Atopic dermatitis (AD) is a chronic inflammatory skin condition that affects up to 25% of children worldwide [1,2]. It is characterized by periods of remission and relapse due to variable, often unknown, triggers [3]. Symptoms such as itchiness or soreness can strongly affect children’s daily activities and cause sleep deprivation, which substantially undermine their quality of life (QoL) [4]. The burden of childhood AD extends to their caregivers as well [5,6]. Caregivers’ sleep can be equally affected as parents have to manage their child’s AD symptoms during the night, which can further influence their performance at home and work [7]. In addition, AD, especially childhood AD, requires skilled self-management from caregivers, including coping with a complex regime: caregivers are often required to perform extra duties, such as preparing special meals for their children, engaging in skincare regimens, and cleaning extensively to minimize outbreaks [5,8]. A child with AD can require approximately 2-3 hours of care from their caregivers daily [9]. To prevent relapses, numerous caregivers deem it necessary to cut down on leisure, defer social activities and events, and even sacrifice their jobs to take care of their children [10].

Lack of support leaves caregivers not feeling confident or even failing to manage and provide disease management support for their children’s condition in between visits to the clinic. In clinical practice, it is difficult for health care professionals (HCPs) to educate their patients adequately. Doctors have insufficient time for patient education during the tightly scheduled consultations, and there are limited health educational materials provided in the hospital setting. Caregivers and patients also have difficulty in recalling the health information following a visit to the clinic [11]. The unmet management support needs encourage patients to seek additional information from the internet or friends, and although often most helpful, such apps may also not be trustworthy or easy to use. As such, caregivers must be equipped with supporting tools that are tailored to their needs and help them to properly manage patients’ condition continuously [12].

Digital health tools are increasingly being leveraged to increase patient engagement and facilitate disease self-management for conditions such as diabetes, asthma, dementia, and cancer [13-17]. Smartphone app features can engage patients more in cultivating guidelines-recommended behaviors and providing evidence-based information and advice [18-21]. For example, the disease education feature can give patients insights into their disease condition and enable them to make informed clinical decisions [20,21]; the disease tracking feature can help users detect the possible allergens and help them eliminate aggravating factors and reduce recurrence [20]. It has also been demonstrated to improve clinical outcomes, and 80% of patients consider the service a viable alternative to in-person consultations [22]. Similar to digital management of AD, previous studies comparing AD digital management with traditional in-person care models have demonstrated no significant differences between them regarding clinical outcomes and QoL outcomes [23,24]. In addition, qualitative data showed that doctors found it helpful to use the website in consultations in a pilot randomized control trial [23]. However, when considering the substantial financial savings that digital health interventions in managing AD could reap, digital health could potentially be a promising substitute for in-person AD care [25].

Although many studies document the benefits of digital health tools in terms of supporting users to monitor, treat, and follow-up disease conditions, the quality of apps is a cause for concern, as in many cases, there is a lack of clinical evidence proving the effectiveness of apps. Many of them are not tailored to the user’s needs; this may explain the low adoption and use of apps as part of routine care in clinical practice [26,27]. Working in partnership with stakeholders in the design process can better serve app development. By doing so, developers can more easily identify desirable features for an app as well as obstacles hindering long-term usage; together, app developers and stakeholders could improve the viability, usability, and effectiveness of apps in health care services [28-30].

The co-design approach refers to patients and caregivers working in partnership with staff to improve the viability, usability, and effectiveness of health care services and is considered one of the best ways of improving patients’ experience of services and of guiding developers into understanding the users’ needs [28]. The co-design approach has been used to build self-management apps for chronic diseases such as heart failure [29] and dementia [17,29,31,32]. Co-design has been employed in AD app development before, but relevant web-based information is so sparse that previous experiments cannot easily be repeated by other researchers. Given the fact that self-management is crucial to tackling AD among children, an app involving all stakeholders with detailed procedures is needed to ensure that patients’ needs and voices are sufficiently heard. The objectives of this study are to better understand the needs of and challenges confronting caregivers and HCPs who care for children with AD and to explore the desirable features and content that smartphone apps should offer in supporting AD self-management.

