Digital health technologies are increasingly being used to monitor, manage, and support health and well-being. The use of digital health promises to increase access to health information, improve the quality of care, reduce errors, and stimulate healthy behavior [1]. Digital health delivery is also often referred to as eHealth [2]. Subsets of eHealth are the application of a specific technology in health delivery (eg, mobile health technologies [3]) or the use of digital technology for a specific purpose (eg, remote delivery of health care services: telehealth or telemedicine [4,5]). The COVID-19 crisis has stimulated the use of digital health, and its application is expected to increase in the coming years [6-8].

The term digital health is defined as the “proper use of technology for improving the health and wellbeing of people at individual and population levels, as well as enhancing the care of patients through intelligent processing of clinical and genetic data” [9]. In this paper, we aim to shed light on the term well-being in this definition. Studies on the meaning of well-being date back to ancient Greece. Since then, many disciplines have reflected on this term. Well-being is commonly considered as “a state of persons which designates that they are happy or flourishing and that their life is going well for them” [10]. It is seen as “the highest value to which other values can be subsumed” [10]. In this, values are considered as everything people consider important in life [11].

Well-being is gaining increasing interest in both health care and technology. In health care, health has long been regarded as the absence of disease or infirmity. Its formal definition was changed to well-being in 1948. In that year, the World Health Organization redefined health as “a state of complete physical, mental, and social wellbeing” [12]. In technology, well-being is often considered to be a central value in the design process. Many technologies aim to improve the well-being of their users. To that end, an increasing number of design methodologies exist that aim to guide design for well-being processes [10].

However, there is no clear explanation of what (values) constitute well-being, which leads to differences in understanding well-being and much uncertainty on how to enhance well-being through digital health. Philosopher Philip Brey [10] categorizes uncertainty by means of 4 problems that are paramount in technology for well-being. First, he describes the epistemological problem. This problem refers to the definition of well-being that should be embedded in technology design. This includes questions on how users consider their own well-being and how designers should obtain an understanding of the conceptions of users. The scope problem refers to questions on the who and when of considering well-being in design. A design can improve the well-being of its main user group but could also include needs from indirect users. In addition, the focus of improving well-being could be short term, long term, or both. The third problem, the specification problem, includes all the questions related to the embodiment of well-being in technology design: how should well-being be translated into design requirements, and how can one ensure that users will experience improved well-being when using technology? The fourth problem, the aggregation problem, refers to dealing with conflicts that arise between the contradictory values of well-being. Value conflicts can occur in one user (eg, when choices for optimal short-term well-being do not correspond to choices necessary for long-term well-being) or between users of similar or different groups (when an increase in the well-being of one user leads to a decrease in the well-being of another user). The 4 major questions in technology for well-being challenge the ability to create digital health for optimal well-being.

In this study, we conducted a systematic scoping review to facilitate future practitioners in the process of creating digital health for the well-being of patients. Digital health commonly follows two processes before adoption takes place: design and evaluation. Design is the process in which technology is created with the objective of solving a specific problem in health care. Evaluation is the process of (scientifically) assessing the added value of a technology to decide whether to adopt it as part of standard care. We will reflect on the current practices in both design and evaluation to find an answer to how to deal with the 4 problems paramount in technology for well-being.

The PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) checklist was followed to ensure the reliability of our results [13]. To meet the eligibility criteria, the included papers had to (1) address the design or evaluation of digital health, (2) report increased well-being as the purpose of digital health, (3) address the patient as the main user, and (4) be published in English. Digital health is mainly designed to improve the well-being of patients. We particularly focused on patients as the main users to locate our research in the context of health care. Age or other criteria were not attributed to the type of patient considered. However, we will reflect on the other users, such as caregivers, who are considered in the design and evaluation processes, in addition to the patient. All types of original peer-reviewed research papers were considered. Reviews were excluded from the search as they did not provide in-depth insights into the 4 guiding problems for each digital health case. To obtain a complete overview of all digital health papers, we did not include any inclusion restrictions on the publication period (see Textbox 1 for an overview of the eligibility criteria).

Papers were collected from six databases on February 24, 2021: ACM Digital Library, PubMed, Web of Science, IEEE Xplore, PhilPapers, and Google Scholar. We used a combination of the following terms appearing in the title or abstract of the papers: (wellbeing OR well-being) AND (patient*) AND (design OR moral* OR ethic*) AND (technology OR digital OR ehealth OR mhealth OR telemedicine OR telehealth OR electronic health OR mobile health OR mobile* OR smart* OR internet; Multimedia Appendix 1). The terms moral and ethic were added to the search strategy as these terms might result in insight into various values related to well-being. The term design in this section was used as a synonym for the term technology in digital health.

