The National Sleep Foundation defines obstructive sleep apnea (OSA) as a chronic condition characterized by involuntary breathing cessation while the patient is asleep. It is a common disorder that affects more than 22 million Americans [1]. A previous study estimated that OSA affects 26% of adults aged between 30 and 70 years [2]. Gender, age, family history, obesity, smoking, and alcohol use often make patients more vulnerable to OSA. Furthermore, OSA is increasingly recognized as a risk factor for medical conditions, such as hypertension, cardiovascular diseases, and stroke [3]. Therefore, it is important to diagnose and treat OSA early. Continuous positive airway pressure (CPAP) remains the gold standard for treatment [4]. For patients who do not respond to CPAP, other treatment options may be prescribed based on the severity of the condition. Increasing awareness coupled with better prognosis from timely interventions has led to an exponential surge in the demand for evaluation and treatment services, but there has not been a commensurate increase in resources to deal with this increasing demand. The diagnosis and treatment of OSA are particularly challenging from an operational perspective because of the limited number of sleep clinics [5]. To compound this problem, there is also an acute shortage of clinical expertise for delivering these services, with the ratio of general population to certified sleep specialists being as high as 43,000:1 [6]. The situation is further exacerbated by their inequitable distribution, with most specialists serving in urban areas. Therefore, there is a need to explore other approaches for diagnosing and treating patients with sleep apnea. Telemedicine is a potentially viable alternative.

In this study, we explore this alternative by developing and evaluating a telemedicine platform for the diagnosis and treatment of OSA. Specifically, we addressed the following two research questions:

Following the design science research (DSR) framework [7,8], we developed a telemedicine platform called Ognomy (North American Sleep Management Inc), where the data collection, consultation with physicians, patient education, and monitoring occur using a mobile app. Access to nursing staff, sleep technicians, and physicians is available through video consultation. Artificial intelligence (AI)–based chatbots for self-help and helpline contacts are provided to assist care seekers in using the app effectively. Such a platform is expected to reduce costs and increase access to care while maintaining quality expectations.

The World Health Organization defines telemedicine as “the use of information and communication technologies in the delivery of health care services, the diagnosis of a condition, prevention of diseases and injuries, treatment, research, and evaluation of patients at a distance” [9]. Owing to its advantages, the development, deployment, and use of telemedicine have surged in the United States [10] and other developed nations [11]. Recent studies have shown that telemedicine interventions effectively improve clinical outcomes [12], decrease costs by reducing expensive inpatient service utilization [13], and provide a superior patient experience [14], particularly in the areas of chronic disease management, such as diabetes [15], chronic heart diseases [16], thyroid disorders [17], varicose veins [18], stroke [19], mental and emotional health issues [20], and sexually transmitted diseases [21]. Previous research has also discussed the benefits, drawbacks, and barriers to the adoption of telemedicine. Some common barriers include patients’ poor perception of its efficacy [22], nonacceptance by insurance companies [23], and technology hurdles [24]. However, with the growing maturity of technology platforms and increasing patient acceptance, telemedicine interventions have begun to overcome these barriers. A study by Watson et al [25] found that monitoring patients via telemedicine is as effective as traditional medical processes. Furthermore, some studies [26,27] have explored the use of telemedicine to monitor compliance with CPAP therapy. However, the design of telemedicine interventions that focus on testing and diagnosing patients with OSA has not been well addressed in previous research. Our study addresses this gap by developing a comprehensive telemedicine platform that caters to all stakeholders involved in seeking and providing OSA care.

Overall, we follow a DSR approach, a principal research methodology that has evolved within the information systems (ISs) discipline. It is defined as the development of a system by an iterative process of developing and evaluating artifacts [28]. The primary goal of this study is to define and develop a telemedicine platform that caters to all stakeholders involved in diagnosing, treating, and managing sleep apnea. Previous research has discussed the process of conducting and evaluating design science [29]. We used the framework suggested by Horvath [30]. Our research approach is illustrated in Figure 1.

Furthermore, this framework divides DSR into three phases: (1) exploring the problem, the context, and the activities using activity theory; (2) designing and developing solutions; and (3) evaluating solutions and generalizing design principles.

