Digital Patient Experience: Umbrella Systematic Review

Introduction

Background

Recently, there has been a significant increase in the use of digital health technologies. In addition, many countries currently use digital health technologies to support health care service delivery to overcome the disruptions caused by the COVID-19 pandemic. These include web-based patient consultations and requesting pharmacy and medication refills [1]. Digital health offers care without the risk of exposure to the virus, especially for vulnerable patients such as older adults and patients with chronic diseases [2]. Before the COVID-19 pandemic, there was increasing recognition of the potential of digital health to improve the accessibility of health care in different clinical settings (eg, ambulatory care, acute care, and inpatient care) [3]. Digital health provides an opportunity to both reduce the costs of care and improve patient affordability [4,5], and previous research suggests that digital health has the potential to provide health prevention, consultation, treatment, and management [5-10]. With digital health solutions continuing to grow in both number and functionality, patient interest in digital health has rapidly increased, leading to an expanding reliance on digital health technologies [11].

As digital health has become a more familiar term, it has generated many definitions, and the concept has been expanded to encompass a much broader set of scientific concepts and technologies [12]. These include digital health applications, ecosystems and platforms [13], patient portals [14], mobile health apps [15], eHealth records, and appointment scheduling applications [16]. For the purposes of this study, we will use eHealth, mobile health, telemedicine, telehealth, virtual health, remote health, electronic consultations, and health information systems (HISs) as interchangeable terms for digital health.

Patient Experience in Digital Health

Digital health has the potential to improve patients’ overall health care experience [17-19]. However, there is currently no common concept for describing patient experience (PEx) in digital health. Neither the general PEx nor user experience (UX) adequately reflects the experience of a patient using a digital service. For example, in a hospital setting, the environment’s cleanliness, background noise, and even food provision could affect PEx [20]; however, these factors would not be expected to influence the experience of a patient using a digital service. Similarly, the fact that the system passes usability heuristics does not necessarily mean that the overall experience of a patient using digital health services is positive [21]. Therefore, it is vital to understand the experiences of individuals using digital health and how the design of new technologies can affect them [17,22,23].

The concept of (nondigital) PEx has many definitions in general health care practice and research. The Beryl Institute defines PEx as “the sum of all interactions, shaped by an organization’s culture, that influence patient perceptions, across the continuum of care” [24]. Other definitions and studies note that the core elements of optimized PEx include access to appropriate care, patients’ active participation in care, a good patient-physician relationship, reliable evidence-based care, comprehensible information, physical comfort, emotional support, involvement of family and friends, individualized approaches, responsiveness of services, and continuity of care [19,25-27]. These core elements of PEx help to recognize patients’ priorities when receiving care and in providing patient-centered care. However, patients’ priorities may differ for digital health, in which traditional face-to-face interaction is replaced by human to digital interface interaction. Therefore, to address patient priorities in digital health, it is essential to consider UX in the design of digital health [28]. In this study, we define UX as a person’s perceptions and responses that result from the use or anticipated use of a product, system, or service [18,29]. Usable, useful, findable, accessible, credible, valuable, and desirable products are more likely to succeed in delivering a positive UX [30]. However, the full impact of digital health technologies on PEx or UX still remains unclear [31]; some products even result in negative effects such as increased patient anxiety [32]. Therefore, more insights into the barriers to and facilitators of individuals’ experiences with digital health are required [33].

Objectives

The objectives of this paper were to systematically review (1) the factors that influence PEx in digital health and (2) the design considerations of PEx that are in digital health. The overall aim was to generate a design framework and guidelines for further improving PEx in digital health.

Methods

We performed an umbrella systematic review compiling evidence from multiple systematic reviews [34] on PEx and UX in digital health. This review was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodology, which is an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses [35].

Digital PEx Working Definition

Throughout this study, we use the term digital PEx as a working definition to describe people’s experiences in various digital health contexts. As the study progressed, the definition underwent several revisions, which resulted in a more inclusive final definition.