Eligible participants were recruited among those of a baseline cohort in a prior study, who had provided consent to be recontacted for participation in this study [33,34]. Caregivers who met the following inclusion criteria were approached by making phone calls or sending emails: (1) adults providing unpaid care or assistance to children (age<16 years) with AD and (2) prior experience with using a smartphone app to manage health-related issues. The exclusion criteria were as follows: caregivers who did not own a smartphone or had no knowledge of mobile apps. HCPs who have experience in managing AD pediatric patients were recruited directly from the Dermatology Department at the KK Women’s and Children’s Hospital in Singapore. Digital health experts with experience in the field of digital health (app assessment or app development) were recruited face to face from the Lee Kong Chian School of Medicine, Nanyang Technological University.

Data were collected across three co-design workshops conducted in a specialist hospital in Singapore from September 2018 to March 2019 across two phases (caregivers’ session was held at 7 pm and lasted for 2 hours; HCPs’ session was held at Noon and lasted for 1.5 hours. The design session was held at 7 pm and lasted for 2 hours). The procedure of the co-design workshops is detailed in Textbox 1.

Workshops one and two adopted a focus group discussion format. Workshop one was a focus group discussion with caregivers: 2 researchers (KG and WS) facilitated the group discussion. Additional field notes were collated to capture relevant contextual information (XM and ZL). Semistructured questions were used to guide the focus group discussions. To better serve the research objectives, the semistructured questions used in the previous AD QoL study were designed to address both subjective feelings and objective findings regarding the most disturbing aspects of childhood AD. The topics raised during workshop one are shown in Textbox 2.

Workshop two was a focus group discussion involving HCPs; 2 researchers (JC and KG) facilitated the group discussion. Additional field notes were taken by XX to capture relevant contextual information. Semistructured questions were used to guide focus group discussions. The semistructured questions were designed based on feedback from caregivers in workshop one. They are shown in Textbox 3.

After completion of phase one, a final co-design workshop was conducted (facilitated by JC and XX). As the targeted user of the app is caregivers whose children have AD, the workshop only invited caregivers and digital health experts to design the wireframe prototype. HCPs were not invited in this phase, but we will seek their feedback after the app is developed. After the participants outlined various design requirements in workshops one and two, the facilitator asked participants to design the interface of the app themselves. The app had to address at least one of the top needs mentioned in the previous workshops. The workshop groups discussed and explored the ideas from the drawings and came to an agreement on the top features or ideas they wanted to include in one consolidated master sketch. Groups sketched a single layout that incorporates the top ideas. The teams regrouped and critiqued the wireframe prototype. After the feedback and revision session, the groups discussed and explored the ideas from the drawings and came to an agreement during the development of the consolidated master sketch.

The following questionnaires were administered during the co-design workshops: Demographic questionnaire and technology acceptance questionnaire (Multimedia Appendix 1, [35]) and workshop evaluation form (Multimedia Appendix 2).

Workshops were audio-recorded, transcribed, and managed using NVivo software (QSR International, version 11) [38,39]. The transcript data were first broken down into concepts and reorganized to summarize key issues from each workshop. XX, KG, and JC reviewed and discussed concepts immediately after each workshop so that emerging issues could be explored in subsequent workshops. XM transcribed the audio recordings and processed the observation logs, photos, and written products to contextualize and gain more details about the discussion. Thematic analysis was chosen to analyze the transcriptions from the workshops [22]. An inductive strategy was employed, which allowed for themes to be generated from the data, which were then organized into higher-order themes. The NVivo software program was used to store and code the transcript. First cycle coding was performed on the raw data. Subsequently, researchers (XM and KG) met and compared each other’s codes. An interrater reliability (IRR) test was performed between the researchers’ findings (XM and KG); whenever an acceptable IRR was not achieved, the inconsistencies were jointly discussed until a consensus was reached.

All eligible participants provided informed consent before the workshops. Ethical approval was obtained from the Institutional Review Board of Nanyang Technological University before the commencement of the study (Nanyang Technological University Institutional Review Board: 2018-09-053).

Patients, the public, or caregivers were not directly involved in the development of this study, but they were recruited as participants in this study.

The participant recruitment procedure and reasons for exclusion are listed in Figure 1. The characteristics of the caregivers are summarized in Table 1. Out of a total of 270 eligible caregivers, 18 caregivers participated (18/270, 6.7%). The caregivers’ sample included a diverse range of demographics (Table 1): the majority of them were women (14/18, 78%), the mean age of the participants was 37.87 (SD 7.81) years, and the children’s disease duration was 4.72 (SD 4.57) years (range: 1.66-10.83). Pediatric dermatologists (n=8) and pharmacists (n=2) were recruited for the second workshop: their mean age was 35.50 (SD 6.69) years and their mean practice time was 9.76 (SD 6.57) years. Four digital health experts with a mean age of 37.80 (SD 8.12) years attended the third workshop.

