two weeks later, participants are asked to log in to the intervention, and only if no symptoms of exercise intolerance are identified by the participant, they are invited to continue the intervention. Static tailored messages on walking behavior (ie, feedback on performance) are also provided to participants in all four profiles (Strategy 1 information) based on their responses of walking behavior assessed only at baseline. Other tailored messages based on “yes” versus “no” responses to questions after intervention login pertain to the identification of personal reasons for walking (Strategy 2 reasons), personal strengths (Strategy 3 strengths), personal goals that are SMART (Specific, Measurable, Attainable, Realistic, and within a Timeframe) (Strategy 4 action plan), social support, and solutions to barriers (Strategy 4 action plan).

The primary outcome is change in steps per day measured by an accelerometer between randomization and 12 weeks. We chose 12 weeks as the primary endpoint rather than 5 weeks to assess the persistence of steps performed beyond the end of the intervention’s 4-week period. The accelerometer step count data are concealed, uploaded wirelessly to a server [50], and downloaded in the first author’s computer. Similar to step counts measured by a previously validated pedometer, step counts measured by the accelerometer worn on a shoe had less than 2% error compared with observed step counts measured by hand tally counter [50]. Participants are instructed to clip the accelerometer on one of their shoes during waking hours. If they are not wearing shoes, they are then instructed to clip it at their waistline clothing (belt or pants) as recommended by the manufacturer.

The steps per day mean will be estimated using ≥3 valid step-days within the 7-day wear period, which is an accepted norm in adult populations [51]. A valid step-day will be determined by a wear time of ≥10 hours per day [51]. Fewer than 3 valid step-days will be treated as missing data.

Secondary outcomes are change in steps per day measured by an accelerometer between randomization and 5 weeks, and in self-reported energy expenditure for walking, and for moderate to vigorous physical activity measured by the French short version International Physical Activity Questionnaire (IPAQ) [52] between randomization, and 5 and 12 weeks. For self-reported energy expenditure, we retained six of the seven items that provided a single continuous score of Metabolic Equivalent of Task (METs)-minutes per week in the last 7 days. The score of energy expenditure for walking is the product of days performed in walking, minutes performed per day, and 3.3 METs. The score of energy expenditure for moderate to vigorous physical activity is the sum of two products: the product of days performed in moderate intensity physical activity (eg, carrying light loads or bicycling at a regular pace), minutes performed per day, and 4.0 METs; and the product of days performed in vigorous intensity physical activity (eg, heavy lifting, or fast bicycling), minutes performed per day, and 8.0 METs. International studies found that the reliability (test-retest) and criterion validity (self-report vs accelerometer data) of the IPAQ generally score around .80 (reliability) and .30 (criterion validity), which is comparable with psychometrics of other self-report physical activity questionnaires [53].

Two outcomes assessed only at 5 weeks for Perceived autonomy support (PAS) were drawn from a French version of the Important Other Climate Questionnaire: (1) from a significant other (PAS-SO) and (2) from the intervention (PAS-WEB). These measures assess autonomy support felt from a significant other (PAS-SO) and from either research website visited (PAS-WEB). The two scores are the mean of responses of 6 items for each PAS (significant other [SO] vs intervention [WEB]) rated between “not at all true” (1) and “very true” (7). Higher scores represent greater levels of PAS. Reported Cronbach alphas across three assessments were between .86 and .89 [54].

Self-determined motivation is assessed at baseline and 5 weeks by the French version of the Treatment Self-Regulation Questionnaire. This measure assesses reasons to attain the recommendation of walking 150 minutes per week. We retained the 12 items that assess controlled motivation (6 items) and autonomous motivation (6 items). The two scores are the mean of responses of 6 items for each motivation (controlled vs autonomous) rated between “not at all true” (1) and “very true” (7). Higher scores represent greater levels of controlled and autonomous motivation. Reported Cronbach alphas across four populations were between .73 and .91 for controlled motivation, and between .85 and .93 for autonomous motivation [55].

Perceived competence is assessed at baseline and 5 weeks by the French version Perceived Competence Scale [53]. This measure assesses confidence to attain the recommendation of walking 150 minutes per week. The score is the mean of responses of 4 items rated between “not true at all” (1) and “very true” (7). Higher scores represent greater levels of perceived competence. Reported Cronbach alphas across two assessments were .93 and .96 [54].