Search Strategy

We searched Scopus, PubMed, and Web of Science for studies published between January 1, 2000, and December 16, 2020. The search time window was limited to 2000 as the term digital health was first introduced by Frank [36] in 2000. To be inclusive, we used broad interchangeable search terms with varying combinations of digital health, PEx, and UX:

1. Category 1: “patient experience” OR “health experience” OR “user experience” OR “customer experience” OR “client experience”
2. Category 2: “ehealth” OR “e-health” OR “mhealth” OR “m-health” OR “telehealth” OR “tele-health” OR “digital health” OR “virtual health” OR “remote health” OR “telemedicine” OR “telemonitoring” OR “teleconsultation”
3. Category 3: “patient digital experience” OR “patient experience in digital health” OR “e-patient experience” OR “epatient experience” OR “online patient experience”

After combining categories 1, 2, and 3, limits were set to restrict studies to English-language literature reviews published in journals after 2000. The final search strategy was ([category 1 AND category 2] OR category 3) AND (DOCTYPE [review]) AND (PUBYEAR>2000) AND (LIMIT-TO [SRCTYPE, “journal”]) AND (LIMIT-TO [LANGUAGE, “English”]). Google Scholar was used as an additional database to manually search for additional related references based on the snowballing method during the review process.

Selection Criteria

Eligibility criteria were developed for title and abstract screening and refined for full-text screening. The following inclusion criteria were proposed by TW and GG and adjusted by MM and RG:

1. No duplicated articles
2. Full text available
3. English language
4. Only completed peer-reviewed journal articles
5. Only review articles
6. Related to digital health (ie, use of information and communication technology in health) and PEx, UX, or health care experience

Screening Process

The collected articles were included in the final analysis if they met all the inclusion criteria after a 2-stage screening process: first, a title and abstract review, followed by a full-text review. In the screening process, 2-round, small random samples (20%) were independently reviewed by 2 reviewers (TW and GG) who evaluated the eligibility of the articles against the selection criteria. The interrater reliability and clarity of the selection criteria were assessed using the Fleiss-Cohen coefficient until it reached the required strength (≥0.60). Uncertainties around paper inclusion and exclusion were resolved by discussions with the research team (TW, GG, MM, and RG) when necessary.

Data Extraction and Thematic Analysis

Articles meeting the eligibility criteria were imported into ATLAS.ti (Scientific Software Development GmbH; version 9.0.7; 1857) for data extraction. Data were extracted for the following aspects: (1) study characteristics, including authors, year of publication, research aims, review methods, target users, and digital health intervention (DHI) characteristics; (2) the overall impression of digital PEx (eg, the foci or types of findings regarding digital PEx); (3) influencing factors of digital PEx; and (4) design considerations for improving digital PEx.

We used the Braun and Clarke 6-phase thematic analysis method [37] to analyze the extracted data; these include (1) familiarization with the data, (2) generation of initial codes, (3) searching for themes among codes, (4) reviewing themes, (5) defining and naming themes, and (6) producing the final report (analytical themes). A total of 4 researchers participated in the review process. After data familiarization, a set of a priori categories was defined by TW and refined by all the coauthors (Table 1). The coding was based on the Performance of Routine Information System Management (PRISM) framework [38], which states that routine HIS performance is affected by the system’s inputs (ie, technical, behavioral [39], and organizational determinants) and progress. Please note that other elements of the framework (outputs, outcomes, and impact) are discussed in another study addressing the evaluation of digital PEx (work in progress).

Group discussions among the authors were used to reach an agreement on the produced a priori categories. TW quoted the relevant data across the included reviews, generated initial codes based on a priori categories, and then searched for themes among codes. Frequently used terms in the included reviews were used as inspiration to generate subsequent codes and themes. The latter process was independently and randomly validated by GG, MM, and RG.

Results

Overview

Figure 1 shows the flow diagram of the systematic search. A total of 173 records were generated after the computer search; 58 (33.5%) duplicates were removed, and the titles and abstracts of 115 (66.5%) articles were reviewed. Subsequently, 53.9% (62/115) of full-text articles (including 4 additional records collected through snowballing) were reviewed for inclusion. Ultimately, 45 studies were included in the review for data extraction.

