Medical progress increases the availability of treatment options and life-sustaining possibilities. Consequently, it also increases the necessity to make choices between medical options that have different impacts on patients’ future lives. When patients are not capable of making those decisions, health care professionals and surrogates are requested to choose, even if they have no clue about the patient’s values and priorities. To avoid such distressful and suboptimal situations, early advance care planning (ACP) is increasingly being recognized as an important condition for adequate treatment [1,2]. ACP is “the ability to enable individuals to define goals and preferences for future medical treatment and care, to discuss these goals and preferences with family and health care providers, and to record and review these preferences if appropriate” [3]. An ACP procedure may result in written advance directives (AD); however, for such document to be of any use, it remains important that it is precise and that patients discuss its content with their surrogate decision-maker and update it regularly [4].

Despite the acknowledged importance of ACP, few patients discuss their fears, priorities, and the type of care they would like to receive with their families and professional caregivers [1,5,6]. This is partly because of a lack of knowledge about ACP, a lack of recognition that ACP is relevant for them, and a lack of assistance in this psychologically heavy process [6-8]. Studies also indicate that even health care professionals find it difficult to initiate such discussions [9-11].

ACP can be seen as a process of behavior change involving steps, including awareness and knowledge acquisition, thinking and commitment to act (eg, talking to someone and writing AD), action, and regular updates [12,13]. It has been suggested that patient-centered, computer-based infographics could enhance interest, understanding, recalling, contemplation, and actual sharing of decisions related to ACP [14]. They may assist in the ACP process and save clinicians’ time, as patients may be able to obtain relevant information and input to start the ACP process in a family context. Free web-based ACP tools have been developed, with promising preliminary results [15-20]. Mobile apps for ACP are available in some countries. Their potential is recognized; however, existing products have limited features and operate mostly in English only [21,22]. We did not find comparable tools available in French. A recent review concluded that overall, mobile apps provide insufficient content and features: they are helpful to users who are ready to complete AD rather than those seeking to learn about the ACP process. Moreover, they are poor in terms of design quality, layout, and functionality. Only one app (in English) was assessed as “fairly easy to read” [21].

To support the ACP process for French-speaking Swiss residents, our interprofessional research team comprising ethicists, physicians, nurses, patients as partners, information technology professionals developed a free and easy-to-use solution, an app called Accordons-nous (Let’s agree).

We developed the content and structure of the app between January 2019 and July 2021 using mixed methods—a Delphi procedure and multiple user tests involving laypeople and patients as partners. Multimedia Appendix 1 [23-25] provides the details about the process of making the app.

Accordons-nous aims to address known barriers to ACP: lack of awareness; insufficient health literacy; difficulty in starting the discussion; identifying one’s own preferences, values, and goals of care; and writing personalized and comprehensive AD and updating them regularly within a smooth process.

The app is adapted to the local context of French-speaking Switzerland, designed to be used by patients and their family caregivers independently at home or to support ACP discussions with health professionals. Its content is adapted to the health literacy of ordinary patients. It includes (1) content to make users aware of the importance of anticipating health-related decisions and motivate them to engage in their ACP; (2) essential information on the ACP procedure, legal issues, definitions of technical terms in a format, and vocabulary accessible to common users; (3) discussion prompts to help users engage in ACP discussions with their family, friends, and professional care providers; (4) a card game to help patients clarify their values and priorities in life and reflect on the conditions that would make life not worth living; and (5) a detailed and comprehensive advance directive form that can be tailored to individual life and health situations.

The app, developed in Angular 11 [26], is structured in 3 main sections (Figures 1-3)—Je m’informe (I get informed), J’en parle (I talk about it), and J’écris (I write)—accessible by clicking on a menu at the bottom of the screen. Each section contains several pages accessible via a drop-down menu.

The primary objective of this study is to measure whether Accordons-nous is easy to use and understand by common users. The secondary objectives are to identify the usability problems encountered by common users and to collect their critical opinions on the relevance of the tool—do they think that the app helps to engage in ACP procedures, and would they recommend the app to its target populations?

Nielsen and Landauer [27] reported that 98% of usability problems could be detected through feedback from 10 users. On the basis of this, we aimed to obtain at least 10 complete responses. We asked the Patients as Partners Project at the Geneva University Hospitals (Hopitaux Universitaires de Genève [HUG]) to recruit patients fulfilling our criteria. Participation was free and voluntary. The patients received and signed informed consent forms before the study.

We used the following inclusion criteria:

Being incapable of understanding the meaning of the questions and tasks required for the study after 3 iterations of the task instructions was the only exclusion criterion.

We submitted our research protocol to the Geneva Commission Cantonale d’Ethique de la Recherche (Req-ID 2020-01397), who decided that the project does not fall within the scope of the Human Research Ordinance and therefore alleviated the need for ethical approval (decision date 12.1.2020). This study was conducted in accordance with the Declaration of Helsinki. Willing participants were informed of the study aims and procedures and signed an informed consent form. During the data collection and analysis phase, only three authors (CS, FE, and CC) had access to the recorded interviews. The recordings were destroyed after analysis. No identifying information was kept in the remaining work files.

