This study adopts a qualitative inquiry approach, which seeks to build an understanding of the world from the experiences, beliefs, and ideas of people while acknowledging that multiple interpretations of reality may coexist [19]. This approach was chosen because it allows for an understanding of patients’ experiences using a digital platform. Semistructured interviewing was chosen as the data collection method as it is useful for gaining an understanding of participants’ perspectives related to a domain of interest while remaining open to exploring other relevant topics during the course of the interview [20]. The study focuses on a single digital platform to enable a richer understanding of the relationship between specific platform features and patient experiences. We sought to recruit patients who had used this platform for at least 1 month to ensure sufficient experience for the study.

We used the SRQR (Standards for Reporting Qualitative Research) reporting guideline when editing this manuscript [21], and its associated checklist is included in Checklist 1.

The study was conducted in compliance with the Declaration of Helsinki, including anonymizing research participant data throughout the analysis and reporting process, ensuring that participants did not feel any undue pressure to take part in the study nor to answer interview questions, and holding interviews at times and locations that were considered convenient for each participant. Written informed consent was obtained from all participants before data collection. Participants were not compensated. The study protocol and all associated methods were approved by the University of Ottawa’s Office of Research Ethics and Integrity (S-11-19-5244).

The study focuses on patients’ experiences with a digital platform referred to as the Chronic Care Platform for the purpose of this study. The platform is used to deliver a chronic disease management program for individuals who have suffered or are at risk of suffering from cardiovascular disease. The program was created and is managed by a hospital providing specialized secondary and tertiary care and had been underway for several years when the study started and remains available. The program targets 2 populations: patients who have received heart disease–related care at this hospital (post discharge) and individuals at risk of experiencing heart disease (prevention). The goal of this program is to help both populations improve their risk factors in order to prevent future life-threatening events.

The Chronic Care Platform is designed to facilitate access to the program and allow the hospital to serve a larger number of individuals. The platform is accessible through a web browser and an app offering features desired by patients to manage their conditions [22]: an online library with varied educational material related to medication, prognosis, and care; a personal health record where patients can record, access, and manage their health information; care and wellness plan; and online rehabilitation support with exercise guidelines. Additionally, the Chronic Care Platform allows scheduling meetings with a health coach for patients who are given access to as part of the program, sharing information with individuals such as family members and physicians as desired by the patient, and interacting with other users through a forum. The platform also supports the development of care and wellness plans, and the tracking and monitoring of varied metrics such as blood pressure, steps, weight, exercise, and mood. It provides visual displays to understand trends related to these metrics (eg, weight gain, daily exercise duration). Information can be tracked through manual entry or by syncing with wearable and health-monitoring devices and apps. As such, the Chronic Care Platform reflects typical digital platforms for self-management.

Participant recruitment followed a convenience sampling strategy. Patients enrolled in the chronic care management program, delivered through the Chronic Care Platform, were invited to participate in the study. Invitations to participate in the study were sent to platform users by the health care program coordinator using the platform messaging system. The message invited interested patients to contact the researchers directly. Recruitment material was also posted on the walls of the hospital where the chronic care management program was being administered. Patients who were using or had used the Chronic Care Platform for at least 1 month were invited to participate in the study. No additional inclusion or exclusion criteria were used in order to allow the recruitment of participants with varied experiences. However, since participants were recruited within the chronic care management program, they were adults (aged 18 y or older) at risk of heart disease or with chronic heart disease. All interested and eligible participants were recruited for the study since the experiences of any user of the platform were of interest to the study.