‎View this figure

Study Characteristics

Embase, MEDLINE, PubMed, PsycINFO, CINAHL, and the Cochrane Library were the most common databases for the included reviews. Of these, 62% (28/45) were systematic review articles. The remainder included scoping reviews (6/45, 13%), literature reviews (3/45, 7%), integrative reviews (3/45, 7%), narrative reviews (2/45, 4%), comprehensive overviews (1/45, 2%), review of systematic reviews (1/45, 2%), and umbrella reviews (1/45, 2%). More than half of the included reviews (24/45, 53%) conducted quality assessments. The reviews included >1400 studies, which mainly or partially reported qualitative and quantitative analyses of PEx in digital health. The data analysis methods varied and included thematic analysis (8/45, 18%), meta-synthesis (5/45, 11%), meta-ethnography synthesis (2/45, 4%), taxonomy (1/45, 2%), hermeneutic synthesis (1/45, 2%), qualitative evidence synthesis (1/45, 2%), and state-of-the-art survey analysis (1/45, 2%).

Among the included reviews, some focused on specific populations, such as children (3/45, 7%), college students (1/45, 2%), younger people (1/45, 2%), adults (7/45, 16%), or older adults (4/45, 9%). Others either focused on the general population or did not mention the target population. The most common health issues across the included articles were chronic diseases (17/45, 38%), including chronic obstructive pulmonary disease, heart failure, cardiovascular disease, cancer, diabetes, and hypertension. Mental health problems (7/45, 16%), including depression, anxiety, psychological well-being, psychotic disorders, and schizophrenia, were the second most common health issues. The remainder either focused on other issues (8/45, 18%), including audiology, asthma, reproductive health, maternal health, newborn health, child health, adolescent health, surgery, postpartum, somatic diseases, or palliative care, or did not mention any specific health issues (14/45, 31%). Some papers (8/45, 18%) also provided multistakeholder perspectives, including health care professionals, providers, surgeons, clinicians, staff and organizations, implementers (such as health policy makers, clinicians, and researchers), and the participation of information technology.

The degree of detail provided about the interventions varied greatly across the studies. Phone-based apps, websites, handheld sensing devices, and ambient assisted living health care systems were common digital health deliveries. Interaction techniques included synchronous, asynchronous, and hybrid models. Diverse intervention platforms, systems, or functions were used to deliver various health care services, including supporting disease management (14/45, 31%); patient-to-physician communication or consultation (9/45, 20%); symptom monitoring (9/45, 20%); information transmission (4/45, 9%); health promotion activities (3/45, 7%); screening, diagnosis, or self-assessment (2/45, 4%); behavior changes (2/45, 4%); self-education (1/45, 2%); and decision-making (1/45, 2%). Multimedia Appendix 1 [28,40-83] provides detailed information regarding the characteristics of the included studies.

Overall Impression of Digital PEx

Our study revealed great diversity in the perspectives and definitions describing patients’ experiences and characteristics when using digital health, presenting a variety of influencing factors and design considerations for digital PEx. The included studies showed different foci regarding digital PEx, including influencing factors (21/45, 47%) [28,40-59], digital health performance (19/45, 42%) [40-43,46,48,49,56,57,59-68], patient perceptions (9/45, 20%) [28,45,47,49,69-73], evaluation methods of digital health or digital PEx (8/45, 18%) [43,64,74-79], and design considerations (9/45, 20%) [48-50,53,54,59,80-82]. The findings and conclusions of the 45 reviews showed a great diversity. Most studies presented a set of themes (19/45, 42%) [28,44,45,48,49,51,54-57,59,62,69-73,76,79] or descriptions only (16/45, 36%) [40-42,46,47,58,60,61,63-67,74,75,83]. Other studies concluded with a theory-based description (5/45, 11%) [52,68,77,78,80], framework (4/45, 9%) [28,49,50,82], model (2/45, 4%) [53,69], method (2/45, 4%) [43,81], or checklist (1/45, 2%) [59]. Only a few studies transformed findings into design considerations (9/45, 20%) or visualized or structured their results into frameworks, models, checklists, or methods (9/45, 20%). Limited information was found on participant dropout reasons during the interventions [28,41,43,51,53,63,69,71]. The overall impression of the researchers on the DHIs was positive. In 51% (23/45) of reviews [41-44,48,49,52-54,57,59,61-63,65-69,72,73,77,80], the DHIs either showed promising results or at least results comparable with face-to-face health care services. Only 4% (2/45) of reviews [47,60] reported concrete evidence of the negative impact of current DHIs on digital PEx. In general, digital PEx was addressed because of the interactions between the DHIs and the patients involved and how the service was organized and carried out.