The caregivers’ perceptions of using self-management smartphone apps were assessed to test perceived usefulness, ease of use, and user satisfaction (Table 2). A total of 89% (16/18) of the participants found it easy to get mobile tracking on children’s disease, and 89% (16/18) of them felt confident in using a smartphone app to manage children’s AD. A total of 78% (14/18) of the participants believed that using a smartphone app to manage AD would increase the caregiver’s QoL.

During the workshops, various functionality requirements were discussed among participants. The most prevalent requirement across the groups was knowledge of the symptoms and on managing the symptoms, medication usage, and triggers. They also expressed a desire to brief caregivers and children about the issues that they might face in the future following diagnosis. This could potentially guide people in transitioning from raising awareness to making concrete actions and plans and developing coping strategies that could help them master everyday challenges.

The HCPs also emphasized that such information should be provided in the right amount in the local language. Excessive information provided during medical consultations may overwhelm both patients and their caregivers. However, inadequate information may also leave them feeling directionless, a feeling which induces them into “going online for relevant information which could lead doctors to spend more consultation time in correcting the wrong online [information].”

The goal is to ensure that users fully understand the requirements and thus reduce confusion over information. To accomplish this, features such as podcasts, chatbots, or web-based lectures were proposed as the potential modes for delivering such information. HCPs also highlighted that the apps should provide other patients’ stories or feedback to debunk the myths concerning or against steroids.

The other requirement discussed during the workshop was the need for patients and caregivers to communicate with HCPs. In this context, participants proposed that technology could be used throughout the whole process of the health care journey, from preparation for the clinical consultation, facilitation of clinical consultation, disease status follow-up, to real-time management after medical consultation. Here, the main goal of the technology would be to help caregivers identify and manage flare-ups; to mitigate confusion; and, at the same time, facilitate HCPs to gain a better understanding of patients’ needs and limitations.

Participants also opined that the technology should enable the user to provide support to others, in this case, by sharing knowledge and experiences with their peers in Singapore. They suggested that apps should be designed to provide social support and facilitate connectedness with others. For example, when other caregivers encounter difficulties when managing flare-ups, their peers could support them and detail their own experiences. Besides, peers can also comfort caregivers emotionally by recognizing their efforts and mitigating their guilt. This might be accomplished using a web-based forum with messaging, chat, or call functions. These were proposed as modes through which other caregivers can provide support for each other.

In addition, both caregivers and HCPs suggested that the app should be personalized. The personalization should embody 2 aspects: (1) features and content should be tailored to each individual’s needs and (2) features and content should suit the local context. In this way, the app could help them with prioritizing and making new choices and with defining their own goals and the small steps necessary for reaching these goals. During the workshops, the participants advised that the app should be personalized to address every patient’s unique circumstances. They recommended that the app should provide all the information needed as soon as a client first uses the app (quote 18 in Multimedia Appendix 3).

Image analysis was also mentioned during the workshop. It was recommended that caregivers could take photographs and use the app to analyze each photograph to measure the severity. They also advised that the app should give advice or suggestions corresponding to the extent and severity of the skin condition, so that clients can get immediate help without seeing the doctor.

The wireframe prototype contained features designed to address the challenges and confusions mentioned in the previous workshops. The must-have features and nice-to-have features recommended by participants, challenges addressed by the features, and how the final wireframe prototype reflects the features are summarized in Table 3.

As shown in the master sketch of the prototype, the prototype allows users to log in with their account (Figure 2) and record lesions’ photos of AD in different body parts and provides instructions relating to treatment, symptoms, triggers, and people’s emotions (Figure 2). The Journal section enables caregivers to track the patient’s disease condition and could prove useful to their doctor during a visit to the clinic (Figure 2).

Another core feature of the prototype is the education element. Participants designed both the chatbot feature and forum feature, both of which are for educational purposes (Figure 2). The chatbot allows users to access relevant information quickly and easily, which is presented in a clear, understandable format. The forum feature builds a bridge between caregivers and their peers by letting them exchange local information and, to some extent, allowing peers to provide informational and emotional support to users (Figure 2). Participants also designed a notification feature, through which the prototype could remind users to avoid the possible triggers around them and decrease AD relapse (Figure 2).