The first author and a research assistant conducted semistructured interviews with the 24 recruited patients. The first author had experience conducting such interviews and provided training to the research assistant prior to starting interviews. They did not have any preexisting relationship or relationship of authority with the participants. Following the same predefined interview protocol for each interview, patients were asked about their motivation for using the platform, how they used the platform, and the resulting value perceived alongside any challenges in realizing this value (Table 1). Each interview concluded with an open-ended question that encouraged patients to share anything they thought to be relevant about the Chronic Care Platform. Nine interviews were conducted in person and the remainder by phone. In-person interviews were held at locations chosen for their convenience by the participants; these locations included private rooms at the hospital delivering the chronic disease management program and public spaces such as coffee shops. Participants chose a space or table that they felt was appropriate for the interview in the latter case.

Interviews lasted between 20 and 60 minutes, with an average of 30 minutes. They were audio-recorded with the consent of participants and transcribed verbatim by a research assistant and verified for accuracy by the first author. Any identifying information was removed from the transcripts prior to data analysis, and each participant was labeled with a number (eg, P1). As the number of participant interviews increased, substantial redundancies emerged in reported motivations, usage, and perceptions, indicating that the dataset had reached saturation [23]. However, data saturation was not used as a criterion for concluding data collection, and all recruited participants were interviewed.

Template analysis was used to analyze interview data. It is a type of thematic analysis that allows both the inductive identification of themes from data and the deductive application of theory-informed themes [24,25]. Each participant interview was thus first coded inductively using labels that reflected the collected data, and similar codes across participant interviews were then merged into initial themes that reflected broader meanings. This process allowed us to discern emerging patterns of use through recurring themes regarding patients’ motivation for using the platform (eg, “understanding,“ “accessing care”), how the platform was used (eg, “tracking,” “self-educating”), and the kinds of benefits and challenges that patients experienced (eg, “being encouraged,” “irrelevant information”). Two researchers were involved in this process, comparing code application, discussing any discrepancy, and clarifying label and theme definitions.

We then turned to existing theories that could provide a relevant explanation for these patterns. We initially reviewed behavior change theories, including the Theory of Planned Behavior [26] and the Health Belief Model [27]. However, while these theories focus on explaining what may drive individuals’ intended and observed behaviors (eg, the concept of “perceived benefits” in the Health Belief Model), they did not provide sufficient coverage of the emerging themes, nor did they help to achieve the study’s objective. These theories and others that are frequently cited in informatics, such as the Technology Acceptance Model [28], propose concepts that can be used to predict the adoption of healthy behaviors and supporting technology. However, their focus on behavioral intentions does not provide the concepts needed to unpack the behaviors themselves such as the platform uses that were present in our data (eg, “self-educating”). We thus turned to Self-Care Theory [16], which provided a natural fit to our data and emerging themes.

Self-care refers to purposeful actions or activities that are performed deliberately and continuously by an individual to maintain their life, healthy functioning, and well-being [29,30]. These activities can be initiated independently by the individual or in collaboration with health care professionals [30]. The notion of self-care thus aligns with common definitions of self-management [31]. In the context of chronic conditions, self-care is key to empowering people in assuming responsibility for their own health [32]. This understanding of self-care has been formalized in Self-Care Theory, which distinguishes between an individual’s capacity to engage in self-care behavior (self-care abilities), and the actions that can be performed to realize this care (self-care behaviors [17,18,33]).

Self-care abilities encompass 4 groups of factors: cognitive, psychosocial, physiological, and sociocultural [17,18,34] (Textbox 1). While some authors also include demographic factors such as age, we do not include them here as they can be viewed as basic conditioning factors whose impacts are better captured through specific self-care ability domains [34]. Self-care behaviors refer to 3 interrelated processes that are initiated by individuals on their own or in collaboration with health care professionals: self-care maintenance, self-care monitoring, and self-care management [17,18] (Textbox 1).

The Self-Care Theory concepts (in italics in Textbox 1) were deductively applied to identified themes and their data by 2 researchers, comparing code application, discussing any discrepancy, and clarifying code definition. Relationships between concepts were also identified when present in the data (eg, “Cognitive abilities—supporting—self-care maintenance”). Each interview was then coded by 1 researcher, with frequent revision of code application by the second author and inputs on the interpretations by the third author. Both the inductive and deductive phases of data analysis were informed by the research team’s training and experience in digital technology and platforms developed and used in health care contexts (LL, MdR) as well as in cardiovascular health and prevention (KAM). Their combined viewpoints allowed the team to identify meaningful information that could have been glossed over from any single perspective and to compare varied assumptions and interpretations.