Influencing Factors of Digital PEx

An influencing factor is an aspect of the existing situation that influences other aspects of the situation, and it is formulated as an attribute of an element that is considered relevant and can be observed, measured, or assessed [84]. In this study, influencing factors refer to specific factors that lead to a positive or negative experience (digital PEx). Some factors have either positive or negative consistent and concrete impacts, whereas others have double-edged impacts; that is, impacts that are different per individual or change over time. Among the included papers, a common understanding of the potential influencing factors was captured from 3 aspects—behavioral, technical, and organizational determinants—following the categorization of the PRISM framework. These determinants were each classified into 3 categories, resulting in nine categories: patient capability, patient opportunity, patient motivation, intervention technology, intervention functionality, intervention interaction design, organizational environment, physical environment, and social environment. Multimedia Appendix 2 [28,40-83] presents an overview of the themes identified for each category, the influencing factors per theme (positive, negative, and double-edged), and references. Most factors appear to be related to technical determinants, followed by behavioral and organizational determinants. For technical determinants, we summarized 3 categories with 13 themes, containing 58 positive, 35 negative, and 13 double-edged factors. For example, DHIs with multiple behavioral change techniques appeared to be more effective [42,56,80] and reported higher patient satisfaction [54,57]. Behavioral determinants included 3 categories with 9 themes containing 11 positive, 21 negative, and 5 double-edged factors. For instance, some studies mentioned a lack of confidence in patients’ own ability to use the technology [43,45,47,48,71,81], leading to a negative digital PEx. Organizational determinants were classified into 3 categories with 5 themes, including 13 positive and 23 negative factors. For example, unrealistic financial reimbursement and higher costs related to the internet or equipment were practical challenges of using digital health [47,48,51,55,56]. For the behavioral and organizational determinants, we collected more negative factors than positive factors. This is in contrast to the technical determinants, in which more positive factors were identified. Double-edged factors were less than both positive and negative factors for all the 3 determinants. Multimedia Appendix 3 [28,40-61,63-67,69-73,75-78,80-83] provides detailed information and examples.

Design Considerations of Digital PEx

Table 2 provides an overview of the identified themes for each design construct or method, related considerations, and references. To address the abovementioned influencing factors, several the included articles referred to design constructs (personalization, information, navigation, and visualization) [48,49,53,54,59,80] and design methods (ie, human-centered design [HCD] or user-centered design [UCD], co-design or participatory design, and inclusive design) [48-50,54,80-82], either as recommendations or implications for improving digital PEx from a design perspective. Notably, there was an overlap between design considerations and influencing factors. The former focuses on concluding possible design suggestions, recommendations, and implications proposed by the reviewed articles. The latter involves mapping the impacts of interaction design on digital PEx in different contexts; therefore, they refer to different themes and references. Generally, the personalization construct identifies patient profiles and tailors digital health according to patients’ needs and preferences. The information construct addresses the source, language, presentation, content, and architecture of delivered health information. The navigation construct considers the interactive, delivered, and instructional elements of digital health to guide users to different areas of content within digital health. The visualization construct focuses on the aesthetics, attractiveness, visibility, and consistency of digital health appearance and interface. Furthermore, co-design and UCD or HCD were recommended as the most common methods for designing digital health, which involve multi-stakeholders and multi-disciplinaries in the design process to facilitate the designers’ work, as designers need to understand end user needs and be aware of potential barriers to engaging in DHIs. Finally, inclusive design provides flexible design and is usable for a broader population. Notably, the design considerations identified in the included papers are not meant to be applied to every project; the implementation depends on the project’s focus. Designers always need to balance project requirements (such as profits), user needs (such as privacy concerns), and policy regulations (such as data security). For example, peer-to-peer patient communication may not be appropriate for more sensitive health issues.