We conducted a usability test [28] using a think-aloud method [29] comprising 7 tasks. Moreover, we collected quantitative descriptive data using a questionnaire to obtain information about participants’ endorsement of the app and their evaluation of the app’s usability and relevance for ACP.

We conducted the usability test sessions on the web using the Zoom video conferencing platform. Ahead of the test day, participants received the information and consent sheet and the instructions on how to log on to Zoom by email. Once connected, the experimenter (CS) recalled the procedure as described in the information and consent sheet, answered the participants’ questions, and asked for verbal (recorded on Zoom) or written (to be sent after the test) consent. The experimenter also expressed her commitment to respect the study protocol. Verbal consent was recorded separately to protect participants’ anonymity during the analysis stage.

Next, the experimenter instructed participants to install the test version of the Concerto HUG app that contains the module Accordons-nous on their personal smartphone or tablet. The experimenter then instructed the participants to connect their phone or tablet to the same Zoom conversation and share their phone or tablet screen. In this way, while discussing with the experimenter in front of their computer, participants could simultaneously navigate the app via their smartphone or tablet, and the experimenter could see how participants did so (owing to the screen sharing function). Precautions were taken beforehand to avoid personal notifications appearing during the screen sharing.

The experimenter then started the actual test. To ensure that participants fulfilled the inclusion criteria, she began by asking some demographic and technical questions (age, gender, and type of phone or tablet used), as well as participants’ understanding of the topic of AD (Multimedia Appendix 1; general questions).

Next, to ensure that participants did not need to express their personal choices and preferences regarding ACP, the experimenter asked the participants to put themselves in the following fictional scenario:

Participants were then successively given 7 tasks to complete in the Accordons-nous module (Table 1). The tasks were designed to test users’ ease while completing the main tasks that a common user of the app is meant to handle and identify the strengths and weaknesses of the app.

While completing the tasks, participants were asked to express all their thoughts verbally (think-aloud), as if they were speaking to themselves. If participants forgot to share their reasoning aloud, the examiner reminded them to share their thoughts with her after 15 seconds. Otherwise, the examiner would not intervene. If the task was too complex and the participant felt lost for more than a minute or decided to give up, the examiner offered assistance by providing hints to help the participant move forward in solving the task. When required, the experimenter could provide four types of help: (1) information (eg, explanation of how the participant arrived where they are or why the module includes certain content), (2) motivating question (eg, “Right, and where would you go now?”), (3) answer (if the participant asked a question related to the exploration), and (4) guidance (eg, if the participant lingered too long on an irrelevant page, tell him or her what the next step would help find the answer).

At the end of each task, we asked participants, “Does the path to fulfil the task seems logical to you?” and “Do you have any suggestions for improvement?” to receive their qualitative feedback. Once the 7 tasks were completed, participants could easily browse the app at their ease if they wished to discover other functions or pages. To conclude, the experimenter asked a series of sociodemographic questions and questions that included items from two validated scales (System Usability Scale [SUS] [30] and Mobile App Rating Scale [MARS] [31]; Multimedia Appendix 1).

We obtained responses from 10 participants for all tasks and questions. During the study, we had to exclude 2 more participants: one because of technical problems (the sound recording did not work, making it impossible to analyze the results) and a second as he failed to understand 5 out of the 7 tasks that they were asked to perform, despite 3 iterations of the instructions.

Our data set included 60% (6/10) women and 40% (4/10) men of various ages (youngest: 31 years, oldest: 68 years; mean 50.8, SD 14.6 years). Of the 10 participants, 9 (90%) used devices running on iOS, and 1 (10%) used a device running on Android. All participants used these devices daily and showed minimal understanding of the notion of AD, thereby fulfilling our inclusion criteria.

The overall Scolarius score of the app was 87.43/250, meaning that the app was readable for people with an elementary school level of education. However, some pages of the app require high school–level education. Table 3 provides the details of the score.

The usability test revealed a series of issues that were addressed, as described in Table 6. Accordons-nous is an app in an app; that is, a module inserted in the larger host app Concerto. For this reason, some navigability issues could only be addressed by the Concerto team. Table 7 describes the issues that we reported to Concerto and how they have been addressed.

Accordons-nous is the first French-language mobile app for supporting the ACP process. In contrast to existing apps worldwide, which are limited in many respects [21], it includes materials and functions that facilitate each stage of the ACP process. It provides all the relevant information in an accessible language. It includes tools and prompts to facilitate contemplation and discussions about the goals of care and end-of-life issues. It provides guidance on writing or updating comprehensive and personalized AD in a simple process. In line with professional recommendations [22], it can be used as an icebreaker for starting discussions within families and as a follow-up to direct conversations with health professionals.