Barrier self-efficacy is assessed at baseline and 5 weeks by the French version Barrier Self-Efficacy Scale for cardiac patients [35]. This measure assesses confidence to walk for the recommended 150 minutes per week even if one or more of eight barriers listed are experienced. We retained 8 of the 9 items. The item removed referred to the barrier of experiencing angina or chest pain. Instead of overcoming this barrier, we expect that participants treat the pain and consult a health care professional if the pain is not relieved instead of continuing to increase their walking behavior. The score is the mean of responses of 8 items rated between “(0%) not at all confident” and “(100%) very confident.” Higher scores represent greater levels of barrier self-efficacy. A reported Cronbach alpha was .86 in the original 9-item scale [35].

Quality of life is assessed at baseline and 12 weeks by the French version MacNew Heart Disease Health-related Quality of Life Questionnaire for cardiac patients [28]. The 27 items assess, in the previous 2 weeks, global quality of life and its 3 subdimensions: emotional (14 items), physical (13 items), and social (13 items) [56]. Items include reverse scores and overlap across dimensions. The scores for global quality of life and each dimension are the mean of responses that range between 1 (poor quality of life), and 7 (high quality of life). Reported Cronbach alphas were .94, .94, .89, and .90 in global, emotional, physical, and social respectively [57].

Self-reported 7-day point prevalence smoking status, an accepted norm in assessing smoking status outcomes [58], is assessed at baseline and 12 weeks. The following question is used: “Have you smoked a cigarette, even a puff, in the past 7 days?” (p. 4 [59]), answered with “yes” (0), “no” (1), or “never smoked” (2). Point prevalence assessment had a sensitivity of 96.9% and specificity of 93.4% in detecting dichotomous smoking versus nonsmoking status compared with urinary cotinine [60].

Medication adherence is assessed at baseline and 12 weeks by the Self-Reported Morisky Medication Adherence Scale (MMAS-4) [61]. This measure assesses barriers to cardiac medication use in the previous 2 weeks such as forgetting to take them and stopping them because one feels well. The score is the sum of 4 items rated “no” (0) or “yes” (1) such that lower scores indicate better medication adherence. Scores are dichotomized between medium to low (1 to 4) and high (0). The MMAS-4 had a sensitivity and specificity of 81.0% and 44.0% respectively in predicting controlled blood pressure [61].

Secondary prevention program attendance is measured at 12 weeks by self-report rated by “no” (0) or “yes” (1) of at least one visit, since hospitalization, to a secondary prevention program that offers clinical follow-up with a health care professional for general health, medication adherence, healthy diet, smoking cessation, or exercise. No data on baseline attendance are collected because programs may start 4 weeks or later posthospitalization, which falls around the planned time of randomization.

Data for both emergency department visits and hospitalizations are collected from the medical records at 12 weeks at each study center. For each outcome, one or more emergency department visits or hospitalizations for any reason indicate a score of 1 and no emergency department visits or hospitalizations indicate a score of 0.

Angina frequency is assessed at baseline and 12 weeks by the angina frequency scale of the Seattle Angina Questionnaire. This measure assesses frequency of angina pain and nitroglycerin use that we changed from “in the past 4 weeks” to “in the past 2 weeks.” The score is the sum of responses of 2 items rated between “4 or more times per day” (1) and “none over the past 2 weeks” (6), which is then transformed to score between 0 (worst) and 100 (best). Lower scores represent greater angina frequency. A reported significant positive association was r=.31 between greater angina frequency and greater number of refills of sublingual nitroglycerin tablets in the previous year [62].