Design Implications

On the basis of our findings regarding influencing factors and design considerations for digital PEx, in this section, we define digital PEx and present design guidelines for the implementation of improving PEx in digital services.

Definition of Digital PEx

Our review reveals the absence of a commonly used concept for PEx in digital health. An increasing number of studies have been conducted on surveying PEx, satisfaction with, and expectations in varied digital health. With the growing academic interest in this topic and increasing efforts to address PEx in digital health design practice, a common concept with a concise definition will strengthen and align efforts overall. After reviewing the alignment of widely accepted concepts of PEx, UX, and DHIs with our generated influencing factors, we observed that many of our findings are included in the PEx definition offered by The Beryl Institute. Therefore, by including the sum of all interactions shaped by an organization’s culture, which influence patient perceptions across the continuum of care [33] along with the constructs of UX (people’s perceptions and responses [18,29]), DHIs (digital health technologies [13]), and the determinants (ie, technical, behavioral, and organizational determinants) identified in this review, we propose a concise, practical definition of digital PEx to guide the future design of digital health: “Digital patient experience is the sum of all interactions, affected by a patient’s behavioral determinants, framed by digital technologies, and shaped by organizational culture, that influence patient perceptions across the continuum of care channeling digital health.” Compared with the original definition of general PEx, this new definition underlines the digital part of health care delivery and includes 2 new determinants (technical and behavioral) that go beyond the organization’s culture to clarify what can influence patient perceptions while traveling along a digital care pathway.

Design Guidelines for Improving Digital PEx

We developed a design and evaluation framework to help digital health designers or developers improve digital PEx in the design process (Figure 2). This framework was based on the findings of this umbrella review and was inspired by the double diamond model [85,86]. Our framework shows four phases: define design, define evaluation, design ideation, and designevaluation. The first and third phases focus on the design itself, and the second and fourth phases focus on design evaluation. In this study, we focus on explaining the first and third phases. In the first phase, designers must define the design goals by considering the factors that affect digital PEx. In this phase, we provided 3 determinants referring to 9 categories of influencing factors that have 3 types of impact on digital PEx (positive, negative, and double-edged) for designers to discover and explore. Designers can frame their design goals based on the intervention purposes and the selection of influencing factors. For example, if the purpose of the intervention is to improve patient eHealth literacy, designers need to pay more attention to patient capability and frame a design goal to develop suitable intervention functionality for improving patient capability. After defining the design goals, designers can move to the second phase, which is the define evaluation phase. In this phase, designers need to consider evaluation indicators (patient emotional, behavioral, and health outcomes) and evaluation methods (surveys and interviews) that are used to assess digital PEx. Detailed information regarding this phase will be discussed in a parallel study. Following this, we provide 4 design constructs (personalization, information, navigation, and visualization) and 3 design methods (ie, HCD or UCD, co-design, and inclusive design) for the design ideation phase. Personalization [41,54,56,57,59,69,81,87] refers to ascertaining user needs with design goals. It encompasses the design of intervention technology and functionality needs that meet the patients’ ability, opportunity, and motivation to trigger behavior changes and promote health outcomes. UCD/HCD and inclusive design are valuable at this stage for the inclusion of patient perspectives. Driven by user needs and intervention goals, information includes content, communication, and functionality [54,59,81], and navigation comprises forms of delivery, user flows, instructions, and tutorials [54,59,80,81]. This relates to how relevant content presented in multimedia with a clear information architecture can attract patient attention and help them understand and complete tasks efficiently [88]. Co-design and participatory design are multidisciplinary collaborations that are necessary at these 2 stages. Finally, designers need to consider visualization [54,57,59,80,81], which determines the product look. The digital health interface can affect patients’ first impressions when using DHIs. An attention-grabbing, simple, and consistent interface [59], layout (colors and images) [80], and message presentation [59] can all lead to positive UX. The design guidelines (Textbox 1) can be used at this stage to produce design concepts. In addition, this phase contains the digital health design workflow, challenges, and tips from a design practice perspective (which will be presented in an ongoing interview study). Finally, we ended up with this framework by introducing the design evaluation phase, in which designers need to develop tests (based on evaluation metrics) to evaluate design concepts. If the evaluation outcomes do not meet the evaluation standards, designers can return to the design ideation phase to adjust the design concepts or return to the first phase to reconsider the design goals.