A total of 33 potential participants contacted the researchers. Two of them were excluded because they had used the platform less than 1 month, 6 participants did not respond to follow-up communications, and 1 withdrew from the study prior to being interviewed. A sample of 24 participants consisting of 9 (37.5%) women and 15 (62.5%) men was thus recruited. Among them, 79% (19/24) participants were part of the postdischarge program and 21% (5/24) participants belonged to the prevention program. Twelve of 24 (50%) participants were receiving or had received advice from a health coach provided by the program for the first 6 months of use. The frequency at which they used the platform varied from occasional to daily, with 42% (10/24) accessing the platform on a weekly or daily basis. All participants had been using the platform for at least 1 month, 12 (50%) of them having been regular users for 1 year or more. Table 2 provides more detailed information on participant characteristics.

In our analysis, we investigated how motivations for self-care translated into platform use, which subsequently not only generated perceived value but also highlighted challenges for patients in effectively using the platform to achieve the desired goals. Overall, we found that the platform offered support for all self-care abilities and behaviors except physiological abilities (Table 3). The platform, however, shows limitations for self-care management support and important challenges in its support to all self-care abilities and behaviors.

Data show that self-care abilities and behaviors impact each other through the platform’s patterns of use. For example, the informational material supporting the development of cognitive and psychosocial activities (eg, educational material related to stress, recipes) generates learning, which is key to integrating effective lifestyle changes as part of self-care maintenance. Another example is that the platform’s tracking features provide incentive (psychosocial abilities) to enter or import and view health-related data (self-care monitoring), which in turn provides positive reinforcement (psychosocial abilities), supports interactions with the coach (sociocultural abilities), and enables patients to integrate behavioral change into their current lifestyle (self-care maintenance). The results of these interactions are thus positive loops among platform features, abilities, and behaviors.

Conversely, challenges in using the platform and missing features were found to hinder the realization of positive loops among self-care components. For example, a participant (P16) stopped tracking their data since the results were stable and they did not see any reason to continue. Putting this decision in the context of the very limited features provided by the platform for self-care management—features that would include, among others, the ability for users to assess the effectiveness of their lifestyle changes—points to negative impacts of gaps within the Chronic Care Platform that could affect patients’ health outcomes over time.

This study leverages Self-Care Theory [16,18] to analyze patients’ patterns of use—related patient motivations, platform features, patient activities, and resulting perceptions of value—of the Chronic Care Platform. This provides a fine-grained understanding of how platform features can support self-care abilities and behaviors. Such explanations go beyond the insights that can be gained from exploring patients’ willingness or intention to use a platform using the Theory of Planned Behavior and the Health Beliefs Model [4,35], or from overall perceptual measures such as usability, ease of use, and usefulness of digital health technology [13,35].

Understanding the role of the Chronic Care Platform in facilitating self-care provides context to clinical studies that are evaluating the behavioral and clinical impacts of the program being delivered by the platform. An earlier 12-month study of the program conducted with 478 participants showed significant improvements in participants’ cardiovascular disease risk—with an average reduction of 93 (19.5%) in Framingham risk score, improvements in anxiety, perceived stress, physical activity, and more [36]. The program, at that time, offered telephone-based virtual coaching, but was not being delivered using the Chronic Care Platform. A more recent study evaluating the program for postpartum women, delivered using the Chronic Care Platform, showed similar results [37]. A total of 190 women taking part in the 12-month study showed a 102 (53.7%) reduction in Framingham scores alongside a 43 (22.6%) average reduction in Lifetime Risk Scores, significant decrease in body weight, hypertension, and other clinical outcomes, increase in physical activity levels, and healthy eating.