At recruitment in hospital (-T2), sociodemographic and clinical data are collected. Nine items in a paper-based self-report questionnaire assess employment, education, marital status, and other demographics. Other data including medical history, diagnosis, laboratory tests, in-hospital events, cardiac risk factors, intermittent claudication, and documented referral to a secondary prevention program are collected from the medical chart after hospitalization. Also, depressive symptoms are assessed at recruitment (-T2) by the 9-item French version Patient Health Questionnaire (PHQ-9). The PHQ-9 administered at recruitment in hospital allowed us to refer participants with abnormal scores to the treating cardiologist. At baseline (-T1), fatigue is measured by the 7-item short form French version [63]. Based on our literature review prior to commencing our RCT, depression and fatigue were retained as potential covariates rather than outcome variables because of the uncertainty that Web-based interventions in ACS populations can decrease depression symptoms [28] and because it is unknown if such interventions can decrease fatigue as this variable has not been previously tested in the Web-based ACS literature.

During the intervention (T1), intervention adherence data are collected. For TAVIE en m@rche, data are collected on the number of times videos and webpages are viewed and documents are downloaded. Time spent in the intervention will be estimated from these data. For the control group website, data on the number of website visits per person are collected by Google Analytics. Because the control group is provided a single webpage of hyperlinks of publicly available websites, collecting data from these websites is not possible.

First, although our primary outcome of steps per day was retained based on the literature showing the association between increased physical activity and reduced mortality in ACS patients [5-7], we did not plan mortality as primary or secondary outcomes. Trials that aim to improve physical activity outcomes usually establish eligibility criteria to select populations that are capable of attaining the amount of physical activity recommended in the intervention. As such, these populations have few comorbidities resulting in low serious adverse cardiac events or mortality. Two RCTs testing Web-based interventions in ACS patients measured an objective physical activity outcome (steps per day or exercise capacity) and reported serious adverse cardiac or mortality events requiring hospitalization per treatment group [28,29]. Reid et al. reported four hospitalisations for chest pain and no deaths in the experimental group, and six hospitalisations for chest pain, one for cardiac surgery and two deaths in the control group during 12 months of follow-up [28] Lear et al. reported three (9%) major cardiac events (e.g., revascularisation, stroke, and death) in the experimental group, and six (16%) in the control group during 16 months of follow-up [29]. These data suggest that fewer serious adverse cardiac or mortality events are found in favor of Web-based experimental groups, and too few events occur to plan mortality as a primary or secondary outcome within a feasible timeframe.

Second, for our secondary outcome of energy expenditure, we plan estimates from self-reported data instead of from accelerometer data. Accelerometers require the entry of participants’ weight in the devices to produce the estimates. However, we do not collect data on weight from participants at home before randomization (baseline [-T1]) because these data may be missing or unreliable from self-report or from participants’ own weighing scales.

Third, we planned a relatively short follow-up of 12 weeks for feasibility reasons as this RCT is part of a doctoral degree. A longer follow-up on steps per day and other health behavior changes in a future RCT testing TAVIE en m@rche could improve clinical relevance.

Fourth, it is possible that accelerometers are worn by or data from online questionnaires are entered by someone other than the study participant because the outcome data are completed by participants at home. Although we will treat this possibility when examining the outliers, which could reveal some data clearly out of range entered by a different respondent, there are no other provisions made for this limitation.

Fifth, there is a possibility of missing outcome data. Different scenarios for handling missing data will be followed according to best practice in that field [65] by a statistician who is part of an internationally recognized clinical trial reference center (Montreal Health Innovations Coordinating Center). The method of handling missing data will be reported in a future publication of the results.

The platform is limited to using static tailoring rather than dynamic tailoring [20]. However, a recent meta-analysis found that dynamic tailoring has not improved effects on health behavior change outcomes compared with static tailoring [21].

Our sample will likely have similar characteristics as the four other RCTs testing a Web-based intervention using steps per day or another objective physical activity outcome in an ACS population [27-30]. Such populations have no important comorbidities or environmental constraints that would impede performance in moderate-intensity physical activity. ACS populations with important comorbidities are neither eligible to participate in our study nor eligible to receive the recommendation to gradually attain moderate intensity walking beginning the fourth or fifth week after hospitalization, as is recommended in TAVIE en m@rche. Therefore, the eligibility criteria for our RCT is comparable to the ACS population intended for TAVIE en m@rche. Another related cardiac population that could benefit from TAVIE en m@rche, after some minor modifications, are those with stable coronary artery disease, such as stable angina patients requiring elective percutaneous coronary intervention. However, our future results will not be generalizable in stable coronary artery disease populations.