Compared with the original double diamond model, our framework separates the evaluation part from the design part. This aligns with the design research methodology framework [84], which suggests generating success criteria after clarifying design research goals and before producing design support, formulating criteria for success is essential to be able to determine whether the results help achieve this aim. Therefore, we paid equal attention to design and evaluation. In addition, our framework provides detailed reference materials (such as 3 determinants) for each phase to provide designers with more practical support. Notably, in our framework, we retain some typical features of the double diamond model: the first 2 phases are research related, the last 2 phases are practice related, and each phase starts from divergence and ends at convergence.

On the basis of our findings on influencing factors and design considerations, we mapped the combinations of design constructs and design methods into 9 design guidelines to address different influencing factors (Textbox 1), which can be used to guide the design ideation process. Some of the design guidelines uncovered in this study have already been implemented, resulting in a positive digital PEx, such as the digital platform PatientsLikeMe, which aims to empower patients to navigate their health journeys together through peer support, personalized health insights, tailored digital health services, and patient-friendly clinical education [89]. One of the studies pointed out that patients can greatly benefit from using this platform as it improves patient health literacy, and its condition-specific customization may still further improve PEx [90], which aligns with our design guidelines on improving “patient capability” and providing “personalized information.”

‎View this figure

Discussion

Principal Findings

We systematically reviewed review articles on factors that influence digital PEx and considerations regarding how best to design digital PEx. The reviews varied greatly in type, including studies and data analysis methods, as well as in HIS, health issues, target patient groups, intervention content, and structure. Of the selected reviews, 62% (28/45) were systematic reviews, the rest were other types. These included qualitative, quantitative, and mixed methods studies. Thematic analysis and meta-analysis were the most common data analysis methods used in the reviews. We note that the studies described in the selected reviews were extremely heterogeneous, and information about interventions and digital PEx were often mixed and complex, making comparison difficult.

Our results are in line with the findings reported by previous authors [25,30,38] on the factors that affect PEx, UX, or the implementation of digital health. On the basis of the identified influencing factors and design considerations, we developed 9 design guidelines for improving digital PEx. Our findings reveal that among the selected reviews, only a few formulated design strategies or guidelines. This lack of design knowledge transformation makes it difficult for designers or developers to apply the findings directly. This aligns with the studies by Sakaguchi-Tang et al [48] and Søgaard Neilsen and Wilson [80]; the former indicated that the absence of specific design recommendations impairs the design of digital health, with the latter suggesting that there was a lack of understanding of the most beneficial design aspects for some specific digital health and how design principles can best be applied. Moreover, the use of UCD has been recommended in many studies to address UX-relevant issues in digital health [3,80,91], which also supports our findings.

Digital PEx Versus General PEx and UX

We found a lack of a common term to describe PEx in digital health; UX (25/45, 56%) and PEx (17/45, 38%) were the most commonly used terms. Patient UX, patient perceptions, client experiences, patient empowerment, and user engagement were also used to describe similar concepts. Many reviews indicated that there was limited information about UX or PEx in varied digital health and underlined the need for a more holistic view of patient needs and priorities to better shape digital health design strategies and provide tailored digital health [28,40,42,45,60,74].