After collecting all papers from the databases, duplicates were removed. The titles and abstracts of the remaining papers were screened for eligibility criteria using Rayyan software [14] by 2 authors independently (MS and CK; Textbox 1). The full texts of the remaining papers were downloaded and independently screened for inclusion by the same authors. In both phases, the authors discussed disagreements until consensus was reached.

Data were extracted by 1 author (MS), who was guided by the 4 theoretical problems and their subquestions to understand if and how these problems are considered in design and evaluation practices (see Textbox 2 for subquestions for each problem directly derived from Brey [10]). The papers were divided into two groups based on their main focus: design and evaluation of digital health. Sections of each paper were marked when they were related to 1 of the 4 problems and their corresponding questions. All marked sections were grouped by question. Consequently, we compared the results of the various papers for each question to understand how each major problem was considered in practice. To determine what values constitute well-being in the epistemological problem, we followed the principles of thematic content analysis [15]. We coded each definition of well-being and categorized them into 7 overarching values. The results are presented for each problem.

A total of 1111 papers were identified. After removing duplicates and gray literature, 75.79% (842/1111) of papers remained. After title and abstract screening, 69% (581/842) of papers were discarded mainly as they did not study a digital health intervention or did not consider the patient as the main user of the technology. Of the remaining 261 papers, full-text screening resulted in the exclusion of 90 (34.5%) papers for not being original research, 40 (15.3%) papers for only reporting well-being but not considering it in the design or evaluation process, and 14 (5.4%) papers for either not considering patients as the main user or digital health as an intervention (see Figure 1, which was derived from the PRISMA [Preferred Reporting Items for Systematic Reviews and Meta-Analyses] 2020 flow diagram [16]).

Of the 261 papers, 117 (44.8%) papers remained for data extraction, of which 46 (39.3%) papers focused on the design of digital health, and 71 (60.7%) addressed its evaluation. Although some papers described both processes, it was possible to group them according to their main focus. The included papers were published between 1999 and 2021. Approximately 80.3% (94/117) of all the papers were written in or after 2015.

The results were presented for each of the 4 problems (see Table 1 and Table 2 for a summary of the general results and Multimedia Appendix 2 [17-133] for more detailed information).

To understand how the scope problem is considered in the design and evaluation of digital health, we studied the main users of the technology, the additional users, and the time frame considered.

The specification problem relates to the embodiment of well-being in the design of digital health. To achieve this, we extracted the type of technology, the design or evaluation method used, and the procedure for creating design requirements.

The aggregation problem refers to conflicts within the values of well-being. We aimed to understand if such value conflicts were reported and, if so, the type of conflicts that arose and the solution for solving these conflicts.

The enormous growth in papers since 2010 reflects that digital health is increasingly being applied to improve the well-being of patients. In addition, this type of technology is rapidly advancing. Although studies on digital health started with CDs and telephone consulting, today’s studies increasingly cover health care provision via web-based supporting platforms, social media, and remote sensing. In all of these cases, we identified how practice deals with the commonly discussed theoretical problems in technology for well-being. We defined common values related to well-being and identified that well-being was rarely defined by the users themselves. In addition, we identified that the current scope of well-being is generally small, involving few users in a short time frame. We illustrated that many methodologies exist on how to embed well-being in design, whereas only a few methods are accepted for evaluation. Finally, we identified that value conflicts, commonly discussed in theory [10,135], are rarely considered in practice. Our results show that simple solutions exist for solving theoretical problems that are paramount in technology for well-being. At the same time, this theory can challenge current practices for continuous improvement.

Within health care, health is considered a state of complete physical, mental, and social well-being. After the introduction of this definition in 1948, it was scarcely applied in practice [136,137] and regularly criticized for being unrealistic. Such criticisms mostly applied to the word complete, as this word implies that people cannot feel healthy without absolute physical, mental, and social well-being. In an aging society, complete well-being would then only be reserved for a few individuals [138]. To specify the novel definition of health as well-being, the concept of positive health was introduced by various authors [139-142]. Positive health focuses on health through well-being. It is the ability to flourish despite mental or physical diseases. In addition, as a response to assess health from the perspective of well-being, an entire range of validated questionnaires emerged, often termed QoL measurement scales [137]. Although used often in recent times, the terms of well-being, positive health, and QoL are rarely defined, which made Locker and Gibson [143] conclude that the commitment to these concepts is more rhetorical than real.

