Autism spectrum disorder (ASD) is a neurodevelopmental disorder. Children with ASD often face difficulties in initiating and maintaining conversations and find it challenging to interpret verbal and nonverbal behavior, which commonly leads to misunderstanding the intentions of others [1-4]. Children with ASD often do not spontaneously interact with peers and have difficulty making eye contact [5-7], often struggle to make friends [8], are more likely to be excluded by peers [8-15], and are more likely to be victims of peer harassment [16]. Many interventions developed for this target group focus on improving and training social skills [17]. One of the goals of social skill training is to enable these children to better cope with their peers. Although social skill training interventions demonstrate improved social skills in clinical settings, the developed social skills in training are not necessarily applied in children’s daily lives at school [18,19]. In addition, improving social skills does not necessarily mean breaking the negative bias of peers at school [20], which affects the quality and quantity of relationships.

The difficulties children with ASD face in forming and maintaining relationships with peers can lead to social fragmentation at school [21]. Research suggests a discrepancy in the quality and quantity of relationships reported by children with ASD compared with peers [22,23]. Some research even suggests that these children have the fewest friendships of all disabled groups [24]. In addition, according to the children themselves, they do not necessarily link how developing social skills contributes to forming more friendships and greater acceptance by peers at school [25]. Although one of the goals of social skill training is to better cope with peers, and social skills can improve in training, its effects do not always transfer to practice. Interventions that incorporate the peer-group context in natural settings can be beneficial for improving social skills [26] and improving relationships [27]. Nevertheless, there are very few interventions that focus specifically on this issue.

Mutually enjoyed activities that focus on similar interests can be a fitting way for individuals with ASD to connect with others [28,29]. Therefore, activity-based interventions might be preferable for individuals with ASD than instruction-based interventions [28]. Activity-based interventions provide structure and a concrete focal point, allowing interactions to unfold more naturally [30]. Children with ASD seem to have an affinity toward digital technologies owing to their linearity and discreteness and often play video games themselves [25,31]. Several studies indicate that games can provide an enhanced experience compared with more common instruction-based interventions and teaching methods [32-34]. Games offer unique features to motivate, trigger, personalize, and facilitate learning [35,36] and provide a safe context for practicing more complex skills [37]. Furthermore, games can provide immediate and consistent feedback for children with ASD [38].

In autism research, there is great interest in the application of games to pursue therapeutic goals. Within the spectrum of games, a distinction can be made between interventions that gamify certain elements and interventions aimed at providing a good gaming experience. Interventions such as LIFEisGAME [39], Secret Agent Society [40], Let’s Face It! [41], A Sunny Day [42], and TeachTown [43] are based on a gamification approach, offering a reward system for finishing more traditional therapeutic tasks. This approach has 2 concerns: (1) although game elements and therapeutic tasks are combined, they do not provide a coherent experience [44], which is known as fidelity dissonance [45], an incongruence in in-game concepts that causes a disturbance in the game experience, and (2) compared with video games on the market, these interventions offer a poor game experience, which might undermine children’s interest in playing [44].

Interventions that facilitate collaboration and social interaction within a more integrated game design are Invasion of the Wrong Planet [46] and a technological touch-activated Collaborative Puzzle Game [47], offering an environment to collaborate through puzzling. Pico’s Adventures [44] is a game based on full-body interaction to promote social initiation skills. Another example of a serious gaming intervention is Lands of Fog [48,49]. Lands of Fog is a multiuser experience, designed with and for children with ASD to foster social initiation and collaborative behaviors. Within Lands of Fog, players can explore and discover unique characters, objects, and events in a magic world by catching fireflies. The game offers an immersive environment and provides specific mechanics to foster collaboration. This game appears to be a useful instrument to facilitate conversation and promising to promote engagement, socialization, and collaboration [48], which can positively affect children’s lives at school and their relationships. The game was able to unite children with and without ASD through an enjoyable and informal common activity–based experience. The practical disadvantage of Lands of Fog is that it is difficult to implement widely in terms of equipment and the required physical space. The system is structured in a 6-meter circular arena, where a virtual world is projected on a floor.