We present the details of the abovementioned three phases in the following sections.

We used activity theory to guide our understanding of the end user and business context. According to activity theory, an activity is a basic unit of analysis [31], which includes six components: (1) subjects or people involved in the activity, (2) object or intended goal of the activity, (3) tools that mediate the activity, (4) rules imposed by the business environment, (5) division of labor among various stakeholders in the organization, and (6) the community impacted by the activity. Activity theory was an apt choice for our study because it emphasizes developing a comprehensive understanding of the environment in which the system operates (which includes participants and subsystems), including all the complexity that manifests from the interactions of various actors in the real world rather than from the limited viewpoint of a single actor [32]. Furthermore, activity theory allows us to develop flexible systems that dynamically cater to various users’ specific workflows, such as patients, providers, office administrators, and information technology (IT) staff. Contemporary activity theory focuses on the interaction of several systems of activities to investigate a complex social phenomenon. In doing so, it brings to the fore some of the deeply embedded contextual issues associated with the research topic [33].

At the outset, we identified the key stakeholders (subjects) involved in the process of seeking and providing care for sleep apnea, including patients, providers, referring physicians, administrative staff such as billers and coders, and back-office personnel. To develop a deeper understanding of their roles in seeking and providing care, the resources they require, and the challenges they face, we conducted six structured interviews with each of these stakeholders and two brainstorming workshops. Details of interview participants and interview questions are presented in Multimedia Appendix 1. The dimensions of activity theory guided the questions and the line of inquiry. On the basis of interviews and brainstorming sessions, we identified Health Information Portability and Accountability Act (HIPAA) application forms, health forms, information about the diagnosis, and treatment protocols as critical resources. Enrollment, consultation, and monitoring emerged as crucial processes. In the interviews, patients expressed a strong preference for a mobile app for consultation, whereas providers favored a web portal for documentation, reporting, and integration with an electronic medical record (EMR) system. During the interviews, the collection of patient information, such as personal and contact details, HIPAA consent, insurance information, and prior health history, emerged as a crucial design challenge. The crux of the problem was to collect all relevant data from patients without overwhelming them. Owing to their limited interactivity, traditional data entry forms are not very effective in engaging users, especially when a large amount of information is collected. Guided by prior research on the effectiveness of chatbots in improving patient engagement [34,35] and care experience [36], we decided to develop a rule-based AI chatbot for enhanced engagement and better patient interaction.

In addition, stakeholders pointed out several business rules and constraints that emerged from the business context. These rules are subsequently coded into the system to develop a patient-first, state-of-the-art telemedicine platform for patients with OSA. The contextual details are presented in Table 1 along the dimensions of activity, subject, instrument, object, division of labor, community, and rules.

This design science study makes several notable contributions to both theory and practice. We make several theoretical contributions of this study. Although ISs research in the health care field has been ripe with development [47], there have not been many interventions developed for patients seeking sleep apnea care [48]. Furthermore, extant DSR has mostly emphasized the design, development, and evaluation of IT solution artifacts [49]. Exploring and validating the problem space have been insufficiently pursued. In this study, we took advantage of the design knowledge model [50] to bifurcate the design space into problem and solution spaces. Problem space validation establishes the relevance of the research product, whereas solution space addresses technical feasibility and product usability.

This study also informs the ISs research community on how telemedicine interventions can be implemented using the activity theory framework for building and evaluating artifacts. Activity theory in DSR has been used to design data models [51] and mobile apps in an educational context [52]. In this study, we enhanced activity theory by mapping the business activity system to a problem space and the technical activity system to a solution space. Furthermore, we argue that any IT artifact embodies the social and technical aspects of system development. The success of an artifact depends on the interactions between business and technical activity systems.

From a practice perspective, we designed and developed a telemedicine platform and demonstrated the potential for such an intervention. We also showed how novel methods for managing products and developing digital architecture can quickly and effectively scale digital interventions. This study generated transferable insights into the most effective practices in conceptualizing and developing digital health products. A total of four core themes of generalizable guidelines emerged:

We have provided an elaborate discussion on these themes in the following sections.