Our review sheds light on how well-being is conceptualized in the context of digital health provision. First, we found that, unfortunately, a large majority of papers only reported on well-being without actually implementing it, which corresponds to the rhetorical commitment to well-being postulated by Locker and Gibson [143]. In the design and evaluation papers that committed to well-being, we were able to identify seven values commonly considered as part of well-being: healthy body, functional me, healthy mind, happy me, social me, self-managing me, and external conditions. The often-occurring values of healthy body, healthy mind, and social me reflect the definition of health as physical, mental, and social well-being, as proposed by the World Health Organization. The value of happy me was found to greatly reflect the movement of positive health, aiming for health care to make a move from disease toward happiness and flourishing. Remarkably, we also commonly found self-managing me to be an important value of well-being related to digital health. Previously, this value was often not considered to be a part of patients’ well-being. Digital technologies have the ability to locate the center of health from hospital to home and from health care provider to patient. With this ability, they can increase the autonomy of patients. We believe that with the growing use of digital health, the importance given to the value of self-managing me will increase. As a consequence, it might be important to constantly reflect on the interactions among patients, technologies, and values of well-being to optimally design health care for patients’ well-being. The identified values can be used as a source of inspiration for digital health designers and evaluators to co-design technologies for optimal well-being with patients.

We identified more evaluation papers than design papers. This difference might result from design research not always reporting on its processes [144]. In addition, we found differences between the design and evaluation papers regarding the 4 problems. Design papers mostly considered well-being as healthy body and self-managing me. Evaluation papers often considered the values of healthy mind and happy me. Design papers mainly focused on mental disorders. Evaluation papers also studied mental disorders and neoplasms. Other users, such as caregivers and relatives, were regularly involved in design papers but not in evaluation papers. Design papers used numerous design methods, whereas evaluation papers mostly considered pre-post measurements or RCTs as a method. Value conflicts were rarely described in design papers and even less in evaluation papers.

Clearly, differences exist between the processes of design and evaluation, given that they take place in different stages of the digital health development process. Design processes commonly take place outside the care context. These processes are both creative and exploratory. In contrast, evaluation follows the rules of evidence-based health and is located in the health care environment [145]. Evaluation is more bound to standardization and reflection. The different characteristics of both processes result in the need for different methodologies for design and evaluation, as discussed in the specification problem. However, the differences seem to be larger than what can be explained by the stage of development. One would expect that the design and evaluation of digital health consider an equal approach to epistemological and scope problems by adopting the same definition of well-being, the same ways of obtaining this definition, and focusing on the same user groups. Nevertheless, this review shows different results. Design and evaluation vary greatly in their definitions of well-being and the main patient groups that are commonly designed for. This is troublesome, considering, for example, an application for postoperative patients to monitor their health. The design team creates this solution to improve the value of self-managing me for these patients. During the evaluation, this technology is assessed based on its ability to improve the healthy body and the healthy mind. The misalignment between the design input and evaluation outcomes results in suboptimal insight into the potential of the technology, which could hinder its successful implementation. Currently, only a minority of digital health technologies have achieved successful implementation. More digital health technologies would achieve successful implementation when the design is better aligned with evaluation and vice versa [146,147]. In particular, we argue that design and evaluation should consider similar definitions of well-being in the design and evaluation processes and the same user groups.

In most design and evaluation papers, well-being was defined without user input. When users were questioned, they were commonly asked to define only a specific value of well-being instead of explaining the values they considered important. Brey [10] previously noted that well-being could not be objectively determined independently from the user. Our results are worrisome, as the papers did not obtain a real understanding of what users considered to be important values belonging to their well-being. Involving users in defining the values of well-being as a source of design inspiration and evaluation outcome is recommended.