Since its inception in Japan in 2007, escape rooms have been growing in popularity worldwide [50]. Entertainment-focused escape rooms are now available across most continents, including Europe, Asia, and America. In escape rooms, participants are given a scenario where they discover clues and solve puzzles to accomplish a specific goal in a limited time, usually escaping the room [51-53]. Participants often must work together and collaborate to fulfill different tasks successfully. Although research into the impact of escape rooms is still limited, escape rooms have already been used for various serious and learning-related purposes [54-57].

Due to the cooperative and game-like nature of escape rooms, escape rooms seem suitable for providing an informal activity–based experience for training social behavior and collaboration. Escape rooms are immersive and engaging [58], and the concept of puzzling and collaborating is fitting to integrate into a serious game. An escape game might offer a natural setting for players to communicate with their peers and can provide an emerging experience. This might allow interactions to unfold naturally. If properly designed, an escape game can offer a fun experience based on equivalent interactions and relationships, which is conditional on achieving friendships and acceptance among peers [27].

When designing a game with serious purposes, addressing the oxymoron and uniting the fun with the serious aspect is essential but challenging. Mapping out the various stakeholders’ perspectives and goals is often a necessary step because the interest of the fun element and the interest of the serious part is usually divided among stakeholder groups, especially when designing serious games for children. Experts who facilitate social skill training programs are interested in pursuing specific therapeutic goals. At the same time, children want to establish better relationships with their peers and, above all, seek a fun experience in a game. To overcome this lack of consensus, the concept of boundary objects can provide a perspective. The idea of boundary objects, which is growing in interest [59-63], was initially framed to facilitate constructive cooperation between sides or social systems in the absence of consensus [64]. Boundary objects [65] can fulfill an important function in overcoming boundaries by addressing and adapting the local needs and constraints of different stakeholders. As an in-between object, the boundary object belongs to the worlds of both stakeholder groups (Figure 1).

In the example of social skills training for children with ASD, children with ASD and experts who facilitate social skills training are connected and have a shared concern. Both parties are interested in making children function better in social situations; however, children have different goals than experts. Experts mainly focus on improving social skills, whereas children want to establish better relationships with their peers [25]. A well-designed serious game can be a novel platform to facilitate communication and a medium that strengthens relationships between sites [66] by creatively adapting and addressing user needs. Designing a serious game as a boundary object ensures that different stakeholders’ needs are addressed, which facilitates an inclusive design. The active inclusion of children’s goals and children with ASD is essential in designing a fitting serious game [44].

This paper describes an iterative design process for the development of an escape room–based serious game. The purpose of the serious game is to facilitate direct communication between high-functioning children with ASD and their peers, for the development of social skills on the one hand and strengthening relationships with peers through a fun and engaging activity on the other hand. During the design research process, we examined in small steps whether the developed prototypes are feasible and whether they have the potential to achieve the objectives of the serious game. This paper aims to provide insight into the design process and describe the development of a serious game as a boundary object.

This study is structured around the Design Research Framework (DRF; Figure 2) [67,68]. This framework facilitates the development of serious media interventions and serious games through an iterative-incremental process. The focus of these iterations shifts during the process, along with the nonlinear design steps [69].

The DRF gives direction within a design process on a more transversal level and can provide insight into where the focus in the development of an intervention lies. In our earlier research [25], we focused on assessing needs and analyzing content and context; this study focuses mainly on the construction and utilization of prototypes.

Various participatory design activities have been conducted to make the boundary object perspective applicable to the design process. As a boundary object, the serious game has to address both the therapeutic goals and needs as children’s needs while still being fun and engaging. Children with special needs can benefit from an inclusive design approach [70]. However, in practice, participation sometimes remains superficial and participatory design activities end up in sessions where children are mainly responding to, for example, visual aspects. Malinverni et al [44] proposed an inclusive design approach specifically for developing video games for children with ASD, where children and experts both provide input of specific elements that translate to game elements. Although the organization and structuring of sessions were slightly different in this study, this has been one of the central principles in the design process.