These results contribute to the body of knowledge on the impact of digital health interventions on chronic disease self-management and the digital health technology features that are needed to successfully deliver them. For example, 2 systematic studies, one on platform-based interventions for self-management of chronic diseases [38] and the other on the efficacy of smartphone-based digital health interventions [39], arrived at similar results. That is, both reviews found that the majority of included studies showed significant improvements in behavioral outcomes such as physical activity. However, the results were mixed for clinical outcomes such as blood pressure and body weight. While several included studies showed positive outcomes (sometimes with minimal effects, however), the majority failed to demonstrate statistically significant improvements.

Better understanding of the features and role of the digital health technologies used to deliver such interventions may help to understand variability in clinical outcomes. Digital health technology includes various software and devices used to process and exchange health information on the internet, including websites, wearable devices, smartphones, and more [40]. Their most common functionalities relate to communication (eg, messaging), self-monitoring (eg, tracking of health measures), tailored self-care support (eg, personalized goals and reminders), self-care education (eg, links to reference material), care planning (eg, personal care plan), and community forums [39,40]. The Chronic Care Platform reflects the proposition for digital health platforms to integrate these functionalities into an environment that is easily accessible through various devices. A recurring observation across studies of digital health technology and platforms is that despite their obvious potential for self-managing chronic diseases, they are yet to meet patients’ expectations regarding tailored, personalized support with an acceptable combination of automated and human guidance [38,39,41]. The challenge in meeting these expectations is not solely technological, in particular with the growing availability of large datasets and artificial intelligence (AI) capabilities that can be used to truly personalize care plans, prevention recommendations, alerts, dietary recommendations, and more [42,43]. The results of this study, in line with other studies of digital health technology [38,39], rather emphasize that the gap lies in using these features in a manner that creates value and generates patient engagement and retention.

The use of Self-Care Theory also helps to situate commonly identified patient needs. For example, while other studies have also found that patients express the need for personalized and relevant information when using digital solutions to support chronic disease management [4,44], they do not justify why such information is important. Self-Care Theory helps to understand that personalized information is needed to develop the cognitive abilities required by patients to care for themselves. Another example is the request for features facilitating the organization of health community activities to increase social engagement [13], which can be understood as supporting the development of sociocultural abilities in Self-Care Theory. As new technologies such as AI become integrated into self-care digital platforms [44], this understanding is key to differentiating needs that are core to chronic disease management from those that are specific to AI-based systems.

Moreover, the use of Self-Care Theory to make sense of this study’s results helps to identify gaps in the Chronic Care platform’s features. Figure 1 identifies key features of the platform for each self-care ability and behavior, as well as missing features that could address the challenges experienced by participants. A key gap, the lack of platform support for the development of physiological self-care abilities, can be observed across studies of digital platforms for chronic disease management. This points to a missed opportunity to support chronic disease management more comprehensively, for example, by making platforms interoperable with movement-focused sensors that could assess a patient’s initial state and lead to tailored exercise recommendations. Such features would also help to improve the platform’s limited support for Self-Care Management, which shows important gaps in enabling normative judgment (eg, assessing if a trend is healthy or unhealthy) and providing actionable recommendations (eg, recommendation to call one’s physician). The integration of AI capabilities could address several of these gaps by leveraging sensor-generated and other patient data to better support all self-care behaviors, for example, facilitating meal tracking through image recognition (self-care monitoring) [43]; predicting and alerting the patient about low blood glucose events (self-care management) [42]; and providing tailored nutritional intervention messages to prevent future events (self-care maintenance). However, patients’ reticence in using features such as AI chatbots would need to be mitigated [41,45].