Methods for designing and evaluating digital health clearly differ. In most design papers, a participatory design was applied. However, as Orlowski and Matthews [148] argued, in addition to participatory design, a design method for structuring the process is needed. The papers in this review considered a wide variety of design methods to structure the design process. A review of the methods for usability testing of eHealth showed similar results [144]. No consistency seems to exist in the design papers on what design methods to use to design for well-being. The question of what design method to use to create well-being has been raised in the past [149]. Multiple papers have been written on the potential to embed the values of well-being in technology design. Design methods such as value-sensitive design [150] or values that matter [151] have been introduced. In addition, numerous papers have aimed at creating design for well-being methodologies to provide designers with a framework for embodying well-being in design. Examples in the literature are the approaches of emotional design [152]; life-based design [153]; capability-sensitive design [154]; positive design [155]; motivation, engagement, and thriving in user experience model [156]; positive technology [157]; experience design [158]; and positive computing [159]. Given this myriad of design methodologies that particularly aim at our goal of designing for well-being, it is remarkable that none of these methods was found to be used in the design papers. We speculate that this resulted from not knowing about the design methods, finding the methods difficult to apply to health care, or valuing well-known methods over novel ones. Future digital health designers would benefit from heuristics on which method to use in what situation. This can be facilitated by increasing awareness of design methods and transparency and reporting on the reasons for using a certain design method.

The interaction between users and technology and the context of use might affect how well-being is expressed, which does not necessarily correspond to the initial embodiment of well-being in design. This is called the positivist problem [160]. For that reason, an evaluation process is necessary to study the actual effects of technology on well-being. Evaluation papers commonly considered the same set of evaluation methodologies: pre-post measurements and RCTs.

The RCT methodology, which has become the gold standard for effectiveness studies in pharmacological interventions, has been transferred one to one to evaluate the effectiveness of nonpharmacological interventions, including digital health [147]. An RCT only provides valuable information on an intervention’s effectiveness when (1) the intervention and the way of providing the intervention are stable, (2) the intervention can be applied with fidelity, and (3) when it is expected that the outcomes of the intervention are measurable and meaningful [161]. Given these requisites, the use of RCTs in digital health has been subject to several concerns. An RCT requires that the intervention is stable, implying that the digital health technology has been finalized before the start of the RCT. However, design is an iterative process that requires numerous phases of testing and adaptation. Clinical outcomes obtained through an RCT are often required to justify the use of more resources to improve the design. However, when the design is improved based on the clinical outcomes of RCTs, these outcomes become directly outdated. Similarly, when a digital health solution needs to await clinical outcome measures for improvement, the technology might be outdated once the trial is finished [162]. In addition, a digital health solution cannot be directly applied with high fidelity as the solution mediates clinical processes and might require changes to the context of use to be optimally used. However, these effects require further study. The limited study design of an RCT does not permit the study of such mediating effects, and without knowing the optimal context of use, digital health cannot be applied with high fidelity [163].

Several changes and alternatives to RCTs have been proposed to align scientific evaluations with digital health. These include, among others, a multiphase optimization strategy for the RCT, allowing the design to be adapted during the evaluation process [164], evaluating the principle of a solution rather than the specific technology itself [162], and broadening the set of outcome measures into the inclusion of human–technology interactions [165] and legal and ethical evaluations [166,167]. Another interesting alternative is a single case experimental design [168]. This method illustrates that RCTs only provide an average good and not the optimal solution for each individual. They propose to observe a single case over a longer period while manipulating treatment (technology). This study design allows personalizing digital health to an individual’s needs, thereby increasing its efficacy. So far, the papers in this review have barely considered moving beyond traditional evaluation methods. Our results resemble previous results from systematic reviews of evaluation methods and health technology assessments of digital health [169,170]. There is a need for the use of evaluation methods that are better aligned with the complexity of digital health.

In addition, not only RCTs but also the many validated QoL questionnaires to assess the effect of digital health via an RCT need reflection. In the past, these questionnaires were studied thoroughly [137,171]. Guyatt et al [171] rightfully posed the concern of how to select, use, and interpret such questionnaires. We share their concerns and believe that when applying such a questionnaire, it is important to understand and illustrate what values of well-being are measured through the questionnaires and to be aware of the potential and limitations of what is being asked. Similarly, as Blandford et al [147] argued, such questionnaires only have a limited face and construct validity. They do not provide insight into how the social structures of care change and how technology is changing the current values of well-being [172]. In addition, QoL questionnaires were validated for a specific user group in a specific context. When technology is introduced, the entire basis on which such questionnaires are validated might change. Thus, the use of QoL questionnaires requires careful consideration of their selection and application. Our 7 values provide a comprehensive and up-to-date view of well-being. Further work could be done to develop tools (eg, questionnaires) based on the 7 values to complement and support the current evaluation methods to better align with the complexity of digital health.