For children, forming more friendships, connecting with peers, and experiencing more acceptance by peers at school are important goals [25]. To achieve this, equality is considered an important principle. Equality is essential in forming relationships and friendships among children with ASD [27]. Further, mutually enjoyed activities that focus on similar interests can be a fitting way for individuals with ASD to connect with others [28,29]. A serious game that initiates social interaction in a structured way might be an appropriate way to address this issue. To arouse children’s and their peers’ interests, the game should also offer a fun and engaging experience. In addition, children provided input in finding an appealing theme and narrative for the game.

On the basis of an existing social skills training protocol, as described by Dekker et al [71], and literature on joint attention [72], 4 therapeutic goals have been taken as the guiding principles throughout the design process. These goals are turn-taking, cooperation, joint attention, and vocalization. The application and expression of these goals within the prototypes can be considered as the early predictors of success [73]. These principles guide the evaluation of the prototypes.

The sessions and activities described in this study took place within a 40-week school year, which runs from September to halfway through July. In this research, 4 iterations are described, consisting of developing and testing prototypes. The first 3 iterations were conducted within the first 20 weeks, always in a block of approximately 3 weeks. During the first 3 iterations, the authors developed the prototypes in 1-2 days and then tested the prototypes around schoolwork at various times within 3 weeks at schools. There were test periods and school breaks between the iterations, which made testing on location impossible. The final prototype was developed in 14 weeks by a professional development team, including the time required to form the team. This prototype was then tested with children and experts between June and July.

In this study, 4 prototypes were tested in children. After the first test, a creative workshop was held to let children brainstorm about the theme of the escape room. We started with a paper prototype supported by materials from an existing escape room board game and ended with a multiplayer serious game playable on Samsung Galaxy Tab S2 tablets. On the basis of the different goals, the prototypes were evaluated after each test. Test sessions with children were also used to test the puzzles. The goal was to compose puzzles that were challenging and at an appropriate level for the specific age group of 10 to 12 years, by evaluating whether or not children can complete each puzzle given the information presented without any extra hints.

In the fourth test session, we also used the Playground Observation Checklist [74] to evaluate the play behavior. Although this checklist comes with some considerations and more background information on the children would be essential to make a reliable interpretation of the observation, the Playground Observation Test is a useful operationalization of social play behavior. In this phase of the study, the observation list was used to determine whether there were significant differences between playing behaviors among children playing the prototype.

All tests have been carried out at the special primary schools. As there was always a larger pool of children present at the schools during the tests, there was no difference in the number of participants reported, as described in the Methods section.

In this study, we designed AScapeD, a serious game to facilitate social interaction and communication between children with ASD and their peers. AScapeD contains various elements of existing knowledge and practical insights resulting from an iterative design research process. AScapeD was deliberately designed to provide a useful activity-based experience for players. This might be an appropriate way for the target group to directly execute social skills in a safe context. The serious game is designed to be played with children with and without ASD, as this can contribute to socialization [48], can trigger peer support [77], and might increase the likelihood of transfer [45] by allowing the children to play directly in the target context.

The game has been developed and tested in small steps to ensure that it offers an immersive and meaningful experience. There was a focus on achieving a fun and challenging experience during these steps by finding the puzzles’ appropriate level of challenge for children aged 10-12 years. There was also a focus on narrative transportation [76] to involve the players over a more extended period. An escape room’s conceptual structure contributes, even in digital form, to the natural emergence of communication and equal cooperation among children. The iterative process has contributed to translating the therapeutic goals into the game in a constructive way. It was a challenge to construct a game in which communication was necessary. Players had to draw on their ability to take turns, and everybody had to have an equal role throughout the game. Everyone was in the lead for approximately the same amount of time. By using 3 rapidly constructed prototypes, each of which took less than 2 days to complete, it was possible to quickly learn which translations from goals to game mechanics were effective.

On the basis of the first results, AScapeD appears to be a promising serious gaming tool to successfully trigger social interaction and connection in a playful way between children with ASD and their peers. While playing, children with ASD participate actively and equally. During play, the play behaviors of children with and without ASD did not differ significantly. Experts recognized the skills of turn-taking, vocalization, joint attention, and the necessity for cooperation in the game. The serious game appears to be feasible to pursue the formulated goals; however, according to the experts, the children already need to be at a certain base level to apply those skills. Further research should focus on what base level is required. For effectiveness, no conclusions can be drawn yet based on the observations and the applied scale.

