The pilot study used a posttest design with one group of ACS patients hospitalized on the Coronary Care Unit (CCU) in a tertiary cardiology center in Canada. The study was approved by the institutional review board of the Montreal Heart Institute Research Center (project number: 11-1320). Our study is reported in accordance with the CONSORT-EHEALTH checklist version 1.6.1. (see Multimedia Appendix 1) [30]. No content or methodological modifications were made after study commencement.

We recruited a convenience sample of patients hospitalized at the CCU. To be included in the study, participants had to be aged >18 years, be hospitalized for an ACS, be discharged to home within 7 days, understand written and spoken French, and have the physical and cognitive abilities to participate. Patients were excluded if they were hospitalized on the CCU for more than 7 days because this would reflect a more difficult ACS recovery, potentially interfering with their physical abilities to participate.

Enrollment and follow-up occurred from June 2014 to August 2015. Participants were recruited during face-to-face encounters with the project nurse at the CCU. After receiving an explanation regarding the study and providing written consent, patients completed a baseline questionnaire. An individual identification number, username, and password were then provided to participants to allow them to log in to the TAVIE@COEUR Web-based platform. The project-dedicated nurse then showed participants how to use TAVIE@COEUR on a tablet computer and invited them to complete the first session (20 min) at the CCU. Participants were then asked to complete the next two sessions of TAVIE@COEUR at home within 2 weeks of discharge.

All patients received usual care during their hospitalization, including a predischarge teaching session by the bedside nurse on resuming activities of daily living, cardiovascular risk factors, and new medications. This included a pamphlet about medication, prevention of complications, local CR program, and postangioplasty care, which is provided to all patients discharged from the CCU following a coronary event.

TAVIE@COEUR (in French, which translates as YOURLIFE@HEART in English) is a Web-based tailored intervention, including videos in which a virtual nurse interacts with the patient. The ultimate goal of TAVIE@COEUR is to improve patient illness management by providing information and resources tailored to the needs of every patient, while being easily accessible from home. TAVIE@COEUR is completely asynchronous and led by a virtual nurse who guides patients through a learning process regarding the development of illness management skills for self-care and medication adherence (see Figure 1).

Descriptive statistics such as means and standard deviations (SDs) were used for continuous variables, whereas count and percentage were used for dichotomous variables. As this was a pilot study, no inferential analyses were planned; preliminary effect results being provided strictly for illustrating general trends as recommended in pilot study methodology [46].

Participants (N=30) were enrolled in the study from May 2014 to June 2015. Most participants were men, with a mean age of 59 years (SD 9). The majority of the participants were born in Canada and living with a partner (see Table 2).

As shown in Table 3, S1, S2, and S3 were browsed by 30, 17, and 10 participants, respectively. Of these participants, 20 out of 30 completed the mandatory pages in S1, 10 participants completed the mandatory pages in S2, and 5 participants completed the mandatory pages in S3. For those who completed the mandatory pages, the first session lasted a mean of 25 min, whereas the second and third sessions lasted around 16 min.

The general level of satisfaction toward TAVIE@COEUR stayed mostly the same following S1 and following either S2, S3, or both (see Table 4). Items that scored the highest on the acceptability scale were related to the ease of navigation throughout TAVIE@COEUR and the ease of understanding the textual and video content of TAVIE@COEUR. Items that scored the lowest were related to the tailoring of the information and the individual relevance of content in TAVIE@COEUR. Patients identified other areas for improvement such as the pacing of the intervention by the virtual nurse, which they felt was not tailored to their needs. Some believed the content to be less suitable for people who already had a good knowledge of disease and drugs, whereas other patients would have preferred more information on how to prevent the recurrence of cardiac events.

Certain aspects of the intervention scored higher after S1 (tailoring of information and individual relevance), whereas other factors pertaining to the interface scored higher after S3 (ease of navigation and user interface design).

We observed that an overall self-care score 1 month post discharge was higher than the baseline score (see Table 5). Moreover, post-intervention scores for the three dimensions of self-care (medication-taking, recognizing and managing symptoms, and managing changes in the health condition) were all higher than baseline scores.

Responses on the medication adherence scale were skewed toward the score of 0: half of patients reported no omission (score of 0) of medication at 1 month after discharge. Therefore, patients were classified as either reporting no omission (score of 0) versus scores ≥1 omissions. When examining each of the four items of the SRMTS scale, the main reason for nonadherence was forgetting (item 1) to take medication. Seven percent (2 out of 27) of patients scored on either items 2 to 4 on omission, being careless, and stopping medication when feeling better. Patients generally reported a low anxiety level at 1 month post discharge.

We developed and pilot-tested TAVIE@COEUR, a Web-based tailored intervention aimed at supporting illness management in patients hospitalized for an ACS. Feasibility and acceptability results suggest some strengths such as the ease of navigation and the content of TAVIE@COEUR but also underline the need for further development of TAVIE@COEUR. Indeed, we believe that further tailoring of TAVIE@COEUR should be done before undertaking a full-fledged randomized controlled trial (RCT). This may be necessary to increase the engagement of participants to complete all sessions, possibly by interchanging some concrete and more abstract content across the sessions. Preliminary effect results are promising, with improved self-care scores 1 month post discharge related to medication-taking, management of symptoms, and management of changes in the health condition.