As highlighted in several prior studies, software development organizations may be tempted to quickly cruise through the conceptualization and design phases without validating the problem space from an end user perspective [53]. However, such practices usually lead to the development of products that are not grounded in their customers’ needs. Researchers frequently use co-design workshops and structured interviews with different stakeholders to overcome this challenge [54].

In this study, we note that the diversity of opinions early on is especially important, as it provides a good platform for all ideas to emerge. It facilitates the holistic development of an app that is relevant, usable, and valuable to all stakeholders, including patients and clinicians, thus overcoming potential barriers to successful adoption and continued use. This implication is particularly relevant for complex business domains, such as health care, where several subject matter experts, such as medical providers (doctors and nurses), administrators (billing, coders, and operation managers), and sleep technicians, interact with end users (patients).

Extant literature has emphasized that weeding out feature ideas without potential quickly and economically can be a critical success factor in the development of digital platforms [55,56]. To accomplish this goal, development organizations must leverage a combination of novel frugal engineering methods to save time and money. For instance, in our study, we used iterative prototyping with varying fidelity levels to understand the potential problems before building the final product.

The team used layered modular architecture using the software as a service app hosted on the cloud environment. Such architecture helps borrow capabilities from other apps for a fee instead of developing them from scratch. In addition to being robust, these services reduce the development time considerably and do not require a large upfront investment. Furthermore, the app was hosted on the AWS cloud environment, which complied with all technology-related HIPAA requirements. The team also used tools such as EC2 (Elastic Compute Cloud), S3 (Simple Storage Service), and ELB (Elastic Load Balancer) to enhance the ease of deployment and migration and to meet on-demand scaling.

Prior research has highlighted the shortage of software development talent [57]. This challenge is pronounced in nontechnology hubs. Although many universities offer courses in software engineering and several boot camps and training centers have arisen to fill the talent gap, the availability of solid software development talent, especially in the latest technologies such as React JS and dev-ops automation that can deliver apps quickly in short time frames, remains a significant bottleneck.

This study had several limitations. We conducted formative usability tests with only a limited number of end users, including 8 patients, 3 providers, and 3 sleep clinic administrators. As a result, the issues identified in the usability tests may not be generalizable to other settings. Furthermore, all the participants of the usability tests were from the northeastern United States, and hence, the sample of participants may not be nationally representative. However, our focus was not on the generalizability of our findings but on gathering practical insights about the challenges faced by the end users of the app.

Although the participants testing the patient-facing apps exhibited diversity in age, gender, and educational accomplishments, they all had access to smartphones and were reasonably comfortable using them. Our findings might have been significantly different if we had focused our attention on individuals who lacked technology proficiency or who had cognitive or physical challenges. However, given the ubiquity of mobile devices and the large number of smartphone users, we deliberately focused on this market segment.

Another limitation is that not all the issues identified during the design review, system test, and formative usability tests were addressed in the current version of the app. We refined the design artifact based on the priority of the concerns raised, as we were constrained by time and budget. All the critical issues and errors associated with the patient-facing features, such as appointment management and video consultation, were addressed. However, issues associated with EMRs and computerized physician order entry integration; physician reports; and administrative features, such as insurance card verification, were deferred for future releases.

Against the backdrop of a surge in demand for sleep apnea care, the telemedicine platform offers a scalable and economical alternative. With an aging population and the escalating cost of care, telemedicine alternatives have become increasingly imperative. This is especially true for patients seeking care for sleep apnea because of its increasing prevalence. This paper discusses the design, development, and evaluation of a telemedicine intervention that aids providers and care seekers in virtual consultation, conducts tests, and prescribes treatment virtually. Such a telemedicine platform breaks access barriers while ensuring high-quality care. During the evaluation phase, we noted that this platform is economically viable, technically feasible, and highly usable by all stakeholder groups. Future research can extend this line of inquiry by developing artifact instances for other sleep-related issues, such as insomnia. An interesting contribution would be to demonstrate how telemedicine interventions are accepted across different cultures and geographies, especially where there is a lack of awareness about sleep issues. Future work could also address related issues of e-consent and explore the legal and ethical guidelines for using such platforms.