The workshop evaluation comments indicated that the co-design workshop was successful in creating and generating new ideas and content for smartphone app development. All the participants agreed that the workshop was organized, that its aims were clear, and its activities had been carefully prepared. All the participants felt that the workshop methods used were appropriate for the audience and said that they enjoyed the workshop. However, some issues were also identified in the evaluation feedback. For example, 21% (7/32) of participants highlighted the need to improve workshop time management and felt that the time should be better utilized; they even suggested a longer session.

For this study, we organized a series of co-design workshops, which were attended by caregivers living with children with AD, doctors, pharmacists, and digital health experts. During the proceedings, participants jointly explored their needs, preferences, and perceptions in relation to the proposed smartphone self-management app. Phase 1 identified themes that reflected caregivers’ and HCPs’ needs and concerns for AD self-management support, along with suggestions concerning the necessary features and contents of an AD self-management app; the data collected during phase one, therefore, informed phase two, which in turn comprised a smartphone app designing activity. The purpose of the interactive designing exercise was for participants to develop a wireframe prototype. The co-design workshops were designed to encourage social interaction, stimulate creativity, and thus help us to articulate the root of the clinical problems. The findings from the workshops align with those in previous studies, in that co-design interaction can foster a co-creative space and lead to considerable contributions by and involvement of people with chronic disease conditions [40,41].

The wireframe prototype has the potential to solve current AD management challenges. First, information gathered by disease diary and disease monitoring features (Figure 2) in prototype are quite comprehensive with both visual AD symptoms (by taking photos), personal feelings (sleep disturbance, itchiness), psychological impact (mood), and suspected allergens, which could be very helpful for doctors in making precise clinical decisions. In addition, involvement of HCPs in the app development process could ensure that the information provided is reliable. Third, the forum feature (Figure 2) could potentially become a peer support platform. By communicating with each other, caregivers can support each other technically and emotionally and, thus, enable a smooth transition following a consultation and empower caregivers to manage their children’s disease by making them feel competent and capable.

Some issues were also identified during the workshop. First, it is important to ensure that all the participants fully understand and can keep pace with workshop information and activities—or, to use the colloquialism, read off the same page. The participants may interpret questions or tasks differently, and there is a risk of digressing the topic or narrating the same topic repeatedly, which can lead to the workshop being less productive in terms of producing the ideas or thoughts anticipated. Second, at the start of third workshop, we delivered a briefing presentation of the current research and technology used for AD management, to encourage the participants to allow their own needs and challenges to guide app co-design. However, it should be mentioned that the briefing presentation might have limited their thinking, and they may have been inclined to follow the information shared during the briefing session.

This study can inform future research studies on co-designing smartphone apps for patients affected by chronic diseases. Participants with a heterogeneous background were invited to co-design the app. This could be considered as a strength of this study as the app design by them will be understandable and beneficial for the wider population regardless of user background. Previous studies have explored treatment needs widely: Batchelor et al [42] tried to identify the uncertainties in eczema treatment that are important to stakeholders. Teasdale et al [43] and Swallow [44] discussed the caregivers’ views of steroid usage [43,44]. However, this chapter goes further in exploring personal experiences, emotional needs, and feelings among both caregivers and HCPs and considering how their points of view could be linked to features of smartphone apps.

However, the scope of the study is limited as the target audience of the app consists entirely of caregivers whose children have AD. However, that limitation alone does not preclude us from being able to satisfy the aims of this study because the content and features caregivers mentioned are to some extent interchangeable. Some of the issues and features caregivers mentioned can also be favored for both children and adults. Second, the co-design process is not rigorously following a participatory co-design framework as continuous evaluation of the prototype was not performed [45]. However, considering that this study has gone through the main process of the framework, this study is still aligned with the overall goal of co-design methodology [29].

Another limitation is that participants were observed to discuss some topics repetitively. Thematic saturation of the workshops was probably less than satisfactory as other themes were barely discussed in phase one, and it might have emerged that we conducted more workshops. However, this limitation can be partly explained as those viewpoints raised repetitively were valued more than other topics from the caregiver’s perspective and, therefore, should be prioritized when designing an AD self-management app. In addition, as the final wireframe prototype is designed by both digital health experts and caregivers, it is difficult for us to differentiate caregivers’ perspectives and digital health experts’ opinions in phase two.

In summary, this study employed a co-design approach to jointly create an AD self-management app wireframe prototype, together with caregivers, HCPs, and digital health experts. The participants’ perceptions and preferences as well as the co-design approach used in this study form a novel contribution to the canon of literature on smartphone app design and the use of digital health tools in an engaging way. Future studies should work on app development as well as invite more stakeholders with separate sessions to test its functionality and usability.