Although acceptability scores regarding the TAVIE@COEUR Web-based tailored intervention were generally good, the low completion rate of the 3 sessions underline the significant challenges related to optimizing engagement and adherence in Web-based interventions for illness management. Participants appreciated the ease of navigation throughout the Web-based platform, the ease of understanding of the textual and multimedia content, and the applicability of the information in TAVIE@COEUR. However, relatively low scores were observed for other dimensions of the acceptability questionnaire, such as individual relevance and tailoring of the information; some participants felt that TAVIE@COEUR was not tailored enough to their needs. Moreover, feasibility was not optimal since the global participation rate in the Web-based intervention decreased over the three sessions: 20 out of 30 participants completed S1 (at the CCU), 10 completed S2 (at home), and 5 completed S3 (at home). The lower-than-normal participation rate in the second and third sessions of TAVIE@COEUR may be explained by a variety of factors, including lower motivation to continue the intervention up to the end and lack of individual relevance of the intervention content in the third session. Indeed, it is possible that the content of S3, focusing on social skills and interactions with health professionals, was seen less relevant to participants’ needs, especially because they received all crucial information on medication adherence during the first two sessions. However, the mean duration of sessions was consistent throughout the intervention, suggesting that participants who completed S3 were similarly engaged in the intervention.

More and more electronic health (eHealth) apps and Web-based interventions are addressing health behavior change and are being evaluated in patients with various chronic illnesses [47-49]. Patients’ engagement in completing Web-based interventions is often reported in eHealth intervention research as being about 50% of patients recruited [50,51]. This means that dropout rates of around 50% are observed in several Web-based interventions, limiting their effectiveness in initiating and maintaining changes in health behaviors in patients with cardiovascular diseases [47,48,51-54]. This phenomenon is explained at least, in part, by the fact that most existing Web-based interventions, even those that are called individualized, do not take into account the navigation behavior, knowledge, preferences, and individual objectives of the real-time users [51].

We believe the true challenge lies in developing new strategies for engaging participants with various information needs and individual capabilities in Web-based interventions, such as TAVIE@COEUR. In this study, tailoring was ensured by participants following their own paths, choosing to skip or view the videos, and selecting the appropriate resources to match their needs. However, the general theme of the intervention (illness management geared toward medication-taking) was similar for all participants. We observed that, whereas all participants went through the same life-threatening coronary event and associated medical treatment, their information needs varied widely, more than anticipated with the tailoring allowed in TAVIE@COEUR. Some participants felt the content of TAVIE@COEUR provided more than enough information for their needs, whereas other patients would have liked different resources not related to medication-taking because they were already accustomed to taking medication regularly.

The findings of this study suggest the need for more variety in the options offered to each individual to better match the individual needs of each patient [55,56]. For instance, intelligent tutoring systems (ITS) show great promise to provide tailored resource to patients by adapting the structure and content of the training to learners’ knowledge, preferences, and individual capabilities by using some components of artificial intelligence [57]. ITS are generally conceptualized using four models: (1) the interface model, by which the participant interacts with the ITS; (2) the domain model, which represents the knowledge to learn (clinical and theoretical content); (3) the learner model, which takes into account the knowledge, preferences, and individual capabilities of each participant; and (4) the tutor model representing the pedagogical strategies used throughout the ITS [58]. The combination of these four models, through the use of machine learning and data mining, allow for the tailoring of the instruction to each participant [58]. Whereas ITS with various levels of complexity have been evaluated in academic settings with positive results, very few ITS have been assessed with patients in health care settings [59,60]. We can hypothesize that user-modeling at the beginning of TAVIE@COEUR could allow to better target patients with illness management problems related to medication-taking and redirect those who are okay to other resources related, for example, to nutritional guidelines or ACS pathophysiology. In this sense, we believe a widest range of content should be developed to better represent the interests and needs of patients affected by cardiac diseases.

Future research should also focus on developing new strategies for engaging participants in Web-based interventions, such as interactive feedback based on real-time emotional and cognitive responses, online support groups, serious games, and blended Web-based or in-person interventions. Increasing the accessibility of such interventions is also an important challenge; researchers should make sure that Web-based interventions are adaptive to the devices preferred by each patient (ie, mobile phone, tablet, or computer) [28,29].

The strengths of the study include adherence to the pilot study protocol and the development of an innovative and complex Web-based tailored intervention.

The limitations of this study mostly relate to pilot study characteristics; the single-group, pre-post study design did not allow for causal inferences or for having enough statistical power to detect statistically significant differences. However, this is expected in a pilot study [45,46]. High dropout rates were observed, particularly in the second and third session of the intervention, leading us to believe that changes must be made to the TAVIE@COEUR Web-based intervention before undertaking an RCT. Finally, we observed in our results that the Morisky SRMTS was not designed for properly quantifying the percentage of medication adherence [38]. Future studies should consider using a different tool or scale for measuring more precisely adherence, such as the medication possession ratio [6,52].

Web-based tailored interventions such as TAVIE@COEUR show potential to support illness management for patients in all clinical settings. However, significant challenges remain, and future research must be conducted to optimize factors related to the tailoring of such interventions to patients’ knowledge, needs, interests, and individual capabilities. We strongly believe the right combination of these factors could contribute to the way in which care is provided in health care settings for years to come. With continued population acceptance of Internet use, Web-based interventions are expected to grow exponentially in content and in complexity. Health professionals have an important role to play in keeping these interventions focused on patients and their families to optimize clinical outcomes.