Influencing Factors Are More Complex Than Facilitators and Barriers

The information provided about digital PEx–influencing factors was complex and heterogeneous. Digital health is often treated as a whole, whereas digital PEx is affected by the additive effect of varying digital health factors. A single change in a factor may affect everything else. We found that without a concrete interaction context, factors could be regarded concurrently with facilitators or barriers. For example, regular contact with health care providers (HCPs) could be perceived to increase a sense of reassurance or perceived as a burden to patients’ daily lives [28]; some patients experienced digital health as time consuming or an additional burden, whereas others experienced it as time saving or convenient [69]. Some influencing factors may have a soft or indirect influence on digital PEx [44,76]. For instance, users who are completely unaware of privacy or security risks may have excellent experience with digital health that fails to meet privacy or security requirements [76]. A lack of concrete solutions to address these barriers was mentioned [48]. It is likely that digital health cannot serve all populations equally well [71], which aligns with the results of a scoping review that investigated the inequities caused by the adoption of digital health technologies [92]. Some researchers indicated that older adults can also experience benefits by using digital health [53], whereas others suggested that telehealth is, at best, a partial solution for younger and fitter subpopulations [47,71]. Again, although some mentioned that patients preferred using personal devices [49,55,61,64], others noted the opposite [64].

Unclear Benefit From the Different Elements in Digital Health

It is likely that some patient groups benefit more than others from specific DHIs. For example, one of the reviews suggested that in telemedicine treatment for type 2 diabetes, behavioral change and continuous management were the keys to success [43]. However, it was unclear precisely which elements of digital health resulted in patients’ satisfaction or dissatisfaction and how they could be addressed [83]. Moreover, we found limited data and even contradictory results on which factors affect digital PEx the most, which elements should be considered first when developing DHIs, and who benefits more from them. The latter is commonly mentioned [48,53,70], with some authors suggesting that patients with unstable chronic diseases might benefit the most [47,93]. However, another review indicated that even if patients are provided with the latest state-of-the-art technology at home, the intervention will not be beneficial if it remains unused [43]. Patients who are less activated are likely to have less positive experiences than those who are highly engaged [74].

Lack of Multiple Perspectives During the Design of Digital PEx

Clear communication between experts, designers, and patients regarding their understanding of digital PEx is required. Some reviews acknowledged the need for a multistakeholder perspective on digital PEx [55,69]. However, we found circumstances in which this was not possible. For example, in some cases, UCD for DHIs was conducted on nonpatient users either because of ethical reasons or relevant regulations [43], and in others, apps that are not specifically designed for patients with cancer were being used for this patient group [64]. HCPs are often isolated from the decision-making process to incorporate digital health into their current service provision [28]. Moreover, a lack of clinician perceptions of digital health use was also reported [40,54]. Furthermore, no studies focused on exploring designers’ views, opinions, experiences, or values in addressing PEx or UX in the design of digital health. There was little information on whether experienced designers had worked with patients in their design process.

Over- or Underestimated Results

Some studies suggested that a lack of interest was the main reason for patients’ refusal of digital health and that reasons for patient withdrawal were patients not wanting to use equipment, deteriorating health, and technical problems [94,95]. We need to gain better insights into the reasons for patients choosing not to engage in or withdraw from digital health, as these will significantly inform future DHI development and design [43,53,69]. However, it is likely that most studies only included patients who had already agreed to or were using digital health technologies; those who refused to use, withdrew from, or had no accessibility were excluded [28,51,63,69]. One of the reviews suggested that this would result in over- or underestimated results of DHIs’ effects on digital PEx, as participants who completed the intervention may differ from those who did not [41]. Another review found that patients only reported positive themes associated with remote monitoring, which may indicate a selection bias [71].

Conflicts Between Benefits and Cost for Developing DHIs

The provision of digital health can reduce the treatment burden and better integrate care into patients’ daily routines [69], which is consistent with our findings; we found that most reviews had a positive perspective of DHIs. However, in one of the reviews, it was suggested that although there was agreement among most professionals that health information technology can have a positive impact on PEx, when weighing the benefits against the potential cost, demonstrating this will be challenging [44,47]. Moreover, unnecessary high-frequency monitoring could result in a waste of health resources and an increased workload for HCPs [52]. Compared with existing health care services, the application of new technology needs to demonstrate clinical evidence of improved health conditions [43]. However, there were discordant findings in terms of the benefits of using DHIs. For example, there was no concrete evidence that telemedicine consultations were quicker than face-to-face consultations [40,57,68,83]. In another case, the impact of DHIs on health care use was not examined [57]. In conclusion, only user-friendly and quality-certified DHIs should be provided to patients [64]; health care organizations should not shift their focus from the basic and inexpensive strategies that affect patient care. Care is needed: new technology should not overwhelm the patient or ignore patient needs [44].