AScapeD results from an iterative design process structured around the DRF [67,68]. The DRF gives direction within a design process on a more transversal level, where constructing and improving prototypes in an iterative way is a central guiding theme, which was also a leading principle in this study. The DRF provides focus and direction on the central design process; however, other resources are necessary to address specific audience characteristics and mechanisms to pursue specific goals. In this study, we could quickly learn from prototypes to find a fitting and constructive translation of predefined goals into the serious game. In addition, it has been of added value to actively involve the children in the entire process and incorporate their goals and wishes in the serious game. By approaching the serious game as a boundary object throughout the entire design process, trying to address the local needs of different stakeholders, there was a constant focus on creating an inclusive design.

Involving children with a particular vulnerability in the research process is always a challenge. Nevertheless, in human-computer interaction design, it is usual to actively involve end users in a design process. Inspired by the inclusive design approach of Malinverni et al [44], we have followed this in this study by enabling children to trigger their strengths and interests and letting them play the prototypes of the serious game and let them respond to it and let them give input on specific design choices. This made it easy for children to get quickly involved in the process and made the input and feedback questions on the different prototypes very tangible.

As a result of approaching the serious game as a boundary object, the process has not led to a new tool for a specific activity system. The applied process has led to a mediating tool that contributes to different involved activity systems objects without attempting to achieve consensus between them (Figure 12). AScapeD adapts the various stakeholders’ local needs and constraints and obtains a different meaning from the various activity systems. For children with ASD, a serious game is a tool that contributes to their goal of better connecting with peers. For social skills trainers, it is a tool that allows them to see how children with ASD put social skills into practice. For special primary school teachers, it is a tool that can be of added value for the development of group dynamics. The skills included can be beneficial for each child. Turn-taking, joint attention, and cooperative skills can be beneficial for each child and the group dynamics in a classroom, even if there is no developmental issue such as with children with ASD. In conclusion, for peers of children with ASD, it is a tool to have a fun experience and establish relations with peers.

From this study, we identified 5 important insights that could be useful as guidelines for future work. The following suggestions emerged from this study:

This study’s inclusive approach offers a specific perspective for shaping innovation in health, mainly when innovation affects multiple activity systems. Many frameworks on implementation in health have been developed within disciplines [81]. These frameworks are, therefore, very suitable for innovation within an activity system to improve its tools. Innovation in more complex contexts involving multiple stakeholders often leads to high failure rates or a lack of impact [82]. The boundary object perspective focuses on bridging the gap between different goals among stakeholders.

For the development of the serious game, different groups of participants participated in the study. The serious game seems promising based on the first results. However, further research is necessary to investigate whether the game is applicable in practice and whether the serious game is of added value in the longer term. On the basis of this study, the game seems suitable for high-functioning children with ASD, given that they could actively participate in the game, had an active role in solving the puzzles, and could actively communicate and collaborate during the game. The puzzles’ level is tailored to children who have the IQ to follow a regular school curriculum. Scalability is, therefore, more likely to occur toward application in other contexts, where the experts call the application in a home setting and schoolteachers, for example, the application in different classes to strengthen group dynamics, rather than toward application to other target groups with, for example, children with a different diagnosis of autism or a lower IQ.

More long-term research with the serious game will have to show whether the serious game can fulfill its potential in the longer term. In a follow-up study focusing on effect rather than feasibility, the correct background variables need to be mapped out more closely. In addition, the application of a measuring instrument requires more consideration.

This paper presented the iterative design process of AScapeD, a serious game based on the concepts of an escape room. AScapeD triggers social interaction and connection in a playful way between children with ASD and their peers. An escape room’s conceptual structure contributes to the natural emergence of communication and cooperation between children within a fun and engaging activity. AScapeD results from an iterative design process, where many insights were gained by learning from the application of rapidly constructed prototypes in practice. Applying prototypes contributed to finding a constructive translation of children’s goals and therapeutic goals into the serious game. Children were actively involved in the study by participating in playful test sessions that triggered their strengths and interests by letting them play the serious game prototypes and allow them to respond to it. By approaching the serious game as a boundary object throughout the entire design process, attempting to address the local needs of different stakeholders, there was a constant focus on creating an inclusive design.