Other gaps pertain to the quality or the way in which functionalities have been designed. For example, the platform does contribute to the development of psychosocial abilities through information and feedback, but could have more impact if the feedback had real-life implications for participants. Similarly, while the ability to share data with selected individuals generated positive value perceptions, the forum only partially supported the development of sociocultural abilities. In both of these examples, the gaps are not functional—information is provided as planned, the forum functions as intended—but rather relate to the inability of available features to meet participants’ expectations.

An apparent disconnect between patient expectations and what is offered through the platform was also apparent in relation to features that facilitate interactions with a health coach, which were positively evaluated by many participants. More generally, results showed that several participants sought reassurance and feedback from health care providers through the platform, which is in line with similar studies [41]. However, virtual care for chronic disease management is often seen as a way to increase efficiency, whereby scarce human resources can be leveraged to care for a greater patient population by transferring self-care responsibilities to the patient while maintaining or improving patient outcomes [46]. Providing features that enable normative judgment and personalized, actionable recommendations may partially address patients’ need to be reassured in their progress. Beyond such automated support, the benefits of added human support should be evaluated against costs when creating digital health interventions for self-management of chronic conditions [38].

Nevertheless, when making design choices for digital health platforms in this context, it is important to account for the ripple effects of individual features across several self-care abilities and behaviors. For example, the support provided by a health coach improves sociocultural abilities for self-care, but these improved abilities in turn affect almost all other abilities and self-care behaviors. Conversely, the lack of mechanisms to make rewards meaningful (gap in psychosocial abilities support) could diminish a patient’s motivation to track their progress (self-care monitoring behavior), hindering their ability to detect and react to changes in health status (self-care management behavior). The same functionalities are now typically integrated across digital health technology and platforms for self-care [39,40], and these functionalities are expected by patients [41]. The challenge of designing new solutions hence does not lie, in the choice or creation of discrete functionalities, but rather in the interlinking of these functionalities to trigger virtuous cycles of self-care abilities and behaviors. This can be achieved in varied ways, for example, by codesigning digital health platforms with target end users to understand if and how features support desired patterns of use and by ensuring software and hardware interoperability (eg, automated data import from self-tracking wearable devices) to facilitate seamless flow among self-care behaviors.

Despite its contributions, the study has several limitations. The study is based on a single digital platform within a single health care setting (cardiovascular disease), which could limit the applicability of the results to other platforms. This limit was however mitigated by analyzing data using a relevant theory that provides generalizable explanations for self-care. As such, the patterns of use identified in this study and resulting recommendations for fully realizing patient value are likely to apply to other digital platforms aiming to support self-care for patients suffering from other chronic diseases. The small number of participants (n=24) also limits the generalizability of results, while being in line with similar qualitative studies [41]. Nevertheless, future research on digital health platforms for chronic disease self-management would benefit from adopting a mixed methods approach with quantitative analysis of representative datasets. Methodological limitations related to analysis and interpretation of the data in qualitative studies were mitigated by involving 3 researchers in the coding process and by iteratively moving between interpretations and data to ensure that findings are grounded in the data as collected. The use of multiple researchers in data analysis corresponds to investigator triangulation, while the consideration of alternative theories as a source of concepts for the deductive analysis phase corresponds to theoretical triangulation [47]. Both types of triangulation used in the study aimed at increasing the trustworthiness of study results.

This study examined patients’ patterns of use of a digital platform to understand how the platform both enabled and limited their ability to self-manage their chronic disease. Digital platforms are increasingly being used in the context of chronic disease self-management, without a full understanding of how their features may influence patients’ experiences and generate desired health outcomes. We contribute to this issue by analyzing what motivates patients to use a chronic self-care platform, how they subsequently use the platform’s features, and how this use generates both positive and negative values. The resulting patterns of use are interpreted as supporting the development of required self-care abilities and the realization of core self-care behaviors [16,18]. Experienced challenges lead to the identification of gaps in platform features that limit patients’ ability to fully care for themselves. This study thus provides a fine-grained understanding of how a given platform can generate positive impacts and how it may be improved to fully support self-care.