Limitations

First, when undertaking a review of review articles, some important details included in the original studies may have been lost, which increases the possibility of reporting bias. We also noted differences in the interpretation of terms and methods between the reviews. There is a lack of consistency in the terminology used to describe the functions of DHIs, HISs, or digital PEx itself. For example, in some cases, “eHealth” and “mHealth” were used as interchangeable terms [75], “persuasive technology” and “behavior change techniques” were presented as having a similar meaning [43], and “patient engagement” and “patient activation” were also regarded as being the same [74]. This inconsistent use of terms may impede knowledge translation and dissemination [57]. To counter this, we summarized the varied factors with unified descriptions to build a common understanding of the digital PEx–influencing factors.

Second, the intervention types and patient groups varied widely among the reviews, limiting meaningful comparisons between different studies. In addition, the digital health landscape is rapidly evolving, and the technology infrastructure is constantly shifting [41], as are the continuous updates of the UX design area. It is important to keep the influencing factors updated or adapted as the technology develops. Possibly, relevant original studies may have been excluded because of our focus on review papers. However, our approach to conducting an overarching review provides readers with a quick overview of the relevant digital PEx studies and a basis for further research.

Third, our umbrella review did not account for the multimodal relationships between subthemes or the potential overlap between subthemes within different domains. For example, different subthemes, such as “personalized design” in “interventions’ interaction design” also interconnect with “interventions’ technology” and “interventions’ functionality.” Moreover, our review process did not aim to address the question of whether some influencing factors are more important than others or how different aspects of DHIs influence them. This warrants further investigation as we suspect that differences may exist between the influencing factors, as some elements in digital health are more likely to increase or inhibit a positive digital PEx.

Finally, as we used qualitative thematic analysis to synthesize the findings and generate themes, the generated themes could have been influenced by the authors’ previous research experiences and personal understanding. By asking other researchers to repeat the coding process, the resulting themes are likely to be different. However, to minimize the potential coding bias, the generation of categories was based on the PRISM framework; 4 researchers with different backgrounds, including design, medical, and human factors, were involved in the iterative coding process, group discussion, and independent and random validation, and existing theories were used.

Further Research

The goals of this umbrella review were to systematically review the influencing factors that affect digital PEx and the design considerations for improving digital PEx that are summarized in the existing literature. We must conclude that, currently, much remains unknown, and the topic of digital PEx is relatively new. We propose 6 directions that require further research. The first direction is to develop frameworks or models that translate digital PEx–related research findings into design practices or implications. For example, in this study, we used design guidelines and a design framework to summarize the findings. The second direction is to identify those who will benefit more from which elements in DHIs and which influencing factors could be addressed by combining design constructs and design methods. The third direction is to further examine how designers understand and address digital PEx in the digital health design process. To address this, we conducted a qualitative study on how designers address digital PEx in design practice. The fourth direction is to standardize evaluation indicators, methods, or tools for assessing digital PEx; we are currently evaluating digital PEx in a parallel study. The fifth direction is to quantify the balance between the benefits and costs of developing user-friendly and validated DHIs. The sixth direction is to identify participants’ reasons for dropping out and their impact on the reported digital PEx–related results.

Conclusions

To the best of our knowledge, this is the first study to propose the term “digital patient experience” as a common phrase to describe PEx in digital health and define digital PEx by synthesizing the reported PEx or UX of varied DHIs from multiple reviews. Multimedia Appendix 4 shows more details about the structure of this study. In this review, information on influencing factors was identified and summarized into 9 categories (ie, patient capability, opportunity, motivation, intervention technology, functionality, interaction design, organizational, physical environment, and social environment). These categories were classified into positive, negative, and double-edged factors based on their positive, negative, and diverse impacts on digital PEx. Our review uncovered 4 design constructs (personalized, information, navigation, and visual design) and 3 common design methods (UCD or HCD, co-design, and inclusive design) as design considerations for addressing digital PEx. Finally, we proposed a design and evaluation framework and design guidelines to help digital health designers and developers address digital PEx throughout the entire design process.