On the basis of the theoretical and practical reflections of this review, we sketched an ideal practice for anyone involved in the design and evaluation of digital health for well-being. This ideal practice begins with the formation of multidisciplinary teams working together from the start of the design process to successful implementation. After the composition of the team, we advise demarcating the scope of the project. Which users will be affected by the technology? What users will be studied, and who will be left out of the study’s scope? Furthermore, we encourage the upfront identification of the moments in time during which the effect of technology is evaluated. The process should continue with a clear study on the definition of well-being per user group in which users are closely involved. To prevent well-being from becoming a buzzword, the team should consider constantly aligning the following design and evaluation processes to the found values of well-being. Instead of two linear processes, design and evaluation take place simultaneously. Evaluation outcomes are the source of design input and vice versa. As many methods exist for designing for well-being, the team should decide together which method is most suitable. The reasoning process is reported to facilitate other teams to make their own decisions. To justify the complexity of digital health, the type of evaluation method is carefully considered. The chosen method enables the evaluation of the user–technology–value interaction in an authentic context of use. In addition, the method facilitates obtaining insight into individual experiences for the personalization of the solution.

In this paper, we aimed to bridge the gaps between the practices of design and evaluation and the theory produced on digital health technologies aimed at improving well-being. This endeavor is both a strength and a limitation. By highlighting the differences among the fields, we enable the design to consider the context of evaluation and vice versa, inspire practice to consider better theoretical insights, and guide theorists to acknowledge the need for pragmatic decision-making in practice. The application of our recommendations would result in individual, social, and economic benefits. First, at the individual level, digital health would foster a culture of inclusiveness through personalization. It would better meet the needs and values of individual patients, caregivers, and relatives, regardless of their education, age, gender, or culture. Second, on a social level, digital health would fit within current health care practices and align with the care processes of health care personnel. Third, economic benefits will arise from the alignment of design with evaluation by preventing a waste of resources and leading to a more successful uptake of digital health.

A limitation of our study is that it was difficult to report insights from all disciplines using a common language. The attempt to place all insights from practice into theoretical frameworks might have resulted in missing important items or mistakenly interpreting certain practical phenomena as belonging to certain theoretical concepts. For example, we found that a great majority of design papers aimed to create digital health technologies for lifestyle management of the older population, regardless of their medical condition. By categorizing all design papers into the International Classification of Diseases for Mortality and Morbidity framework, these older adults were divided into three different categories (mental disorder, disease of the nervous system, and symptoms not classified elsewhere). Although design papers often considered lifestyle management and older adults, this insight was invisible to those looking at the graph only.

A second limitation is that we aimed to include a wide variety of digital health technologies, although our results mostly identified digital health solutions in the domain of support platforms via apps or the internet. In addition, almost all solutions were designed for use at home. This is remarkable, as our search strategy did not focus solely on the home. Obviously, the terms telemedicine and telehealth refer to technologies for use at home. Nonetheless, we also included the term digital in our search strategy, which we expected would result in papers on digital technologies used in hospitals and primary care settings. Our search strategy had missing terms that would result in papers on technology use in hospitals, technology use in hospitals is not yet common, or it is not discussed in relation to well-being. Although we believe that our insights can also be applied to a wider range of digital health technologies in varying contexts, in the future, it is interesting to expand our research to varying technologies and extend it to the inpatient care context to understand the differences in designing technologies in diverging contexts. Furthermore, as we only included papers focused on well-being improvement, we might have missed the results on in-person value conflicts that are reported in papers studying the harmful effects of digital health on well-being. Finally, we rarely found sets of papers that focused on the design and evaluation of the same technology. The ability to compare the design and evaluation processes of similar technologies would have led to more reliable results. In the future, better aligning design and evaluation would ideally also result in more consistent reporting.

In this review, we have shown how current practices deal with the major problems that are paramount in digital health for well-being. We identified major gaps in design and evaluation regarding their conceptualization of well-being, types of users studied, and methods used to design and assess well-being. The comparison of empirical practice with theoretical frameworks also showed how both fields have found pragmatic ways of circumventing or dealing with the problems of digital health for well-being. By illuminating the differences between design and evaluation, as well as practice and theory, and providing recommendations for best practice, we expect to have set the first steps to slightly bridge some gaps. As digital health technologies are gaining an increasingly important role in the future, we believe that multidisciplinary collaborations are required to be improved by moving beyond the common dichotomy of design and evaluation. Only then it is possible to transcend from digital health toward digital well-being.