In our usability test, participants succeeded in the 7 think-aloud tasks in 80% (56/70) of the cases without help from the experimenter. However, in 16% (11/70) of the cases, help was needed, indicating that the app needed some improvement. An analysis of the navigational errors and difficulties encountered revealed no usability catastrophe but a series of major usability problems. These can be addressed with minor modifications (Table 6).

Our graphs indicate that participants performed better at completing the tasks over the course of the test, although the latter tasks were not necessarily easier to complete. We think that participants gradually became familiar and at ease with the logic of the app, which indicates that a few minutes of use are enough to grasp our tool. The easy-to-use character of the tool is particularly important [15] as ACP is mainly relevant for older patients [3] who are less used to digital devices. Our oldest participants (aged 66 and 68 years) fulfilled the think-aloud tasks well, indicating that age was not necessarily a limiting factor for using our tool. This is in line with previous results indicating that older patients are likely to use a platform for ADs with adequate design and support [15]. Moreover, as we designed Accordons-nous as a tool to help engage in discussions with family and health care providers, even patients who are not accustomed to digital devices may obtain support from surrounding care providers.

The app Accordons-nous obtained high scores on readability (overall score of 87.4 on the Scolarius test, corresponding to elementary school level). This is one of the best scores compared with other similar apps internationally [21]. However, the language difficulty varies across different pages of the app. Owing to the inherent complexity of the ACP topic, this difficulty can only be partially addressed; we ensured that all technical terms in the app were clickable to pop up the definition and added 2 introductory motion design videos summarizing the main information in simple terms and with illustrations.

Overall, our users provided very positive feedback on Accordons-nous. They assessed the app as likely to raise awareness of the importance of ACP and AD, increase knowledge of such topics, change attitudes and behaviors intention toward improving their own ACP, and encourage seeking further help to fulfill one’s ACP. Moreover, participants expressed a clear willingness to recommend the tool to relevant stakeholders, including patients, health professionals, and patients’ caring relatives. These are exciting results indicating that Accordons-nous is relevant and contains easily understandable content.

However, among the few critical views expressed, it is worth noting that 10% (1/10) of patients disagreed that Accordons-nous may encourage further help seeking in engaging in ACP, and 30% (3/10) had no opinion on that question. This result may indicate that patients expect the tool to offer the possibility of sending direct calls for help (eg, sending a message to a task force of professional ACP facilitators). Indeed, during the test, 20% (2/10) of participants erroneously thought at first glance that the section I talk about it would provide such direct support service. For practical and organizational reasons, we were not able to provide direct access to private counseling with Accordons-nous, which is a limitation of the app.

Our analysis of the main problems encountered by participants while completing the 7 think-aloud tasks indicates that some participants struggled with cognitive load: they failed a task or needed help to complete it as their attention was caught by the high density of information contained in some pages of the app. For instance, some participants got distracted by the long list of works of art that they discovered on the page works of art.

Other difficulties were because some participants did not spontaneously grasp the logic of navigating the app. In particular, some participants did not immediately see that the app contained 3 sections accessible via the navigation bar at the bottom of the screen (this was needed to complete the task conversation starters), whereas other participants encountered difficulties with the logic of the drop-down menu that allows navigating different pages (this was needed to complete tasks legal obligation and conversation starters). These difficulties indicate that some improvements needed to be made to the tool, especially for users who are not familiar with smartphone apps.

To address the main navigation difficulties and cognitive load because of the high information content, we included 2 introductory motion design videos in the app. In the first video [36], we explained the aim of the app, why ACP is useful, how the app is structured (3 sections), and the main navigation functions (menu to navigate the 3 main sections and drop-down menu buttons). In the second video [23], we explained how confidentiality issues are managed and how to write and store ADs with or without the help of the app.

Regarding task data confidentiality, several participants struggled to find the information as they searched it in the I get informed section rather than in the I write section. For this task, half of the participants reported that the path to complete it was more or less logical (rather than very logical). This indicates that data about confidentiality should be duplicated in both sections to be more easily found. This is particularly important as previous studies indicate that users are concerned about privacy issues [37]. We adapted the content of the app accordingly (Table 6).

Our results must be interpreted while considering the limitations listed in Textbox 2. Some limitations are common to usability studies, whereas others are minor issues related to small deviations from the original study protocol.

The tool we developed is a novel solution for promoting ACP and AD. Accordons-nous is the first French-language mobile app developed by an interdisciplinary team of professionals in collaboration with target users. It includes a variety of content for prompting discussions related to medical emergency situations and end-of-life issues. It provides support for writing and easily updating AD on a smartphone or tablet. Considering the complexity and sensitivity of the process of ACP and given that we expect most users to be older people, we put special emphasis on producing easy-to-understand information, discussion prompts, and simple navigation principles. The results of our usability test with patients were very satisfying and helped us make the necessary final adjustments to our tool before making it available to the public. Further usability and efficacy tests involving health care professionals would help define whether the tool is also suitable for this population.